Effects of sugar and fiber consumption in minority adolescents and self-tracking as a potential dietary intervention tool

PDF

Download

Open document

Flip pages

Summarize

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content Effects of sugar and fiber consumption in minority adolescents
and self-tracking as a potential dietary intervention tool

by
Gillian A. O’Reilly

A Dissertation Presented to the
FACULTY OF THE USC GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(PREVENTIVE MEDICINE: HEALTH BEHAVIOR RESEARCH)

August 2016

i

Acknowledgements
I would like to extend my gratitude to my committee members, Dr. Donna Spruijt-Metz,
Dr. David Black, Dr. Jimi Huh, Dr. Jennifer Unger, and Dr. Jaimie Davis, for their guidance and
support throughout my pursuit of my doctoral education and during the dissertation process. I am
thankful to Dr. Unger and Dr. Davis for insights and feedback on this work, and to Dr. Huh for
being generous with her time, encouragement, and statistical knowledge over the past several
years. I am very grateful to my advisors, Dr. Spruijt-Metz and Dr. Black, for being such
dedicated mentors. They have consistently pushed me to think harder and aim higher throughout
my doctoral training, and their guidance has helped me to grow both personally and
professionally.
I would like to acknowledge the Health Behavior Research program for providing a
supportive and nurturing learning environment that allowed me to thrive as a student and
researcher. I am particularly thankful to Marny Barovich for being so helpful in countless ways
over the past several years. Thank you to the friends I have made at USC, especially Kimberly
Miller, Stephanie Dyal, Lauren Martinez, Eleanor Tate Shonkoff, and Cheng Wen, for providing
me with friendship and support throughout this experience and for making graduate school
enjoyable and fulfilling.
Finally, I am deeply grateful to my family for their support. I thank my parents, Ann and
Ray Mallen, for providing me with every opportunity to succeed and helping me to pursue my
ambitions. I am grateful to my sister Maeve O’Reilly, for all of the wisdom and encouragement
she has imparted to me throughout this process. I owe my greatest thanks to my husband Adam
Gentner for all that he has done to help me through this journey. I could not have accomplished
so much without his unwavering patience, kindness, and love.
ii

Table of Contents
List of Tables .................................................................................................................................. v
List of Figures ................................................................................................................................ vi
Abstract ......................................................................................................................................... vii
Chapter 1: Introduction ................................................................................................................... 1
Background & Significance ........................................................................................................ 1
Prevalence of overweight and obesity in the United States. ................................................... 1
Disproportionate prevalence of overweight and obesity in minority populations. ................. 3
Implications of overweight and obesity and public health burden. ........................................ 4
Etiology of obesity. ................................................................................................................. 7
Contributions of dietary sugar and fiber intake to overweight and obesity. ......................... 14
Sugar and fiber consumption: what is known about their effects. ........................................ 17
Self-tracking as a potential intervention approach to support change in dietary behaviors. 22
Gaps in the Literature................................................................................................................ 24
Introduction to the Dissertation Studies .................................................................................... 27
Specific aims and hypotheses ............................................................................................... 28
Theoretical model for the dissertation. ................................................................................. 31
Data source and study samples. ............................................................................................ 33
Chapter 2: Effects of high sugar/low fiber versus low sugar/high fiber meal consumption on
subsequent ad libitum sugar intake in overweight minority adolescents ...................................... 34
Introduction ............................................................................................................................... 34
Planned Specific Aims and Hypotheses ............................................................................... 39
Methods..................................................................................................................................... 41
Study Sample ........................................................................................................................ 41
Experimental Design ............................................................................................................. 42
Measures ............................................................................................................................... 45
Statistical Analyses ............................................................................................................... 49
Results ....................................................................................................................................... 53
Discussion ................................................................................................................................. 57
Strengths and Limitations ..................................................................................................... 62
Conclusions ........................................................................................................................... 63

iii

Chapter 3: Impact of negative mood on hunger ratings in response to high sugar/low fiber versus
low sugar/high fiber meals ............................................................................................................ 64
Introduction ............................................................................................................................... 64
Specific Aims and Hypotheses ............................................................................................. 69
Methods..................................................................................................................................... 70
Study Sample ........................................................................................................................ 70
Experimental Design ............................................................................................................. 72
In-lab Procedures .................................................................................................................. 73
Measures ............................................................................................................................... 74
Statistical Analyses ............................................................................................................... 76
Results ....................................................................................................................................... 80
Discussion ................................................................................................................................. 85
Conclusions ........................................................................................................................... 91
Chapter 4: Self-tracking of weight-related behaviors: a promising intervention approach for
sustaining weight-related behavior change ................................................................................... 93
Introduction ............................................................................................................................... 93
Specific Aims and Hypotheses ............................................................................................. 97
Methods..................................................................................................................................... 98
Study Design ......................................................................................................................... 98
Study Sample ...................................................................................................................... 100
Measure ............................................................................................................................... 101
Data Analysis ...................................................................................................................... 103
Results ..................................................................................................................................... 105
Sample Characteristics ........................................................................................................ 105
Quantitative Results: Differences Between Self-Trackers and Non-Trackers ................... 105
Qualitative Results .............................................................................................................. 106
Discussion ............................................................................................................................... 122
Study Limitations ................................................................................................................ 126
Conclusions ......................................................................................................................... 126
Chapter 5: Summary and Conclusions ........................................................................................ 128
Summary of Aims and Findings ............................................................................................. 128
Implications and Future Research ........................................................................................... 134
Limitations .............................................................................................................................. 137
Contribution to the Literature ................................................................................................. 139
iv

Alphabetic Bibliography ............................................................................................................. 140
Appendix A. FAME study ad libitum snack tray contents ......................................................... 156
Appendix B. Study 3 Interview Protocol .................................................................................... 157
Appendix C. Study 3 Questionnaires .......................................................................................... 159
Appendix D. Study 3 Coding Scheme ........................................................................................ 167

v

List of Tables
Table 2-1. FAME study sample characteristics .................................................................42
Table 2-2. HSLF and LSHF test meal contents .................................................................43
Table 2-3. HSLF and LSHF test meal nutrient compositions ............................................44
Table 2-4. Data structure for dyadic multilevel analyses ..................................................52
Table 2-5. Snack period negative mood, habitual sugar intake, and impulsivity sample
characteristics .....................................................................................................................54
Table 2-6. Results from within-person mediation analysis examining the mediating
effect of negative mood on the relationship between meal condition and ad libitum
sugar intake ........................................................................................................................55
Table 2-7. Results from model examining the moderating effect of impulsivity on
the relationship between meal condition and ad libitum total sugar intake .......................56
Table 2-8. Results from model examining the moderating effect of habitual total
sugar intake on the relationship between meal condition and ad libitum total
sugar intake ........................................................................................................................57
Table 3-1. FAME study sample characteristics .................................................................71
Table 3-2. HSLF and LSHF test meal contents .................................................................73
Table 3-3. HSLF and LSHF test meal nutrient compositions ............................................73
Table 3-4. Results from preliminary analyses of ICC and growth trajectory of hunger
during LSHF meal condition .............................................................................................79
Table 3-5. Results from preliminary analyses of ICC and growth trajectory of hunger
during HSLF meal condition .............................................................................................79
Table 3-6. Average negative mood scores, hunger scores, and grams of habitual total
sugar intake .......................................................................................................................81
Table 3-7. Results from mixed model HSLF condition .....................................................84
Table 3-8. Results from mixed model LSHF condition .....................................................85
Table 4-1. Sample characteristics ....................................................................................105

vi

List of Figures
Figure 1-1. Trends in overweight and obesity among U.S. adults from 1960-2012 ............1
Figure 1-2. Trends in obesity among U.S. children and adolescents from 1970-2012 ........2
Figure 1-3. Theoretical model of the dissertation research................................................32
Figure 2-1. Study 1 conceptual model ...............................................................................39
Figure 2-2. Visual analogue scale (VAS) for negative mood ...........................................47
Figure 3-1. Visual analogue scale (VAS) for hunger ........................................................74
Figure 3-2. Visual analogue scale (VAS) for negative mood ............................................75

vii

Abstract
Diets that are characterized by high sugar and low fiber intake are widespread in the
United States, and are particularly prevalent among African American and Hispanic adolescents.
Sugar consumption may have physiological, psychological, and behavioral effects that reinforce
subsequent sugar intake at the detriment of more healthful foods, such as those high in fiber.
Examining key acute effects of sugar consumption in overweight minority adolescents may
reveal targets for improving dietary interventions. A promising tool that could supplement
dietary intervention efforts is self-tracking. However, self-tracking adherence tends to be low and
decline over time, so research is needed that provides insights into factors that can promote long-
term self-tracking adherence. The objectives of this dissertation were to: 1) investigate the acute
impact of high sugar/low fiber (HSLF) versus low sugar/high fiber (LSHF) test meals consumed
in a laboratory setting on subsequent ad libitum sugar intake in a sample of overweight minority
adolescents, 2) examine the impact of negative mood on hunger ratings during the HSLF and
LSHF test meal conditions in the same sample of adolescents, and 3) use a mixed methods
approach to explore factors that facilitate self-tracking adherence in a sample of individuals who
self-track weight-related behaviors.
Findings from study 1 demonstrated that consumption of HSLF and LSHF test meals had
different effects on the amount of ad libitum sugar consumed by overweight minority
adolescents in the laboratory setting. Participants consumed an average of 8 more grams of sugar
during the ad libitum snack period that was provided after condition-specific breakfast and lunch
meals when they were in the LSHF condition compared to when they were in the HSLF
condition (p = 0.002). Contrary to the hypotheses, negative mood reported during the snack
period did not mediate the relationship between meal condition and ad libitum sugar intake in the
viii

LSHF condition. Additionally, the amount of sugar participants reported habitually eating in
daily life did not moderate the meal condition-ad libitum sugar intake relationship. However,
impulsivity did moderate the relationship; higher impulsivity was associated with higher ad
libitum sugar intake in both meal conditions (LSHF: p<0.01; HSLF: p<0.01). These findings
suggest that 1) when participants were restricted from eating sugar during the LSHF breakfast
and lunch meals, they compensated by eating more sugar when they were later given a choice of
foods to eat during the snack period, and 2) higher impulsivity predisposed participants to
consume more sugar when they had free access to sugar during the snack periods, regardless of
whether they were provided access to sugar or restricted from sugar intake during the breakfast
and lunch tests meals.
Study 2 findings demonstrated that negative mood was associated with perceived hunger
during both the HSLF and LSHF meal conditions. At the beginning of the in-lab visit for both
meal conditions, when participants were in a fasting state, greater negative mood was associated
with greater perceived hunger (HSLF: p<0.05; LSHF: p<0.05). Perceived hunger decreased
throughout the day during both meal conditions, and negative mood was associated with the
change trajectory of perceived hunger during the LSHF condition. Specifically, during the LSHF
condition, greater negative mood was associated with a more pronounced decline in perceived
hunger throughout the day (p<0.05). Contrary to the hypotheses, habitual total sugar intake did
not moderate the relationship between negative mood and perceived hunger. These results
demonstrated that negative mood was associated with perceived hunger during the test meal
conditions, and that direction of this relationship differed based on whether the relationship was
captured in a cross-sectional or repeated measures manner.
ix

The third study of this dissertation used a mixed methods qualitative/quantitative
approach to discover factors that contribute to self-tracking adherence. The quantitative
component of the study examined how a group of self-trackers differed from a group of non-
trackers on key characteristics, and found that the self-tracking group had higher autonomous
motivation and higher competence for engaging in healthful dietary and exercise behaviors
(motivation dietary: p<0.0001; motivation exercise: p<0.05; competence dietary: p<0.01;
competence exercise: p<0.01). The qualitative component of the study used in-depth interviews
to gain insights into factors that help adherent self-trackers initiate and maintain a practice of
self-tracking weight-related behaviors. Findings provided useful knowledge about strategies that
aid self-tracking adherence, including paying attention to immediate feedback from tracking
tools and reviewing past self-tracking data to understand goal progress. This first exploratory
study of factors that facilitate self-tracking adherence revealed many insights that remain to be
tested in future intervention research.
This dissertation makes important contributions to the body of research on the effects of
dietary intake in overweight adolescents and to the emerging self-tracking literature. Overall,
findings from the three studies that comprise this dissertation demonstrate that sugar-rich, fiber-
poor meals can have acute negative effects in overweight minority adolescents, and that self-
tracking adherence is modifiable, so it has the potential to be taught to adolescents and utilized to
aid dietary behavior change in future interventions.
1

Chapter 1: Introduction
Background & Significance

Prevalence of overweight and obesity in the United States. Overweight and obesity
has been a pressing public health issue in the United States for several decades. Overweight and
obesity rates increased rapidly from the 1970s until recent years, and remain alarmingly high
(Flegal, Carroll, Ogden, & Curtin, 2010). Figure 1-1 shows trends for overweight and obesity
among adults according to the Centers for Disease Control and Prevention (CDC) (Fryar,
Carroll, & Ogden, 2015). Figure 1-2 shows trends for obesity among children and adolescents
according to the CDC (Fryar, Carroll, & Ogden, 2014).

Figure 1-1. Trends in overweight and obesity among U.S. adults from 1960-2012

2

Figure 1-2. Trends in obesity among U.S. children and adolescents from 1970-2012

According to the latest available National Health and Nutrition Examination Survey (NHANES)
data, 31.8% of youth aged 2-19 years and 68.5% of adults in the U.S. are overweight or obese
(C. L. Ogden, Carroll, Kit, & Flegal, 2014). Of those, 16.9% of youth and 34.9% of adults meet
body mass index criteria for obesity (C. L. Ogden et al., 2014). Evidence indicates that trends in
obesity growth are beginning to slow; comparisons of obesity prevalence from 2003-2004 and
2011-2012 NHANES data show signs of stabilizing rates (Finkelstein et al., 2012; Flegal,
Carroll, Kit, & Ogden, 2012; Flegal et al., 2010; Cynthia L. Ogden, Carroll, Kit, & Flegal, 2012;
Skinner & Skelton, 2014). Obesity rates may even be declining among some populations and in
certain geographic locations, including low-income preschool-aged children (Pan, Blanck,
Sherry, Dalenius, & Grummer-Strawn, 2012), young children in Massachusetts (Wen et al.,
2012), and youth in Pennsylvania (Skinner & Skelton, 2014). In contrast to young children,
however, the overall prevalence of overweight obesity in older youth and adults has remained
high and unchanged over the past decade (C. L. Ogden et al., 2014). Some estimates even project
that overweight and obesity could continue to grow; forecasts predict that the prevalence of
3

obesity could increase by as much as 33% by 2030 (Finkelstein et al., 2012). Despite public
health efforts on multiple fronts, overweight and obesity prevalence in the general population,
and in certain subpopulations, is proving to be difficult to reduce. This is particularly salient for
certain subpopulations of youth and adults.
Disproportionate prevalence of overweight and obesity in minority populations. The
problem of overweight and obesity disproportionately affects African American and Hispanic
groups. NHANES 2011-2012 data indicates that 76.2% of African American adults and 77.9% of
Hispanic adults are overweight or obese, compared to 67.2% of non-Hispanic white adults (C. L.
Ogden et al., 2014). Of those, 47.8% of African American adults and 42.5% of Hispanic adults
are obese, compared to 32.6% of non-Hispanic white adults (C. L. Ogden et al., 2014). This
disparity is also seen in adolescents. In the 12-19 year age group, an estimated 39.8% of African
American adolescents and 38.1% of Hispanic adolescents are overweight or obese, compared to
31.2% of non-Hispanic white adolescents. An estimated 22.1% of African American and 22.6%
of Hispanic adolescents are obese, compared to 19.6% of non-Hispanic white adolescents (C. L.
Ogden et al., 2014). The high prevalence of overweight and obesity among adolescents and the
disproportionate prevalence among minority youth is a particularly important public health issue
because overweight and obesity in youth tends to track into adulthood (Singh, Mulder, Twisk,
Van Mechelen, & Chinapaw, 2008), and adolescence is a critical period for establishing lifelong
weight-related behaviors (Baranowski et al., 2000; Li & Wang, 2008). Thus, the particularly high
prevalence of overweight and obesity among minority adults underscores the need to address
overweight and obesity in minority adolescents. The need to address this issue in these
populations is also highlighted by evidence that overweight and obesity during adolescence is
associated with greater risk for many weight-related health issues in youth and adulthood.
4

Implications of overweight and obesity and public health burden. Overweight and
obesity are linked with a variety of comorbidities. Low health-related quality of life has been
found to be associated with overweight and obesity across age groups (Ford, Moriarty, Zack,
Mokdad, & Chapman, 2001; Swallen, Reither, Haas, & Meier, 2005; J. Williams, Wake,
Hesketh, Maher, & Waters, 2005). Risk for mood disorders such as major depression, bipolar
disorder, anxiety, and substance use disorders is higher for obese compared to normal weight
youth and adults (Mustillo et al., 2003; Onyike, Crum, Lee, Lyketsos, & Eaton, 2003; Petry,
Barry, Pietrzak, & Wagner, 2008; Reilly et al., 2003; Simon et al., 2006). Overweight and
obesity are also independent risk factors for a host of serious physical health conditions including
cardiovascular disease (Poirier et al., 2006; Reilly et al., 2003; Van Gaal, Mertens, & Christophe,
2006), hypertension (Faloia, Giacchetti, & Mantero, 2000; Sorof & Daniels, 2002), asthma
(Beuther, 2009; Beuther & Sutherland, 2007), nonalcoholic fatty liver disease (Angulo, 2007;
Fabbrini, Sullivan, & Klein, 2010), and type 2 diabetes (Michael I. Goran, Ball, & Cruz, 2003;
Kahn, Hull, & Utzschneider, 2006). Such diseases, which were known only to occur in adults in
the past, have now become common among children and adolescents because of the rise of
overweight and obesity in youth (Centers for Disease, Prevention, Centers for Disease, &
Prevention, 2011; Yanovski & Yanovski, 2011). Many studies have also shown that obesity
increases risk for certain types of cancer, including pancreatic cancer, colon cancer, and breast
cancer (Renehan, Tyson, Egger, Heller, & Zwahlen, 2008; Wolk et al., 2001). Further, excess
adiposity is associated with a substantial increase in risk for mortality; relative risk for death
increases across the body mass index (BMI) range of overweight and is even higher for obese
individuals (Calle, Teras, & Thun, 2005). This indicates that overweight and obesity represent a
serious health issue across the lifespan.
5

Overweight and obesity are particularly critical health issues during youth because of
consequences that can impede development and affect future risk for health conditions. Poor
self-esteem and adverse impacts on academic performance are more prevalent among overweight
and obese adolescents compared to normal weight adolescents (Swallen et al., 2005). Such
consequences hinder educational attainment and lifetime earnings (Gortmaker, Must, Perrin,
Sobol, & Dietz, 1993; Karnehed, Rasmussen, Hemmingsson, & Tynelius, 2006), underscoring
the life-long, adverse impact of overweight and obesity on overall well-being. Growing evidence
indicates that being overweight or obese during youth increases the risk for psychological and
physical health issues in adulthood. For example, the longitudinal Northern Finland 1966 Birth
Cohort Study, which followed participants from 14 years to 31 years, found that individuals who
were obese at 14 years old were more likely to have physician-diagnosed depression at 31 years
old (Herva et al., 2006). Participants in the Bogalusa Heart Study who were overweight or obese
as adolescents had greater cardiovascular risk factors as young adults, including adverse levels of
systolic and diastolic blood pressure, cholesterol, insulin, and blood glucose, compared to
participants who were normal weight as adolescents (Srinivasan, Bao, Wattigney, & Berenson,
1996). One study that examined the association between childhood obesity and cardiovascular
morbidity in a 57-year follow-up cohort study using the Carnegie Survey of Family Diet and
Health found greater risk for heart disease and cardiovascular mortality for those who were
overweight or obese as youth compared to those who were normal weight as youth (Gunnell,
Frankel, Nanchahal, Peters, & Smith, 1998). This association may be due to tracking of
overweight and obesity into adulthood (Gunnell et al., 1998). Clearly, it is necessary to
determine how to prevent and reduce excess adiposity during the critical developmental period
of adolescence.
6

These issues are even more crucial for minority populations. Risk for obesity-related
health conditions is higher in African American and Hispanic populations compared to other
ethnic groups. For example, compared to non-Hispanic white adults, Hispanic adults have a 66%
higher risk and African American adults have a 77% higher risk for being diagnosed with type 2
diabetes (Centers for Disease et al., 2011). African American adults also have higher rates of
cardiovascular disease than white adults (Folsom et al., 2011). Risk for diseases related to
overweight and obesity are also higher among African American and Hispanic adolescents
compared to non-Hispanic white adolescents (M. I. Goran, 2001; Linder & Imperatore, 2013).
Given that these populations are at higher risk for weight-related chronic illnesses, it is crucial to
investigate problem behaviors that lead to overweight and obesity among these groups to inform
targeted interventions.
The health consequences of overweight and obesity impart a high economic burden. The
rise in medical expenditures that have occurred over the last several decades in the U.S. can be
partly attributed to the rise in overweight and obesity. Between 1987 and 2001, 27% of the rise
in healthcare costs was due to obesity and related illnesses (Thorpe, Florence, Howard, & Joski,
2004). A recent report by the Centers for Disease Control and Prevention estimates that in a
single year, healthcare costs related to obesity were as much as $147 billion (Centers for Disease
& Prevention, 2010). Overweight and obesity have adverse economic consequences for
individuals as well. A systematic review of direct medical costs of overweight and obesity
estimated that in 2008, per-person medical costs were $266 for overweight and $1723 for obesity
(Tsai, Williamson, & Glick, 2011). Evidence indicates that obese individuals have higher
medical expenditures than normal weight individuals. One study found that, compared to normal
weight individuals, average annual medical expenditures are 36% higher for obese individuals
7

(Finkelstein et al., 2012). The economic cost of obesity and related diseases emphasizes the
burden of this health issue to public health as well as to individuals and highlights the need for
effective interventions to reduce and prevent obesity in current and future generations.
Etiology of obesity. The etiology of obesity is complex, as myriad distal and proximal
factors contribute to risk for weight gain (Rennie, Johnson, & Jebb, 2005). Many of these factors
interact to influence the most proximal and modifiable determinants of adiposity – physical
activity and dietary intake. The following provides a summary of key determinants of overweight
and obesity.
Physical environment factors. Environmental contexts influence overweight and obesity
risk because they help to shape several behaviors that are most proximally associated with
weight gain. There is an “obesogenic environment” in many residential neighborhoods in the
United States. An “obesogenic environment” can be characterized by food deserts (areas where
there is poor access to food outlets that provide fresh foods such as grocery stores), food swamps
(areas where there is an abundance of access to food outlets that provide pre-prepared foods such
as fast food restaurants), and poor access to safe places to exercise and play (Kirk, Penney, &
McHugh, 2010). These environmental characteristics promote energy imbalance and ultimately
weight gain by supporting excess energy intake and impeding energy expenditure through
physical activity (Hill & Melanson, 1999). Over time, increased energy intake and reduced
energy expenditure lead to a state of positive energy balance which causes increased adiposity
(Hill & Melanson, 1999). For example, it has been postulated that a small daily positive energy
balance caused by as little as 100 kcal/day of excess energy consumption may lead to
incremental weight gain (Hill, Peters, & Wyatt, 2009). This may be because higher body weight
leads to a higher energy requirement to maintain physiological homeostasis (Hill et al., 2009;
8

Swinburn et al., 2009). Greater energy intake is needed in order to meet a new, higher energy
requirement (Hill et al., 2009). The process of small increases in positive energy balance 
small amount of weight gain  increased energy requirements  compensatory increased
energy intake may lead to an incremental increase in weight gain over a period of years (Hill et
al., 2009). Consumption of diets that are high in sugar and low in fiber are likely to lead to
excess energy intake (Saris, 2003; Te Morenga, Mallard, & Mann, 2013). Consequently,
minority adolescents, whose diets tend to be higher in sugar and lower in fiber than adolescents
of other ethnicities, may be at a higher risk for this incremental weight gain. This may at least
partly explain the higher prevalence of overweight and obesity among minority adolescents.
The built environment, characterized by land use, availability of public transportation
and activity options, and factors related to urban design (Handy, Boarnet, Ewing, &
Killingsworth, 2002), has become a focus of research in recent years about the influence the
physical environment has on energy intake and energy expenditure. Low physical activity levels
have been found for residents of neighborhoods that lack access to recreational facilities and
green spaces, possibly due to a lack of opportunities for leisure time physical activity (Ellaway,
Macintyre, & Bonnefoy, 2005; Giles-Corti & Donovan, 2002; Hill & Melanson, 1999). Studies
have demonstrated that low walkability, often defined by few pedestrian access points, poor
street connectivity, high traffic levels, and lack of safe sidewalks, is associated with lower levels
of physical activity within residential areas (Giles-Corti & Donovan, 2002; Hill & Melanson,
1999; Saelens, Sallis, & Frank, 2003). Other neighborhood characteristics have also been
associated with obesity risk factors. Safety, for example, has been identified as a key trait of the
neighborhood environment that influences physical activity. Studies have found that residents in
neighborhoods with lower safety (either perceived or objective) engage in less physical activity,
9

such as walking for pleasure or exercise, than residents in neighborhoods with higher safety
(Parkes & Kearns, 2006; Saelens et al., 2003). These neighborhood characteristics are not only
linked with behavior risk factors for weight gain, they have been directly linked with overweight
and obesity; one study of physical activity using NHANES data found that people who live in
areas with higher walkability and lower crime have lower BMIs than people who live in areas
with lower walkability and higher crime (Doyle, Kelly-Schwartz, Schlossberg, & Stockard,
2006). Built environment and neighborhood characteristics may be particularly important
influences on weight-related behaviors in African American and Hispanic youth, as youth from
these ethnicities are more likely to live in neighborhoods with low access to exercise facilities
and high crime (Lovasi, Hutson, Guerra, & Neckerman, 2009).
The food environment may also influence obesity risk by hindering or facilitating access
to healthful food choices. In general, Americans have access to an overabundance of palatable,
energy dense but nutrient poor foods (i.e. “empty calorie” foods that are characterized by a high
caloric content and a low nutrient content; generally processed foods that are high in fats and
added sugars) (Booth et al., 2001). This accessibility can represent a challenge to one’s efforts to
maintain a healthful diet and reasonable portion sizes. The physical availability of food outlets in
the neighborhood environment may influence exposure to healthful and unhealthful foods.
Access to supermarkets has been found to be associated with greater intake of fruits and
vegetables and lower intake of energy-dense, nutrient poor foods, dietary patterns which lower
the risk for overweight and obesity (Morland, Wing, & Roux, 2002). Conversely, poor access to
supermarkets may be linked with lower fruit and vegetable intake and increased risk for weight
gain and excess adiposity (Rose & Richards, 2004). Proximity to fast food restaurants may also
be associated with higher obesity rates (Maddock, 2004). Low density of supermarkets and high
10

density of fast food restaurants and other convenience food outlets, which negatively impact
healthful nutrition, are common characteristics of low socioeconomic neighborhoods (Hinkle &
Wu, 2003; Morland, Wing, Roux, & Poole, 2002). The physical environment and food
environment represent a possible important influence on physical activity behavior, dietary
intake behavior, and obesity risk.
Social environment factors. Evidence indicates that social factors can contribute to
overweight and obesity risk. Social influences on adiposity can range from peer group norms to
more complex interactions between sociodemographic factors and the physical environment. The
influence of immediate social environment on overweight and obesity risk is exemplified by
findings that family and peer groups can influence weight-related perceptions and behaviors.
Family environment factors, such as home food availability (Rosenkranz & Dzewaltowski,
2008), parental feeding practices (Johannsen, Johannsen, & Specker, 2006), and family mealtime
practices (Videon & Manning, 2003), can influence risk for overweight and obesity. Evidence
indicates that individuals within peer groups change their behaviors based on exposure to
common influences (Cohen-Cole & Fletcher, 2008). Norms about body weight, physical activity,
and eating behaviors established within peer groups and behaviors of friends also affect an
individual’s weight-related behaviors (Ball & Crawford, 2006; Cohen-Cole & Fletcher, 2008;
Eisenberg, Neumark-Sztainer, Story, & Perry, 2005). For instance, determining how much one
should eat or drink may be influenced by norms regarding consumption and acceptable quantities
of food within social groups (Cohen-Cole & Fletcher, 2008; Eisenberg et al., 2005). The impact
of common social contexts on overweight and obesity risk is demonstrated by findings that body
weight may correlate among friends. In a recent study examining overweight in social networks
among 617 11-15 year old adolescents, those adolescents who were overweight were found to be
11

twice as likely to have overweight friends compared to normal weight adolescents (Valente,
Fujimoto, Chou, & Spruijt-Metz, 2009). Another recent study examined obesity rates in over
12,000 participants from the Framingham Heart Study over a span of 32 years (Christakis &
Fowler, 2007). The study found that in adult sibling pairs, if one sibling became obese over the
course of the study, the chance of the other sibling becoming obese increased by 40% (Christakis
& Fowler, 2007). Among spouses, if one partner became obese, the chance of the other partner
becoming obese increased by 37% (Christakis & Fowler, 2007). This evidence indicates that
obesity may “spread” throughout certain social networks over time (Christakis & Fowler, 2007).
Sociodemographic factors have long been recognized as one of the important upstream
determinants of overweight and obesity. For example, sociodemographic factors and social
position can impact overweight and obesity risk by determining access to opportunities for
healthful weight-related behaviors. One study found that poorer neighborhoods have around
three times fewer supermarkets compared to wealthier neighborhoods (although this study did
not control for size of neighborhood, which may impact access) (Morland, Wing, Roux, et al.,
2002). Consequently, people of lower socioeconomic status may be more affected by unhealthful
food environments than people of higher socioeconomic status, as they may have less access to
nearby outlets that sell healthful foods (Morland, Wing, Roux, et al., 2002). This may be one of
the reasons why there are higher rates of obesity among individuals of lower socioeconomic
status in developed countries (Ball & Crawford, 2006). It may also at least partly explain why a
greater proportion of African American and Hispanic individuals are overweight and obese in the
United States, as a higher proportion of racial/ethnic minorities in this country are from lower
socioeconomic backgrounds (Bennett, Wolin, & Duncan, 2008).
12

Disparities in overweight and obesity among these populations may also be due to
different sociocultural norms. Evidence indicates that African American and Hispanic
individuals may have lower rates of perceived overweight and obesity, as well as greater social
acceptance of overweight and higher body weight ideals, compared to white individuals, which
may influence weight-related behaviors among these groups (Bennett et al., 2008). From such
evidence, it is apparent that the social environment influences overweight and obesity risk in
both immediate, more proximal ways and in more systemic, distal ways. Overall, however,
individuals do not have much control over their physical and social environments, making these
factors difficult to modify on the part of the individual. Therefore, these factors may not
represent viable targets for interventions aimed and reducing and preventing overweight and
obesity.
Biological factors. Biological factors represent more proximal contributors to overweight
and obesity risk. Resting metabolic rate (the minimal rate of energy expended by the body per
unit time at rest), for example, can affect energy balance and propensity for weight gain (Hill &
Melanson, 1999). Physiological signals are also implicated in weight gain. Circulating levels of
hormones including leptin and ghrelin can influence energy intake through their effects on
hunger and satiety (Woods, 2004). Genetic factors may increase susceptibility for overweight
and obesity in some individuals. Estimates for the heritability of BMI have been found to range
anywhere from 30 to 70% (Herbert et al., 2006). Recent evidence suggests that gene-
environment-behavior interactions may play a role in determining adiposity (Bouchard, 2008; Qi
& Cho, 2008). The prenatal and perinatal time period may also be an important determinant for
the development of overweight and obesity (Rhee, Phelan, & McCaffery, 2012). Exposures that
occur in utero and shortly after birth have been found to play a significant role in predisposition
13

for weight gain later in life. For instance, gestational diabetes, excess gestational weight gain,
maternal smoking, and no or short breastfeeding have been associated with greater risk for
overweight and obesity during childhood and adulthood (Monasta et al., 2010; Rhee et al., 2012).
Recent research has focused on explaining how such environmental exposures during early life
influence later risk for overweight and obesity through epigenetic mechanisms, which alter gene
expression (Rhee et al., 2012). Future research is needed to understand how gene-environment
interactions impact overweight and obesity risk and development.
Behavioral factors. Lifestyle factors are perhaps the most modifiable proximal
determinants of overweight and obesity because these are the factors that can be most directly
controlled by the individual. It is well established that dietary intake, physical activity, and
sedentary behavior are particularly important determinants of adiposity, as these behaviors
directly impact energy balance. Weight gain occurs due to incremental positive energy balance
that accumulates over time (Rennie et al., 2005). Positive energy balance occurs when energy
intake exceeds energy expenditure (Jebb, 2007; Rennie et al., 2005). Past research has pointed to
dietary fat intake as the main dietary culprit driving excess energy intake in the United States
(Bray & Popkin, 1998). However, despite the growth in availability of low-fat foods, overweight
and obesity in this country persist, suggesting that fat consumption may not be the main dietary
culprit for the obesity epidemic (Ludwig, 2000; Saris, 2003; Volek, VanHeest, & Forsythe,
2005). This is supported by findings from studies that have demonstrated inconsistent
relationships between high-fat diets and weight gain (Kant, Graubard, Schatzkin, & Ballard-
Barbash, 1995; Larson et al., 1996), and studies that have demonstrated less weight loss on low-
fat diets compared to low-carbohydrate diets (Samaha et al., 2003). Such evidence suggests that
dietary fat may play a lesser role in creating positive energy balance compared to other dietary
14

habits. High consumption of dietary sugar and low consumption of dietary fiber are gaining
attention as strong contributing factors to overweight and obesity and to risk for obesity-related
chronic conditions, especially among children and adolescents.
Contributions of dietary sugar and fiber intake to overweight and obesity. Growing
evidence supports the links between high sugar consumption, low fiber consumption, and
increased body weight. A recent meta-analysis and systematic review that examined results from
60 randomized controlled trials and prospective cohort studies concluded that in adults and
youth, increased intake of added sugars (either ad-libitum or via intervention) was related with
increased body weight, while reduced dietary sugar intake was related with decreased body
weight (Te Morenga et al., 2013). Fiber intake, on the other hand, has been found to be
protective against overweight and obesity. Findings from epidemiologic studies indicate that
fiber intake is associated with lower body mass, and findings from intervention studies show that
fiber intake leads to weight loss (Kimm, 1995; Lissner, Lindroos, & Sjöström, 1998; Slavin,
2005). This evidence supports that sugar consumption increases risk for overweight and obesity,
while fiber consumption decreases risk for overweight and obesity.
The mechanism linking sugar intake and weight may simply be due to consumption of
sugar leading to increased energy intake beyond energy needs, since foods high in dietary sugar
are usually energy dense (Saris, 2003; Te Morenga et al., 2013). Other, more complex metabolic
consequences of dietary sugar intake may be responsible (Te Morenga et al., 2013). Foods high
in sugar typically generate a high glycemic response, which in turn stimulates carbohydrate
oxidation and hinders fat oxidation (the body’s process of burning stored carbohydrates and fats
as fuel), physiologic responses that promote increased fat storage and thus weight gain (Brand-
Miller, Holt, Pawlak, & McMillan, 2002; Saris, 2003). In contrast, high fiber foods, which are
15

typically characterized by low glycemic index, induce a lower, more stable glycemic response
(D. J. A. Jenkins & Jenkins, 1985). High fiber diets may also support weight loss and weight
maintenance by promoting satiety and decreasing food intake because they take increased time to
digest (Slavin, 2005). There is strong evidence that high-sugar diets are associated with
detrimental behavioral and physiological consequences. Consumption of sugar-laden foods may
lead to feelings of fatigue and low levels of physical activity in some individuals (Thayer &
McNally, 1992). Furthermore, diets high in added sugar (sugars or syrups that are added to foods
during processing or preparation) are associated with poor metabolic health (J. N. Davis et al.,
2007; Stanhope, Schwarz, & Havel, 2013) and increased risk for type 2 diabetes (Willett,
Manson, & Liu, 2002). In contrast, fiber intake is associated with health benefits, including
decreased risk for obesity and type 2 diabetes (Montonen, Knekt, Järvinen, Aromaa, &
Reunanen, 2003). Decreased risk for type 2 diabetes may be linked to fiber intake through the
effects that fiber intake has on improved glycemic control and insulin sensitivity (Anderson et
al., 2009; Anderson, Smith, & Gustafson, 1994). Fiber intake may lead to improved glycemic
control and insulin sensitivity by slowing absorption of nutrients in the gut, thereby reducing
blood glucose concentrations (D. J. A. Jenkins & Jenkins, 1985; Pereira et al., 2002; Potter,
Coffman, Reid, Krall, & Albrink, 1981). Lower blood glucose concentrations require a lower
insulin response for absorption of glucose by the tissues (Pereira et al., 2002). It is postulated that
over time, lower circulating insulin levels may cause an up-regulation of insulin receptors on cell
surfaces, leading to increased insulin sensitivity and improved glycemic control (Song,
Sawamura, Ikeda, Igawa, & Yamori, 2000).
Despite the adverse effects of sugar and the protective effects of fiber, the intake of these
dietary components by most individuals in the United States does not meet recommendations.
16

The average intake of added sugars in the United States has risen over the last several decades
due to increased availability of sugars in the food supply; between 1970 and 2005, availability of
dietary sugars in the U.S. increased by 19% (Johnson et al., 2009). Estimates put typical intake
of sugar at 22.2 teaspoons per day, equaling 355 calories (Johnson et al., 2009). This is well over
the recommendations by the American Heart Association of 6 teaspoons (100 calories) per day
for women and 9 teaspoons (150 calories) per day for men (Thompson & Veneman, 2005).
Typical fiber intake does not meet national recommendations, either. According to
recommendations by the American Heart Association, the Institute of Medicine, and the Panel on
Dietary Reference Intakes for Macronutrients, adults should consume between 25 and 38 grams
per day (King, Mainous, & Lambourne, 2012) and adolescents should consume between 26 and
38 grams per day of fiber (Slavin, 2005). However, recent estimates have found that the average
American adult consumes only 15.9 grams per day (King et al., 2012). Typical fiber
consumption is even lower among obese individuals at 14.6 grams per day, less than one-half of
the upper recommended daily intake (King et al., 2012). In a recent study of a nationally
representative sample of youth, more than 60% had average daily fiber consumption that was
less than the recommended amount (Brauchla, Juan, Story, & Kranz, 2012). It is clear that,
among the general population of the United States, sugar and fiber consumption does not
correspond to current dietary recommendations.
Consumption of sugar and dietary fiber may be particularly problematic among African
American and Hispanic youth, who are at higher risk for overweight, obesity, and diet-related
chronic diseases than youth from other ethnic backgrounds (Cruz et al., 2004; Dabelea et al.,
2007; Eaton et al., 2010; Michael I. Goran & Ventura, 2012; Hasson et al., 2012; Hasson et al.,
2013; Nesbitt et al., 2003; Toledo-Corral et al., 2011). While adolescents tend to have
17

unhealthful diets and consume high amounts of sugar across ethnicities (R. R. Briefel & C. L.
Johnson, 2004), the typical diets of African American and Hispanic adolescents are less healthful
than the typical diets of non-Hispanic white adolescents. Evidence indicates that Hispanic and
African American adolescents tend to consume diets that are higher in energy-dense,
micronutrient-poor foods (such as those high in added sugars) than adolescents of other ethnic
backgrounds (J. N. Davis et al., 2010; Mendoza, Drewnowski, Cheadle, & Christakis, 2006). In
fact, researchers from the USC Childhood Obesity Research Center have found that added sugar
intake is particularly high in Hispanic youth living in Los Angeles, averaging 100 grams per day
and accounting for 25% of total daily energy intake (J. N. Davis et al., 2010). Adolescents from
these populations also tend to consume diets that are lower in fiber-rich fruits and vegetable
(Trevino et al., 1999). Sugar and fiber consumption represent behaviors that are closely related to
overweight and obesity risk and require efforts to modify in these populations.
Sugar and fiber consumption: what is known about their effects. Maintaining limited
sugar consumption and high fiber consumption is difficult for many people, evidenced by the
low compliance with recommendations for these dietary components in the United States
(Johnson et al., 2009; King et al., 2012; Who & Consultation, 2003). It is especially crucial that
interventions are devised that support reduced sugar intake and increased fiber intake in African
American and Hispanic adolescents, given the high sugar intake and low fiber intake as well as
the increased risk for obesity and related chronic diseases in these populations. In order to
understand how to effectively alter sugar and fiber intake in these populations, it may be
important to first understand the effects of sugar consumption and fiber consumption on
physiology, psychology, and behavior. Such insights may aid in the design of targeted
interventions that can help initiate and maintain dietary change. Little research in humans has
18

focused on the effects of sugar and fiber intake on psychological, physiological, and behavioral
factors, and little is known about how these may interact to reinforce unhealthful dietary patterns
that are low in fiber and high in sugar. Research on this topic may reveal why it is it difficult to
improve intake of these dietary components.
Physiological effects of sugar and fiber consumption. The glycemic and insulinemic
responses to sugar and fiber intake have been a major focus of research on the physiological
effects of these dietary components. In normal weight, healthy individuals, subsequent to
consumption of high sugar foods, blood glucose rapidly rises, triggering a high insulin response
which leads to rapid glucose uptake by tissues and subsequent low circulating blood glucose
(Rennie et al., 2005; Roberts, 2000). In contrast, in normal weight, healthy individuals,
subsequent to consumption of high fiber foods, blood glucose rises slowly and elicits a lower
insulinemic response, so glucose uptake occurs less rapidly and circulating blood glucose levels
remain more stable (D. J. Jenkins et al., 1987). It has been postulated that the rapid decrease in
blood glucose in response to consumption of high sugar foods may hasten the return of hunger
after eating in some individuals (Brand-Miller et al., 2002; Rennie et al., 2005). Further, the
hyperglycemic and hyperinsulinemic response to high sugar food is exaggerated in individuals
who are overweight and/or insulin resistant (Brand-Miller et al., 2002; Rennie et al., 2005). The
more gradual glycemic response to high fiber and the bulk of fibrous foods have been
hypothesized to contribute to slower return of hunger after eating high fiber foods in some
individuals (Roberts, 2000). However, the relationship between the glycemic responses to high-
sugar and high-fiber foods are controversial due to mixed findings. Additionally, little is known
about how consumption of high sugar and high fiber foods affects hunger in overweight
individuals, particularly overweight adolescents.
19

Evidence from animal and human studies has shown that hunger is also influenced by a
number of other factors, including the food environment, mood, and signals from gut peptides.
For example, increased appetite and consumption of food has been shown to occur in response to
emotional distress in certain individuals, particularly those who are overweight (Greeno & Wing,
1994; Macht & Simons, 2000; Ouwens, van Strien, & van Leeuwe, 2009). The immediate food
environment, especially the palatability of available food, can also have a strong influence on
perceived hunger; evidence indicates that exposure to palatable foods may stimulate hedonic
hunger, which can be difficult for some individuals to distinguish from homeostatic hunger and
can lead to overeating (Low, Lacy, & Keast, 2014). In recent years, several gut peptides, such as
the hormones ghrelin and leptin, have been identified as mechanisms that modulate physiological
hunger and satiety. Much research on how these factors influence hunger has focused on their
isolated effects. Little is known about how food environment, physiological, and psychological
factors may interact to affect appetite in humans. Findings from animal studies point to the
potential for these factors to interact in affecting appetite. Research on rodents indicates that
stress can increase plasma concentrations of ghrelin, a gut peptide involved in the regulation of
physical hunger (Asakawa et al., 2001; Kristenssson et al., 2006; Schellekens, Finger, Dinan, &
Cryan, 2012). Ghrelin has been implicated in reward-based and stress-based eating behaviors;
studies have demonstrated that stress-induced increases in ghrelin levels stimulate behaviors that
counteract stress, such as consumption of hedonically pleasing foods (like those high in sugars),
in rats (Chuang & Zigman, 2010; Schellekens et al., 2012). However, further research is
necessary to elucidate the impacts of sugar and fiber on hunger, with the consideration of the
aforementioned psychological and physiological factors as mechanisms driving the relationships
between sugar, fiber, and hunger. Investigating how psychological and physiological factors
20

interact with the food environment to influence appetite may be an important avenue to examine,
particularly in overweight Hispanic and African American adolescents, who tend to consume
less healthful diets compared to youth from other ethnicities (Ronette R. Briefel & Clifford L.
Johnson, 2004; Choi, Meininger, & Roberts, 2006; Mendoza et al., 2006; Reynolds & Spruijt-
Metz, 2006).
Behavioral effects of sugar and fiber consumption. Neurochemical responses to sugar
intake have been a focus of much research on the potentially dependence-inducing properties of
dietary sugar. Despite the proliferation of knowledge about the negative impacts of high-sugar
foods, a low-sugar diet is difficult for many people to maintain, possibly because withdrawal
from palatable foods may increase the drive for consumption (Adam & Epel, 2007). Evidence,
mainly from animal studies, indicates that sugar may have a high hedonic (i.e. pleasant,
rewarding) value that reinforces sugar intake. Sugar-rich foods, which are highly palatable,
activate the reward system of the brain by stimulating opioid, dopamine, and endocannabinoid
signaling in the limbic system (Adam & Epel, 2007; Avena, Rada, & Hoebel, 2008). Repeated
activation of the brain’s reward system by consumption of sugar-laden, highly palatable foods
may cause neurobiological adaptations that lead to an increased drive to consume such foods
(Volkow & Wise, 2005). Consumption of high-sugar foods has been shown to occur at the
detriment of consumption of healthful foods, such as those high in fiber, in humans (Adam &
Epel, 2007). The potential for sugar consumption to drive dependence is also supported by what
happens in response to sugar restriction in animals. For example, one animal model of sugar
dependence has shown that rats exposed to a daily intermittent food deprivation condition show
preference for consuming sugar over standard chow when they are subsequently given food
access (Avena et al., 2008). Such reactions may also occur in humans restricted from sugar
21

intake. Evidence also indicates that individuals attempting to alter their diets by restricting
specific foods may relapse when those foods are available to them (Polivy, 1996; Wideman,
Nadzam, & Murphy, 2005). This may be because highly palatable foods can activate the brain’s
reward system, causing behavioral reinforcement for consumption of such foods (Adam & Epel,
2007). Consequently, restraint from eating sugar may exacerbate eating in response to the
presence of sugar-rich foods (Adam & Epel, 2007). It is not known, however, what effects sugar
consumption versus sugar restriction may have on subsequent sugar intake in overweight and
obese adolescents.
Psychological effects of sugar and fiber consumption. Sugar intake may impact mood
due to its hedonic properties. Studies in rats have demonstrated that stress promotes intake of
highly palatable foods, such as foods high in sugar (Dallman et al., 2003; Dallman, Pecoraro, &
la Fleur, 2005). Increased intake of sweet fatty foods has been observed in individuals who tend
to eat in response to emotion (i.e. emotional eating) (Benton & Nabb, 2003; Oliver, Wardle, &
Gibson, 2000). Evidence from animal studies indicates that sugary foods may have a calming
effect because consumption of simple carbohydrates increases synthesis of serotonin, a
neurotransmitter associated with feelings of happiness and well-being, in the brain (Benton,
2002; Wurtman & Wurtman, 1995). This indicates that sugar intake may alleviate negative
mood. However, other evidence points to an opposite influence of sugar intake on mood. Studies
have indicated that diets high in simple carbohydrates (such as foods with high amounts of added
sugars) may be associated with feelings of fatigue and low energy in some individuals (Thayer,
1987, 1989, 2003). If sugar intake does indeed have a positive impact on mood, given the
hedonic reward of sugar it is possible that reducing or restricting intake (for example, by
replacing with high fiber foods) may induce negative mood, especially in habitual sugar
22

consumers. However, little research has been done on the impact of fiber intake on mood.
Additionally, no studies to date have examined the effects of providing access to high-sugar
foods versus restricting access to high-sugar foods, especially in overweight and obese
adolescents who tend to have high habitual sugar consumption (Mendoza et al., 2006; Reynolds
& Spruijt-Metz, 2006).
Self-tracking as a potential intervention approach to support change in dietary
behaviors. Little is known about how to motivate adolescents to initiate and adhere to diets that
are low in sugar and high in fiber. Given the rising ubiquity and accuracy of tools such as
smartphone apps and on-body sensors that can aid in self-tracking of weight-related behaviors,
such as dietary intake (O’Reilly & Spruijt-Metz, 2013), self-tracking may represent a viable
approach to the challenge of helping adolescents make and maintain dietary changes. Self-
tracking has been shown to improve adherence to weight-related behavior change, including
dietary intake (Burke et al., 2005). Self-tracking of weight-related behaviors has also been shown
to lead to greater weight loss and success in weight loss maintenance (L. E. Burke, J. Wang, &
M. A. Sevick, 2011). However, evidence from interventions indicates that adherence to self-
tracking tends to be low and decline over time (Burke et al., 2012; Y. Mossavar-Rahmani et al.,
2004; Tate, Wing, & Winett, 2001; Yon, Johnson, Harvey-Berino, Gold, & Howard, 2007).
Since self-tracking of weight-related behaviors is such an important component of weight-related
behavior change, it is important to understand how to help individuals adhere to this strategy.
Little is known about how to improve adherence to long-term self-tracking. However,
there are communities of people who volitionally self-track a variety of behaviors and metrics
that can be consulted for insights into how to help people initiate and adhere to self-tracking.
Such communities include online fitness and lifestyle change communities and the Quantified
23

Self movement. Many individuals who participate in online fitness communities and in the
Quantified Self movement apply the data they collect from self-tracking to better understand
aspects of their personal health and wellness. Many are keen to share their data collection
techniques and findings. Their motivations, collective and individual, to monitor themselves
long-term have not yet been studied. The participants in online lifestyle change communities and
the Quantified Self movement represent untapped sources of information about volitional, long-
term self-tracking. Insights from these communities could be utilized in future intervention
research to improve adherence to dietary behavior change and enhance weight loss efforts.
While some research has focused on understanding barriers to self-tracking, no previous
research has focused on understanding why and how those who do engage in long-term self-
tracking are able to adhere to this strategy. Potentially modifiable characteristics such as
mindfulness, perceived competence, and motivation orientation may facilitate adherence, but
research is needed to understand how such characteristics may aid in self-tracking. Examining
what motivates individuals within online lifestyle change communities and the Quantified Self
community to volitionally self-track their health metrics may reveal common attributes among
them that could be targeted in other populations, such as minority adolescents with high-sugar,
low-fiber diets, to improve self-tracking adherence in future studies. To date, no formal research
has been done with these communities to understand 1) what attributes are common among those
who volitionally self-track weight-related behaviors, 2) what motivates those who self-track
weight-related behaviors to begin self-tracking, 3) what motivates them to adhere to self-
tracking, and 4) how self-tracking helps them attain personal health- and weight-related goals.

24

Gaps in the Literature
The interplay of psychological, behavioral, and physiological effects of sugar
consumption and fiber consumption has not been fully elucidated in humans. Adolescents,
particularly Hispanic and African American youth, tend to consume diets that are high in sugar
and low in fiber (Mendoza et al., 2006; Reynolds & Spruijt-Metz, 2006). In order to understand
how to help minority adolescents reduce high sugar intake and increase low fiber intake, it may
be important to elucidate how consumption of high sugar foods and consumption of high fiber
foods impact subsequent sugar intake, mood, and hunger. Additionally, it may be important to
understand individual characteristics that modulate how these factors change in response to sugar
exposure versus sugar restriction.
There is some evidence that restricting specific foods may cause people to relapse when
those foods are available to them (Polivy, 1996; Wideman et al., 2005). It is possible that
restricting sugar intake and replacing it with high fiber foods may instigate greater sugar intake
once sugar is again available. One possible mechanism for this could be increased negative mood
caused by sugar restriction. Based on mood maintenance hypothesis (Andrade, 2005; Clark &
Isen, 1982) and mood management theory (Andrade, 2005; Zillmann, 1988), restricting sugar
intake in the presence of high sugar food options may trigger some individuals to consume sugar
in an attempt to counteract negative mood precipitated by sugar restriction. Alternatively,
consumption of sugar subsequent to experiencing negative mood may occur due to depletion of
self-control. The self-control strength model posits that self-control is a limited cognitive
resource that depletes over time with consecutive attempts at exerting self-control (Muraven &
Baumeister, 2000). Any activity that demands self-control draws from the same self-control
resource (Muraven & Baumeister, 2000). So, for example, regulating mood, (which may lead to
25

depletion of one’s self-control resource because it requires overriding or inhibiting the current
mood) (Muraven & Baumeister, 2000) could potentially impact one’s ability to exert self-control
over eating behaviors. It is possible, then, that attempting to regulate negative mood in the face
of sugar restriction could lead to a depleted ability to limit intake of highly palatable, high-sugar
foods. To our knowledge, no studies to date have examined how restricting sugar intake in the
context of increasing fiber intake impacts subsequent sugar intake in overweight minority
adolescents, nor have studies examined possible mechanisms and individual characteristics that
may drive potential associations. Research is needed to demonstrate how these factors interrelate
to reveal why it is it difficult to improve dietary intake of sugar and fiber. Understanding these
factors may be important to help minority adolescents reduce sugar intake and increase fiber
intake.
Maintaining low-sugar, high-fiber diets over the long-term is difficult for many people to
achieve. One potential reason may be the impact of sugar intake on mood and the subsequent
impact of mood on hunger. Sugar consumption has been shown to have a mood-elevating effect
in some individuals (Benton & Nabb, 2003). Additionally, removal of palatable foods such as
those high in sugar from the diet can lead to negative mood in some individuals (Lowe & Butryn,
2007). Evidence indicates that mood can influence feelings of hunger; hunger has been shown to
be more intense during periods of negative mood in some individuals (Macht & Simons, 2000).
According to the psychosomatic hypothesis, negative mood can influence hunger in people who
have difficulty differentiating negative moods from physical signals of hunger and satiety
(Bruch, 1964; Greeno & Wing, 1994). Given this evidence, it is possible that attempting to
reduce sugar intake may lead to negative mood, which may cause increased hunger. To our
knowledge, no studies to date have examined how restricting sugar intake in the context of
26

increasing fiber intake impacts mood and hunger in overweight minority adolescents.
Understanding how restricting sugar intake influences mood and hunger may provide insight into
why it is difficult for many individuals to maintain low-sugar, high-fiber diets. This may be
particularly important to understand in overweight Hispanic and African American adolescents,
who tend to consume diets that are higher in sugar-rich foods and lower in fiber-rich foods
compared to youth from other ethnicities (Ronette R. Briefel & Clifford L. Johnson, 2004; Choi
et al., 2006; Mendoza et al., 2006; Reynolds & Spruijt-Metz, 2006).
Self-tracking, which consists of paying deliberate attention to an aspect of one’s behavior
and recording details about that behavior, is central to initiation and maintenance of weight-
related behavior change. It is known to help people adhere to dietary behavior changes (Burke et
al., 2005), and may be a promising intervention strategy to help adolescents adhere to diets that
are lower in sugar and higher in fiber. Self-tracking aids in weight-related behavior change by
promoting awareness of targeted behaviors and enhancing self-efficacy for behavior change
adherence (L. E. Burke et al., 2011; Y. Mossavar-Rahmani et al., 2004; Wadden, Butryn, &
Wilson, 2007; Wing, 1998). Enhanced awareness of targeted behaviors, such as eating behaviors
and foods consumed, may support goal-directed behavior by cuing memory retrieval of weight-
related behavior goals and intentions (Latner & Wilson, 2002).
Evidence from intervention studies indicates that adherence to self-tracking of weight-
related behaviors tends to be low and decline over time (Burke et al., 2012; Y. Mossavar-
Rahmani et al., 2004; Tate et al., 2001; Yon et al., 2007). Little is known about how to improve
adherence to long-term self-tracking of weight-related behaviors, especially dietary intake. Tools
such as smartphone apps and on-body sensors that can aid in self-tracking of health-related
behaviors are becoming increasingly ubiquitous and low cost (O’Reilly & Spruijt-Metz, 2013),
27

making self-tracking a promising approach to the challenge of aiding adolescents to make and
maintain dietary changes.
While a small yet growing body of research has focused on understanding barriers to self-
tracking, little is known about positive attributes or motivations related to self-tracking
adherence. People who participate in online fitness communities and the Quantified Self
movement who volitionally self-track health-related behaviors and metrics may represent
resources for understanding how to help people adhere to self-tracking as a strategy to improve
dietary behaviors. To our knowledge, no previous studies have utilized online fitness
communities or the Quantified Self community to explore why and how individuals who engage
in long-term self-tracking are able to adhere to this behavior change strategy.
Introduction to the Dissertation Studies
This dissertation addresses the identified knowledge gaps through three studies. The first
two studies use data from an in-lab experimental crossover design feeding trial in overweight
Hispanic and African American adolescents to investigate the interplay of psychological,
physiological, and behavioral effects of sugar and fiber consumption. The third study uses data
from a mixed methods qualitative/quantitative study of individuals who volitionally self-track
behaviors and metrics related to weight to understand factors that contribute to self-tracking
adherence. The overarching aim of these studies is to provide insights into how psychological,
physiological, and behavioral factors contribute to poor dietary patterns characterized by high
sugar intake and low fiber intake, and how self-tracking can be more efficaciously promoted to
aid dietary intake behavior change.

28

Specific aims and hypotheses
Study 1: Effects of high sugar/low fiber versus low sugar/ high fiber meal consumption
on subsequent ad libitum sugar intake in overweight minority adolescents. The overall
objective of study 1 is to examine how sugar restriction impacts subsequent sugar intake in
overweigh adolescents, and how individual-level characteristics modulate this relationship. To
investigate this topic, data is examined from a sample (N=87) of overweight minority
adolescents who participated in a randomized experimental crossover design feeding trial that
included an ad libitum snack period subsequent to condition-specific high sugar/low fiber and
low sugar/high fiber breakfast and lunch meals.
Aim 1: To examine how exposure to the high sugar/low fiber versus the low sugar/high fiber
meal conditions impacts ad libitum (i.e. voluntary, free-feeding) total sugar intake in overweight
African American and Hispanic adolescents.
Hypothesis 1: Participants will consume a greater amount of total sugar during the ad
libitum snack period in the low sugar/high fiber (sugar restriction) condition than in the
high sugar/low fiber (sugar exposure) condition.
Aim 2: To examine the mediating effect of self-reported negative mood measured during the ad
libitum snack period on the relationship between meal condition and ad libitum total sugar
intake.
Hypothesis 2: Higher self-reported negative mood ratings will mediate the relationship
between the low sugar/high fiber (sugar restriction) condition and higher ad libitum total
sugar intake.
Aim 3: To examine if the following individual characteristics moderate the effect of meal
condition on ad libitum sugar intake:
29

a. Habitual sugar intake
b. Impulsivity
Hypothesis 3a: Habitual total sugar intake will moderate the effect of meal condition on
ad libitum sugar intake. Compared to individuals with low habitual total sugar intake,
individuals with high habitual total sugar intake will consume more ad libitum total sugar
in the LSHF condition.
Hypothesis 3b: Impulsivity will moderate the effect of meal condition on ad libitum
sugar intake. Compared to individuals with low impulsivity, those with high impulsivity
will consume more ad libitum total sugar in the LSHF condition.
Study 2: Impact of negative mood on hunger ratings in response to high sugar/low
fiber versus low sugar/high fiber meals in overweight minority adolescents. The overall
objective of study 2 is to examine how restricting sugar intake in the context of increasing fiber
intake impacts mood and hunger in overweight minority adolescents. To investigate this topic,
data is examined from the same sample of overweight minority adolescents who participated the
randomized experimental crossover design feeding trial.
Aim 1: To examine the influence of negative mood ratings on initial hunger ratings and change
in hunger ratings throughout the day in HSLF (sugar exposure) vs. LSHF (sugar restriction) test
meal conditions in overweight minority adolescents, and whether the relationship between
negative mood and hunger is moderated by habitual total sugar intake.
Hypothesis 1a: During both the HSLF (sugar exposure) and LSHF (sugar restriction) test
meal conditions, higher initial negative mood ratings will be associated with higher
fasting hunger ratings, and this relationship will be moderated by habitual sugar intake
(i.e. there will be a significant positive interaction between initial negative mood and
30

habitual total sugar intake on fasting hunger ratings). This is hypothesized to occur in
both the HSLF and LSHF conditions because initial negative mood and fasting hunger
ratings were assessed at time 0, prior to exposure to the test meal conditions.
Hypothesis 1b: Higher negative mood will be associated with an increase over time in
hunger ratings in the LSHF (sugar restriction) condition but not in the HSLF condition.
This relationship will be moderated by habitual sugar intake, such that in the LSHF
condition, higher negative mood ratings will be associated with an increase over time in
hunger ratings for participants with higher habitual sugar intake.
Study 3: Self-monitoring of weight-related behaviors: a promising intervention
approach for sustaining dietary behavior change. The overall objective of study 3 is to identify
factors that aid establishment and maintenance of self-tracking adherence. This study uses
qualitative and quantitative data from an exploratory mixed methods study of people who
volitionally self-track weight related behaviors.
Aim 1: To examine attributes that may be common among individuals who self-track weight-
related behaviors by comparing them to individuals who do not self-track weight-related
behaviors on the following domains: health locus of control, dispositional mindfulness levels,
motivation orientation for healthful diet and physical activity behaviors, and competence for
making changes toward healthful behaviors. This aim was achieved through questionnaires.
Hypothesis 1a: Compared to people who do not self-track weight-related behaviors,
individuals who do self-track weight-related behaviors will have a greater tendency
toward an internal health locus of control.
Hypothesis 1b: Compared to people who do not self-track, people who do will have
higher dispositional mindfulness.
31

Hypothesis 1c: Compared to people who do not self-track, people who do will have a
greater orientation toward autonomous motivation for engaging in healthful dietary and
physical activity behaviors.
Hypothesis 1d: Compared to people who do not self-track, people who do will have
higher perceived competence for making changes toward healthful behaviors.
Aim2: To investigate factors that may contribute to initiation and adherence to self-tracking of
weight-related behaviors, as well as how self-tracking aids health-related goal attainment. This is
achieved through analysis of in-depth qualitative interviews.
Hypothesis 2: The in-depth interviews functioned to elicit descriptive information about
what motivates individuals to begin to self-track weight-related behaviors and factors that
help them to adhere to self-tracking.

Theoretical model for the dissertation. Figure 1-3 shows the theoretical model for this
dissertation. Study 1 is based on prior animal and human studies as well as the self-control
strength model (Muraven & Baumeister, 2000) and affect regulation theories including and
mood maintenance hypothesis (Andrade, 2005; Clark & Isen, 1982) and mood management
theory (Andrade, 2005; Zillmann, 1988). The portion of the theoretical model for this study
provides a framework for testing: a) the relationships between meal condition and subsequent ad-
libitum sugar intake, b) the mediating role of negative mood on the relationship between meal
condition and ad libitum sugar intake, and c) the moderating roles of habitual sugar intake and
impulsivity on the relationships between meal condition and ad libitum sugar intake.
Study 2 is based on findings from prior animal and human studies as well as the
psychosomatic hypothesis (Bruch, 1964; Greeno & Wing, 1994). The portion of the theoretical
32

model for this study provides a framework for testing: a) the relationship between negative mood
ratings and hunger ratings in response to high sugar/low fiber vs. low sugar/high fiber
experimental meal conditions and b) the moderating role of habitual sugar intake on the
relationships between meal condition, negative mood, and hunger ratings.
The framework for study 3 is partially based on Self-Determination Theory (SDT), a
theory of motivation that is concerned with explaining intrinsic tendencies to behave in effective
and healthful ways (Deci & Ryan, 2011). SDT provides a context for examining potentially
modifiable attributes that that may be common among individuals who volitionally self-track.
Study 3 provides a hypothesis-generating framework for investigating factors that influence
motivation for and adherence to self-tracking.
Figure 1-3. Theoretical model of the dissertation research

33

Data source and study samples. Data for study 1 and study 2 are from the FAME
(Food, Adolescence, Mood, and Exercise) study. This randomized experimental cross-over
design feeding study collected data on the acute effects of meals that differed in sugar and fiber
content on physiology, mood, and behavior in overweight and obese African American and
Hispanic adolescents. The study applied two experimental meal conditions, high sugar/low fiber
versus low sugar/high fiber, in separate 8-hour in-lab observation settings. The two experimental
meal conditions were conducted during two in-lab visits separated by a 2 to 4 week washout
period. Participants underwent both conditions and were randomly assigned to a high sugar/low
fiber or a low sugar/high fiber visit order. The study conditions were identical except for the
experimental meal conditions.
Data for study 3 is from an exploratory mixed methods qualitative/quantitative study of
individuals who volitionally self-track weight related behaviors and metrics, including dietary
intake, exercise, sleep, stress, and weight. Participants were individuals who participate in online
fitness communities or the Quantified Self community (self-tracking group), as well as
individuals who do not self-track any weight-related behaviors (comparison group). This study
employed in-depth interviews to explore topics related to self-tracking adherence in the self-
tracking group, as well as questionnaires to compare key attributes hypothesized to be related to
self-tracking adherence between the self-tracking group and the comparison group.

34

Chapter 2: Effects of high sugar/low fiber versus low sugar/high fiber meal consumption on
subsequent ad libitum sugar intake in overweight minority adolescents

Introduction
When consumed, sugar has hedonic properties that may reinforce further sugar
consumption. Evidence indicates that the hedonic properties of sugar are due to the impact of
sugar consumption on the neural reward system. Studies in humans have shown that habitual
consumption of highly palatable foods, such as those high in sugar, leads to a blunted reward
response by reducing availability of dopamine D2 receptors and dopamine release (C. Davis &
Carter, 2009; Parylak, Koob, & Zorrilla, 2011; Volkow, Wang, Fowler, & Telang, 2008). This
downregulation of the reward response suggests that habitual sugar consumers may need
increasingly greater amounts of sugar in their diets to derive a pleasurable response from sugar
consumption. Studies in animals have shown that repeated activation of the reward system by
habitual exposure to sugar can cause neurobiological adaptations that increase drive for
consumption of sugar-rich foods (Adam & Epel, 2007; Volkow & Wise, 2005). Evidence from
rodent studies has demonstrated that consumption of sugar stimulates opioid, dopamine, and
endocannabinoid signaling in the brain’s reward pathway (Adam & Epel, 2007; Avena et al.,
2008). This activation of the reward system leads to gratifying or pleasurable sensations in the
short term and can lead to behavioral reinforcement of sugar intake over the long term (Berridge,
2003). Such neurobiological evidence from human and animal research demonstrates that sugar
intake has a complex, reward-inducing impact that may perpetuate high sugar consumption.
The hedonic value of sugar is exemplified by findings in experimental animal studies
examining the impact of unpleasant feelings and experiences on sugar consumption behaviors.
When rodents are exposed to acute stressors and have a choice of consuming sugar-rich food,
35

they consume more of the sugar-rich food compared to other choices (Adam & Epel, 2007;
Dallman et al., 2003; Dallman et al., 2005). In one study that examined the impacts of food
restriction and stress on food intake in rats, animals that were exposed to repetitive bouts of food
restriction and stress increased consumption of sugar-rich cookies over less palatable chow
(Boggiano et al., 2005). These findings suggest that rats will eat sugar to reduce stress arousal,
and lend support to the idea that sugar consumption leads to hedonic reward (Adam & Epel,
2007; Boggiano et al., 2005). Rodent studies have also found that foods rich in simple
carbohydrates, such as those high in sugar, induce a calming effect by stimulating the synthesis
of serotonin (Polivy, 1996; Wideman et al., 2005), a neurotransmitter associated with positive
mood (E. Williams et al., 2006). This further supports that sugar consumption may be linked to
positive mood responses, which could in turn promote subsequent sugar intake.
Given the high hedonic value of sugar and its effects on neural reward circuitry, it is
possible that restricting sugar may lead to a rebound effect of increased sugar intake. Indeed,
there is evidence that abstinence from palatable foods can result in eating binges of the restricted
food when it subsequently becomes available (Polivy, 1996). For example, in an experiment with
3 to 5 year old children, Fisher and Birch showed that restricting access to children’s preferred
snacks increased their selection and intake of those foods in an ad libitum setting (Jennifer Orlet
Fisher & Leann Lipps Birch, 1999). If sugar restriction does indeed lead to a sugar intake
rebound, it is possible that this could occur through an attempt to ameliorate negative mood
induced by sugar restriction. Studies have shown that removal of sugar from the diets of rats that
are given intermittent access to sugar-rich foods and solutions leads to behaviors that
demonstrate increased anxiety, accompanied by increases in expression of the stress-related
biomarker corticotropin-releasing factor (CRF) (Cottone, Sabino, Roberto, et al., 2009; Cottone,
36

Sabino, Steardo, & Zorrilla, 2009; Parylak et al., 2011). Anecdotal accounts in humans indicate
that restriction of sugar-rich foods can lead to dysphoria (Ifland et al., 2009) as well as symptoms
of withdrawal and cravings (Avena et al., 2008), which can be associated with negative mood
(A. J. Hill, C. F. L. Weaver, & J. E. Blundell, 1991; Lafay et al., 2001). According to affect
regulation theories, including mood management theory (Andrade, 2005; Zillmann, 1988) and
the mood-maintenance hypothesis more specifically (Andrade, 2005; Clark & Isen, 1982), some
people consume palatable food to relieve negative mood (Garg, Wansink, & Inman, 2007).
Removing sugar from the diet, then, may trigger some individuals to consume sugar at a future
time in an attempt to counteract negative mood precipitated by sugar restriction.
Restriction of sugar intake could also lead to increased subsequent sugar intake through
demands on self-control. Self-control is a cognitive resource that one exerts when one attempts
to regulate how one thinks, feels, or behaves (Muraven & Baumeister, 2000). According to the
self-control strength model, self-control is limited and diminishes with repeated exertions
(Muraven & Baumeister, 2000). Consequently, people are more likely to fail at exerting self-
control after recent demands on self-control resources (Muraven & Baumeister, 2000). If sugar
restriction leads to negative mood, it is possible that attempting to regulate negative mood could
cause a diminished capacity to exert self-control in later choosing healthful, low-sugar foods
over less healthful, high-sugar foods. Indeed, this is supported by findings from previous
research that has shown self-control depletion to be associated with increased sensitivity to
immediately rewarding cues, such as palatable foods (Salmon, Fennis, de Ridder, Adriaanse, &
de Vet, 2014). Reactance to sugar restriction could be particularly salient in people who have
high impulsivity. Individuals with higher impulsivity have been found to be less able to inhibit
automatic responses to positively reinforcing stimuli, such as highly palatable foods (R Guerrieri,
37

Nederkoorn, & Jansen, 2008; Logan, Schachar, & Tannock, 1997). Greater impulsivity has also
been linked to unhealthy eating habits. For example, experimental studies have demonstrated an
association between higher impulsivity and overeating in healthy women (Ramona Guerrieri,
Nederkoorn, & Jansen, 2007) and children (R Guerrieri et al., 2008). A particular aspect of
impulsivity, negative urgency (the tendency to act rashly in response to negative feelings), has
been related to increased risk for addictive eating (Murphy, Stojek, & MacKillop, 2014). It is
possible that taxing self-control with attempts to regulate negative mood may lead people with
higher impulsivity to consume high amounts of sugar after experiencing sugar restriction.
The hedonic value of sugar consumption and consequences of sugar restriction may make
positive dietary behavior change, such as reducing sugar intake and increasing fiber intake,
difficult to initiate and maintain. Given that removing palatable foods from the diet can result in
cravings, negative mood, and subsequent eating binges of the restricted food (Avena et al., 2008;
Jennifer Orlet Fisher & Leann Lipps Birch, 1999; Andrew J. Hill et al., 1991; Lafay et al., 2001;
Polivy, 1996), removing high-sugar foods from the diet and replacing them with high-fiber foods
may induce further sugar intake in some individuals. This may be particularly relevant for
individuals with high habitual sugar intake and individuals who have greater impulsive
tendencies. Experimental research has examined the impact of sugar restriction on rodents
(Boggiano et al., 2005), and research in humans has examined the impact of restricting specific
food types in young children (for example preferred snack foods) (Jennifer Orlet Fisher & Leann
Lipps Birch, 1999). However, to my knowledge no studies to date have examined how restricting
sugar intake in the context of increasing fiber intake impacts subsequent sugar intake in humans.
Examining the impact of restricting sugar in the context of increasing fiber on mood and
subsequent ad libitum (voluntary, free-feeding) sugar intake may provide important insights for
38

future dietary interventions. For example, findings could reveal whether interventions aimed at
reducing sugar intake and increasing fiber intake could improve adherence to dietary behavior
change by including intervention components targeting negative mood and impulsive tendencies.
Such insights may be especially important for interventions targeting dietary behavior change in
Hispanic and African American adolescents, who tend to have diets that are higher in sugar and
lower in fiber than adolescents of other ethnicities (Jaimie N. Davis et al., 2010; Mendoza et al.,
2006).
The objective of this study was to use data from a randomized experimental crossover
design feeding trial (the Food, Adolescence, Mood and Exercise (FAME) trial, described in
detail elsewhere (O'Reilly et al., 2015) to examine the impact of low sugar/high fiber meals
compared to high sugar/low fiber meals on mood and subsequent ad libitum sugar intake in
overweight minority adolescents. Figure 2-1 shows that the proposed purpose of these analyses
was to: 1) investigate the acute impact of high sugar/low fiber (HSLF) versus low sugar/high
fiber (LSHF) meals on subsequent ad libitum sugar intake, 2) examine the potential mediating
effect of negative mood on the relationship between meal condition and ad libitum sugar intake,
and 3) investigate whether the mediating effect of negative mood depends on level of habitual
sugar intake and impulsivity.

39

Figure 2-1. Study 1 conceptual model

Planned Specific Aims and Hypotheses
The planned specific aims of this study were:
Aim 1: To examine how exposure to the high sugar/low fiber versus the low sugar/high fiber
meal conditions impacts ad libitum (i.e. voluntary, free-feeding) total sugar intake in overweight
African American and Hispanic adolescents.
Hypothesis 1: Participants will consume a greater amount of total sugar during the ad
libitum snack period in the low sugar/high fiber (sugar restriction) condition than in the
high sugar/low fiber (sugar exposure) condition.
Aim 2: To examine the mediating effect of self-reported negative mood measured during the ad
libitum snack period on the relationship between meal condition and ad libitum total sugar
intake.

40

Hypothesis 2: Higher self-reported negative mood ratings will mediate the relationship
between the low sugar/high fiber (sugar restriction) condition and higher ad libitum total
sugar intake.
Aim 3: To examine if the following individual characteristics moderate the mediating effect of
negative mood on the relationship between meal condition and ad libitum sugar intake:
c. Habitual sugar intake
d. Impulsivity
Hypothesis 3a: Habitual sugar intake will moderate the mediating effect of negative
mood on the relationship between meal condition and ad libitum sugar intake such that
the effect of meal condition on negative mood will depend on habitual sugar intake.
Compared to participants with low habitual sugar intake, participants with high habitual
sugar intake will report greater negative mood immediately prior to the ad libitum snack
period in the low sugar/high fiber meal condition and will then consume a greater amount
of ad libitum total sugar during the snack period.
Hypothesis 3b: Impulsivity will moderate the mediating effect of negative mood on the
relationship between meal condition and ad libitum sugar intake such that the effect of
negative mood on ad libitum sugar intake will depend on impulsivity. Compared to
participants with lower impulsivity, participants with higher impulsivity will consume a
greater amount of ad libitum total sugar in response to negative mood during the low
sugar/high fiber meal condition.

41

Methods
Study Sample
The study sample consisted of participants from the Food Adolescence Mood and
Exercise (FAME) study, a randomized experimental crossover design feeding study conducted at
the University of Southern California (O'Reilly et al., 2015). This study investigated the acute
effects of meals that differed in carbohydrate quality (sugar and fiber content) on physiology,
mood, and behavior in overweight minority adolescents. Participants were 87 overweight and
obese African American and Hispanic adolescents who were recruited from clinics, hospitals,
churches, schools, health fairs, and other gatherings in the Los Angeles area from 2007–2010 by
an ethnically diverse staff. To be included in the study, participants were required to meet the
following inclusion criteria: a) African American or Hispanic/Latino ethnicity (all four
grandparents reported to be African American or Hispanic/Latino), b) male or female between 14
and 17 years old, and c) body mass index BMI ≥ 85
th
percentile for age and sex. Potential
participants were excluded if they: a) had a diagnosis of diabetes, b) were participating in a
weight loss or exercise program, c) used medications that influenced body weight or insulin
sensitivity, or d) had a diagnosis of a syndrome that influences body composition. Prior to study
procedures, informed written parental consent and participant assent were obtained. The study
was approved by the Institutional Review Board of the University of Southern California. Data
were collected between 2008 and 2010. Sample characteristics are presented in Table 2-1.

42

Table 2-1. FAME study sample characteristics
Variable Mean (SD) [range]
Age (years) 16.3 (1.2) [14.2 – 18.5]
BMI z-score 2.02 (0.5) [0.5 – 3.2]
Weight status
1

Overweight 5.7% (5)
Obese 94.3% (83)
Insulin Sensitivity (SI) 1.6 (1.1) [0.1 – 8.2]
Sex
2

Male 48.9% (43)
Ethnicity
2

African American 43.2% (38)
Latino 56.8% (50)

Experimental Design
This randomized experimental crossover design feeding study consisted of two controlled
experimental meal conditions, a high sugar/low fiber (HSLF, i.e. sugar exposure) meal condition
and a low sugar/high fiber meal condition (LSHF, i.e. sugar restriction), during which
participants ate condition-specific breakfast and lunch meals. The meal conditions were
developed using data from focus groups conducted at the University of Southern California with
the purpose of identifying foods typically eaten by the study population. Focus group participants
were 12 African American (5 male, 7 female) and 19 Hispanic/Latino (9 male, 10 female)
adolescents. The contents of the HSLF and LSHF meals were determined based on focus group
participant feedback. Foods included in each meal condition are listed in Table 2-2. The meal
Mean and (SD) reported, unless otherwise indicated
1
: Overweight defined as BMI percentile ≥85 and <95;
obese defined as BMI percentile ≥95
2
: Frequencies (N)
SD = standard deviation; BMI = body mass index

43

conditions were isocaloric and matched for macronutrients except for the experimental
components (sugar and fiber content), and were determined using the Nutrient Data System for
Research (NDS-R 2010, University of Minnesota, Minneapolis, MN). Portions for the test meals
were determined for each participant based on 20% of the participant’s daily caloric needs,
which were calculated using sex, age, height, and body weight using the Dietary Reference
Intake Guidelines Estimation of Energy Expenditure for overweight children ages 3-18 (Hellwig,
Otten, & Meyers, 2006). An unblinded Registered Dietician designed and prepared or supervised
the preparation of all test meals. Meal macronutrients are presented in Table 2-3. Each
participant underwent both meal conditions in separate 8-hour in-lab observation visits that were
conducted 2 to 4 weeks apart. Participants were randomly assigned to a HSLF/LSHF or a
LSHF/HSLF visit order using a stratified block design randomization procedure. The study visits
were identical except for the experimental meal conditions.

Table 2-2. HSLF and LSHF test meal contents
HSLF meals LSHF meals

Regular Pop-tart
(Kellogg NA Co., Battle Creek, MI)

Thompson’s 100% Whole-wheat bagel
(Bimbo Bakeries, Horsham, PA)
Calcium-enriched string cheese
(Sargento Mootown Light String Cheese, Sargento Food Inc.,
Plymouth, WI)
Margarine
(I Can’t Believe It’s Not Butter Light, Unilever PLC/Unilever N.V.,
Englewood Cliffs, NJ)

Tampico juice
(Tampico Beverages, Chicago, IL)
Water treated with soluble fiber powder
(Benefiber Powder, Novartis Consumer Health Inc., Parsippany, NJ)

44

Table 2-3. HSLF and LSHF test meal nutrient compositions

Ad libitum snack period. The ad libitum snack period occurred subsequent to the lunch
meal during both meal conditions. Participants were given access to a snack tray that consisted of
a variety snack foods based on findings from the aforementioned focus groups. The snack food
choices and portions were identical during each meal condition and contained low sugar, high
sugar, low fiber, and high fiber foods (see Appendix A for a list of snacks provided).
Participants were given free access to the snack tray for 3 hours (from 1 hour after lunch until the
end of the 8 hour observation period) and were invited to eat any amount of any of the foods on
the snack tray.
In-lab procedures. Study visits were conducted at the Spruijt-Metz observation
laboratory at the University of Southern California’s Health Science Campus. Participants began
each study visit after a 10-hour overnight fast. Breakfast and lunch meal times were standardized
to 15 minutes during each visit, and participants were instructed to eat the entire meal for both
breakfast and lunch. After finishing the breakfast meal, participants were instructed to choose
from activities available in the laboratory for an 8-hour observation period. Activity options
included a variety of physically active and sedentary activity options. For 8 hours after the start
Macronutrient
grams (% kilocalorie)
HSLF Meal LSHF Meal
54.0 grams Poptart 61.0 grams whole wheat bagel
42.0 grams string cheese 14.0 grams margarine
247.0 grams juice 10.5 grams Benefiber supplement
Fat 11.0 grams (24%) 9.5 grams (24%)
Carbohydrate 64.0 grams (61%) 61.0 grams (68%)
Protein 14.0 grams (13%) 10.0 grams (11%)
Sugar 41.0 grams (39%) 7.0 grams (8%)
Fiber 1.0 grams (1%) 16.0 grams (18%)
45

of the breakfast meal, participants completed self-report questionnaires that assessed appetite and
mood.
Measures
Outcome
Ad libitum total sugar consumption. The outcome, ad libitum total sugar
consumption, was modeled as a continuous variable. Each participant had a measure of ad
libitum sugar consumption from both meal conditions, for a total of two outcome measures. Ad
libitum sugar consumption was assessed at the end of the observation period for the visit by
comparing the quantities of snack foods provided to the quantities of snack foods consumed. The
weight (grams) of each food and the volume (milliliters) of each drink on the ad libitum snack
tray were measured prior to the ad libitum snack period to establish the amount of each type of
food and drink provided to the participant. At the end of the observation period, a Registered
Dietician recorded the post-snack weights and volumes of each food and drink. This information
was compared to the pre-snack weights and volumes to assess the types and amounts of each
snack consumed by the participant. This data was then entered into the Nutrient Data System for
Research (NDSR) (http://www.ncc.umn.edu/products/ndsr.html) in order to calculate the amount
of total sugar in grams consumed by the participant during the ad libitum snack period. The ad
libitum sugar variable was adjusted for ad libitum energy intake by dividing the number of grams
of total ad libitum sugar consumed by 1000 kilocalories. This calculation provided a measure of
nutrient density and allowed for adjustment of possible measurement error.(National Cancer
Institute)

46

Mediator
Negative mood. The construct hypothesized to mediate the relationship between meal
condition and ad libitum total sugar intake was negative mood measured during the snack period.
Negative mood was measured via a four-item visual analogue scale (VAS) adapted from the
Profiles of Mood States (POMS) (Terry, Lane, Lane, & Keohane, 1999). The scale was
composed of one item for each of the following feelings: nervous, worried, anxious, and panicky.
Participants were asked to rate how nervous, worried, anxious, and panicky they felt at the time
of measurement by indicating with a mark on the corresponding 100 mm line that had endpoints
labeled “not at all” (corresponding to 0) and “extremely” (corresponding to 100). Each item
yielded a single score between 0 and 100 (corresponding to the number of millimeters from 0, or
“not at all”, to where the participant placed the mark on each line). The VAS for negative mood
is depicted in Figure 2-2. Participants self-reported their levels of nervousness, worry,
anxiousness, and panic every 30 minutes from the start to the end of the visit. A composite
negative mood score was obtained by averaging the nervous, worry, anxious, and panic scores at
each time point. A total of eighteen negative mood measures were conducted over each in-lab
visit. The average of the six negative mood measures collected during the three-hour ad libitum
snack period was used for the analyses. The composite negative mood measure had good internal
consistency (LSHF Cronbach’s α = 0.93, HSLF Cronbach’s α = 0.84) and good test-retest
reliability (Pearson’s r = 0.80, p<0.01).

47

Figure 2-2. Visual analogue scale (VAS) for negative mood

Moderators
Hypothesized moderators of the indirect effect of negative mood on the relationship
between meal conditions on ad libitum total sugar intake were habitual sugar intake and
impulsivity. Descriptions of the moderator measures are provided blow. Habitual sugar intake
was originally hypothesized to moderate the relationship between meal condition and negative
mood, and impulsivity was originally hypothesized to moderate the relationship between
negative mood and ad libitum total sugar intake. Subsequent to the original planned analyses,
follow-up analyses were conducted to determine whether habitual total sugar intake and
impulsivity moderate the effect of meal condition on ad libitum sugar intake. In these follow-up
analyses, higher habitual sugar and higher impulsivity were hypothesized to moderate the
relationship between the HSLF meal condition and ad libitum total sugar intake.
Habitual total sugar intake. Habitual total sugar intake was measured using a series of
three multiple pass 24 hour dietary recalls one week prior to the first in-lab visit. The dietary
recalls were conducted via phone call by trained interviewers using the Nutrient Data System for
Research (NDS-R 2010, University of Minnesota, Minneapolis, MN). The number of grams of
48

dietary sugar consumed during the 24 hours prior to the dietary recall was calculated using the
NDSR system. Habitual sugar intake was calculated by averaging the reported number of grams
of total sugar intake from each dietary recall session. Habitual total sugar intake was used as a
continuous variable in the analyses.
Impulsivity. Impulsivity was calculated using a subset of items from the Urgency
Subscale of the UPPS Impulsive Behavior Scale measured at the baseline visit (Whiteside &
Lynam, 2001). The Urgency Subscale of the UPPS Impulsive Behavior Scale is a twelve-item
scale measuring negative urgency, the tendency to commit rash or regrettable actions in response
to negative mood (Whiteside & Lynam, 2001). The scale includes items related to the inability to
resist cravings and binging, when experiencing low mood (Whiteside & Lynam, 2001).
Responses are rated on a likert-type scale from 1 (not at all) to 4 (a lot). To score the responses,
the item responses were averaged to generate a score ranging between 1 and 4, with a high score
representing a tendency to engage in impulsive behaviors (such as reacting to cravings) in order
to alleviate negative mood (Whiteside & Lynam, 2001). A composite score was generated by
averaging the responses to each of the 12 items in the scale, and had good internal consistency
(Cronbach’s α = 0.88). Impulsivity was used as a continuous variable in the analyses.
Participant demographics. Prior to the two in-lab feeding visits, participants completed
an inpatient visit at the Clinical Trials Unit at the USC University Hospital where body weight,
height, and demographic data were collected. Age (years) and sex were self-reported by
participants and collected from the protocol flow sheets at the in-patient Clinical Trials Unit
visit. Height (cm) and weight (kg) were measured in triplicate by a Registered Nurse. Body mass
index (BMI) z-score was calculated using the height and weight measurements based on the
CDC age- and sex- specific growth charts (Kuczmarski et al., 2002).
49

Statistical Analyses
Data analyses were conducted using SAS version 9.4 (Cary, NC) and Mplus version 6
(Muthén, L. K., & Muthén, B. O. (1998-2011)). Significance for all statistical tests were set at p
< 0.05. Prior to the analyses, the distribution of the outcome variable, ad libitum total sugar
consumption, was checked for normality and skew, and a square-root transformation was applied
to improve normality.
To test hypothesis 1, repeated measures ANCOVA via SAS PROC GLM was be used to
examine the within-subject main effect of meal condition on ad libitum sugar intake. This
analysis examined whether there were differences in mean ad libitum total sugar consumption
during the LSHF meal condition compared to the HSLF meal condition, controlling for the a
priori covariates sex, ethnicity, age, body mass index (BMI)-z score, and randomization order.
The within-subjects factor was meal condition. The analytic sample for the repeated measures
ANCOVA test of hypothesis 1 included 84 participants. A total of 3 participants were excluded
from the analyses for not having ad libitum intake data for both meal conditions. This statistical
method was used to examine aim 1 because the outcome variable, ad libitum total sugar intake,
was measured twice for each participant (once during each meal condition); thus, the outcome
variables were not statistically independent between meal conditions. Repeated measures
ANCOVA is appropriate for a model with a continuous dependent variable, a categorical
independent variable, and one or more covariates, where all cases have been tested under every
condition (i.e. the outcome variables are not independent). This method accounts for the lack of
statistical independence of the outcome variables between meal conditions. Additionally, the
data was appropriate for this type of analysis because the study design is balanced since all of the
participants completed both meal conditions.
50

To test hypothesis 2, mediation analysis was conducted to examine if greater self-
reported negative mood measured during the ad libitum snack period mediated the relationship
between the meal condition and greater ad libitum total sugar consumption. A priori covariates
included sex, ethnicity, age, body mass index (BMI)-z score, and randomization order. The
analytic sample for the mediation analysis to test hypothesis 2 included 82 participants. In
addition to the 3 participants that were excluded for missing ad libitum total sugar data for both
meal conditions, 2 participants were excluded from the analysis for missing snack period
negative mood data. The mediation analysis was conducted in Mplus in a path analytic
framework approach for examining mediation in two-condition, within-participant repeated
measures (AB/BA crossover) study designs (Montoya & Hayes, 2015). This analytic approach is
similar to conventional between-person mediation analyses, but uses difference scores for the
mediator and outcome variables to eliminate interdependence between repeated mediator
measures and repeated and outcome measures (Montoya & Hayes, 2015). Bootstrap resampling
was used to determine the confidence limits for formally testing the significance of the indirect
effect (MacKinnon, Fairchild, & Fritz, 2007; MacKinnon, Lockwood, & Williams, 2004;
Montoya & Hayes, 2015). The bootstrap method is useful because 1) no assumptions about the
shape of sampling distribution of the indirect effect are necessary (which is an issue because the
sampling distribution of the indirect effect is usually positively skewed and kurtotic) (Preacher,
Rucker, & Hayes, 2007), 2) it is appropriate for smaller sample sizes (MacKinnon et al., 2007),
and 3) it has greater power to detect indirect effects than other methods (Hayes, 2009). If the
confidence limits of the indirect effect derived by the bootstrap method do not include zero, the
null hypothesis of no indirect effect is rejected (Hayes, 2009; Preacher et al., 2007).
51

Aim 3 was modified post-hoc as a consequence of the results of aim 2. The original
purpose of aim 3 was to examine whether habitual total sugar intake and impulsivity moderated
the mediating effect of negative mood on the relationship between meal condition and ad libitum
sugar intake. Rather than testing the proposed moderated mediation, aim 3 was modified to test
whether habitual total sugar intake and impulsivity moderate the effect of meal condition on ad
libitum sugar intake. The updated hypotheses for this aim were:
Hypothesis 3c: Habitual total sugar intake will moderate the effect of meal condition on
ad libitum sugar intake. Compared to individuals with low habitual total sugar intake,
individuals with high habitual total sugar intake will consume more ad libitum total sugar
in the LSHF condition.
Hypothesis 3d: Impulsivity will moderate the effect of meal condition on ad libitum
sugar intake. Compared to individuals with low impulsivity, those with high impulsivity
will consume more ad libitum total sugar in the LSHF condition.
Updated hypotheses 3c and 3d were examined using a repeated measures analysis via
PROC MIXED. The analytic approach was similar to that used for conducting multilevel
modeling with dyadic data, because the structure of the data for the current study is similar to the
structure of dyadic data (O'Reilly et al., 2015). In dyadic data, interdependence of outcome data
between actors and partners occurs within each pair group (O'Reilly et al., 2015). Similarly, the
outcome data from the current study was interdependent between meal conditions within each
participant. Using a dyadic multilevel modeling approach accounts for statistical dependency of
data between paired observations (between actors and partners in the case of dyadic data, or
between meal conditions, in the case of the current study) (Bolger & Shrout, 2007; O'Reilly et
al., 2015). An additional benefit of using a dyadic multilevel modeling approach is that it allows
52

for simultaneous estimation of parameters for each meal condition separately, but within the
same model. This approach improves power over other possible analytical techniques, for
example, using separate regression models for each meal condition (Lyons & Sayer, 2005;
O'Reilly et al., 2015). To conduct this type of analysis, the data structure was converted into a
“stacked” structure such that the meal condition variable (where LSHF = 0 and HSLF = 1) was
separated into two different dichotomous variables, one for the HSLF condition (HSLF = 1,
LSHF = 0) and one for the LSHF condition (LSHF = 1, HSLF = 0) (O'Reilly et al., 2015). An
example of this data structure is presented in Table 2-4.
Table 2-4. Data structure for dyadic multilevel analyses
Subject ID Meal Condition
a
HSLF Meal

LSHF Meal

1 0 0 1
1 1 1 0
2 0 0 1
2 1 1 0
a: 0 = LSHF condition; 1 = HSLF condition
HSLF = high sugar/low fiber meal condition; LSHF = low sugar/high fiber meal condition

Two separate models were run to test hypotheses 3c and 3d: one to examine the
moderating effect of habitual total sugar intake, and one to examine the moderating effect of
impulsivity. Separate models were used to examine each moderator variable due to a smaller
subsample with complete habitual dietary intake data. The dichotomous HSLF and LSHF
variables were entered into the models as separate variables in order to estimate intercept
parameters for each meal condition. Interactions between the separate meal condition variables
and the predictor variable of interest (for example, HSLF*impulsivity) allowed for modeling of
the moderating effect for the variable of interest in each meal condition. This provided parameter
estimates for the moderating effect of the moderator of interest for each meal condition
separately, while retaining increased power by modeling the data from both meal conditions
53

simultaneously. Each model controlled for randomization order, sex, BMI z-score, and ethnicity.
The analytic sample size for the analysis testing moderation by impulsivity level was N =
83. In addition to the 3 participants who were excluded for missing ad libitum total sugar data for
both meal conditions, 1 participant was excluded from the analysis for missing impulsivity data.
The analytic sample size for the analysis testing moderation by habitual sugar intake level was N
= 64. A total of 23 participants were excluded from the analysis. In addition to the 3 participants
who were excluded for missing ad libitum intake data, 20 participants were excluded due to
missing or unusable 24 hour dietary recall data. Of the 20 participants who were excluded due to
24 hour recall data, 6 participants did not have 24 hour dietary recall data, 9 participants had only
1 day of dietary recall (valid recall data had to include at least 2 days of dietary recall data), 4
participants had greater than 1 month between each 24 hour dietary recall data collection (valid
recall data had to have less than 1 month between each recall date), and 1 participant was
excluded due to implausible 24 hour dietary recall energy intake. In order to screen for
implausible energy intake, standardized residuals were obtained from a regression of mean
kilocalories on BMI. A cut-off of ± 3 SD was used to determine implausible energy intake. The
participant who was excluded was excluded due to implausibly low reported energy intake based
on the cut-off.
Results
Sample characteristics for snack period negative mood, habitual total sugar intake, and
impulsivity are displayed in Table 2-5.

54

Table 2-5. Snack period negative mood, habitual sugar intake, and impulsivity sample
characteristics
Variable Mean (SD) [range]
HSLF snack period negative mood 4.89 (7.7) [0-53.5]
a
LSHF snack period negative mood 5.68 (8.0) [0-70.8]
a
Habitual total sugar intake (grams) 106.8 (50.3) [14.9-258.7]
Impulsivity 1.8 (0.5) [1.0-3.6]

Results from the repeated measures ANCOVA test of hypothesis 1 indicate that there was a
significant difference in the amount of total sugar participants consumed during the ad libitum
snack period between the two meal conditions. The mean amount of ad libitum total sugar
consumed during the LSHF meal condition was greater than the mean amount of ad libitum total
sugar consumed during the HSLF meal condition (LSHF mean = 78.63 ± 38.84 grams, HSLF
mean = 70.86 ± 37.73 grams; F = 9.64, p =0.002).
The results from the within-person mediation analysis are presented in Table 2-6. Model
fit indices (chi-square goodness of fit, CFI, TLI, and RMSEA) indicate that the model had
acceptable fit. Results from the mediation analysis show that the indirect effect was not
statistically significant (indirect effect estimate = -0.04, 95% CI: -0.15 – 0.01). This indicates
that negative mood reported during the snack period did not mediate the relationship between
meal condition and ad libitum sugar intake.

HSLF = high sugar/low fiber meal condition; LSHF = low sugar/high fiber meal
condition; SD = standard deviation
a
: Repeated measures ANOVA test showed snack period negative mood in the HSLF
and LSHF meal conditions were not significantly different
55

Table 2-6. Results from within-person mediation analysis examining the mediating effect of
negative mood on the relationship between meal condition and ad libitum sugar intake

Model fit
Parameter Estimate p-value
Chi-square 0.003 0.9592
CFI 1.00 -
TLI 1.00 -
RMSEA
a
0 -
Mediation analysis results
Parameter Estimate (SE) 95% CI
Direct effect
0.84 (0.54) -0.02 - 1.77
Indirect effect -0.04 (0.05) -0.15 - 0.01
Total effect 0.80 (0.55) -0.08 - 1.74
a: p-value for the RMSEA is reported as the probability that the RMSEA is
<0.05
SE: standard error; 95% CI: 95% confidence interval
The moderated mediation analysis originally proposed for aim 3 could not be conducted
due to null findings of aim 2. Alternatively, tests of moderation by impulsivity and habitual total
sugar intake on the effect of meal condition on ad libitum total sugar intake were conducted.
Results from these analyses are presented in Tables 2-7 and 2-8. Results from the mixed model
examining the moderating effect of impulsivity showed that higher impulsivity was associated
with higher ad libitum sugar intake in both meal conditions. The contrast estimate (which
compared the relationship between impulsivity and ad libitum sugar intake between the meal
conditions) showed that there was no significant difference in the relationship between
impulsivity and ad libitum sugar intake between the LSHF and HSLF conditions. Results from
the model examining the relationship between habitual total sugar intake and ad libitum total
sugar intake showed that there was no relationship between habitual total sugar intake and ad
libitum total sugar intake in either meal condition.

56

Table 2-7. Results from model examining the moderating effect of impulsivity on the
relationship between meal condition and ad libitum total sugar intake
a

Parameter Estimate (SE)
HSLF 0.202 (0.004)***
LSHF 0.212 (0.039)***
HSLF*impulsivity 0.034 (0.01)*
LSHF*impulsivity 0.029 (0.01)*
HSLF*randomization order (ref = LSHF then
HSLF)
0.004 (0.015)
LSHF*randomization order (ref = LSHF then
HSLF)
-0.007 (0.015)
HSLF*sex (ref = female) 0.028 (0.015)
LSHF*sex (ref = female) 0.046 (0.015)*
HSLF*BMI-z score -0.002 (0.015)
LSHF*BMI-z score -0.001 (0.01)
HSLF*ethnicity (ref = African American) -0.016 (0.016)
LSHF*ethnicity (ref = African American) -0.032 (0.015)*
Contrast Estimate
LSHF*impulsivity vs HSLF*impulsivity 0.006 (0.012)

a
ad libitum total sugar intake square root transformed and adjusted for ad libitum
energy intake
*p<0.05, **p<0.01, p<0.0001
SE: standard error; HSLF: high sugar/low fiber meal condition; LSHF:
low sugar/high fiber meal condition; ref = reference group; BMI-z = body
mass index z-score

57

Table 2-8. Results from model examining the moderating effect of habitual total sugar
intake on the relationship between meal condition and ad libitum total sugar intake
a

Parameter Estimate (SE)
HSLF 0.276 (0.043)***
LSHF 0.280 (0.045)***
HSLF*habitual total sugar intake 0.124 (0.160)
LSHF*habitual total sugar intake 0.2351 (0.176)
HSLF*randomization order (ref = LSHF then
HSLF)
-0.027 (0.016)
LSHF*randomization order (ref = LSHF then
HSLF)
-0.017 (0.018)
HSLF*sex (ref = male) -0.023 (0.017)
LSHF*sex (ref = male) -0.032 (0.018)
HSLF*BMI-z score 0.005 (0.015)
LSHF*BMI-z score 0.006 (0.017)
HSLF*ethnicity (ref = African American) -0.026 (0.016)
LSHF*ethnicity (ref = African American) -0.036 (0.018)*

a
ad libitum total sugar intake square root transformed and adjusted for ad libitum
energy intake
*p<0.05, ***p<0.0001
SE: standard error; HSLF: high sugar/low fiber meal condition; LSHF: low
sugar/high fiber meal condition; ref = reference group; BMI-z = body mass
index z-score

Discussion
Findings from this experimental cross-over design study demonstrate that the HSLF and
LSHF meal conditions had different effects on how much total sugar the overweight adolescent
participants consumed during the ad libitum snack periods. In accordance with the hypothesis,
participants ate more total sugar during the snack period in the LSHF condition than in the HSLF
condition. This finding suggests that when participants were restricted from eating sugar during
breakfast and lunch in the LSHF condition, they compensated by eating more sugar when they
were given a choice of foods to eat during the snack period. Findings from previous studies
58

support the notion that restriction of certain foods is causally linked to overeating of the
restricted food when the restricted food is available. A study of 3 to 5 year old children found
that those whose parents placed greater restriction on the children’s intake of palatable snack
foods in daily life consumed greater amounts of those snack foods when they were freely
available in an experimental setting (Jennifer O Fisher & Leann L Birch, 1999). Another study in
3 to 6 year old children found that restricted access to palatable foods lead to increased focus on,
preference for, selection of, and consumption of the restricted food over other types of foods in
an ad libitum setting (Jennifer Orlet Fisher & Leann Lipps Birch, 1999). Restricting access to
foods in children has also been shown to lead to a greater self-reported preference for the
restricted food (Birch et al., 2001a). The findings from the current study are novel because they
are the first to reveal that restriction of sugar overall, not just restriction of a specific type of
snack food, can lead to compensatory sugar intake in humans. While studies have demonstrated
compensatory sugar intake in animals, no previous studies have examined and demonstrated this
behavior in humans. This finding is particularly salient for overweight adolescents, an important
target group for dietary interventions to reduce sugar intake. Further, previous research has
focused on consequences of food restriction by parenting practices in young children. To our
knowledge, no previous studies have examined the consequences of food restriction in a
controlled experimental setting in adolescents. Adolescence is a time during which individuals
begin to develop and express greater agency in their decisions, such as which foods they
consume (Bassett, Chapman, & Beagan, 2008). The current study extends previous findings in
younger children by showing that compensatory intake in the face of restriction can also occur in
adolescents, who are in a critical developmental period for solidifying autonomous dietary
habits.
59

The adolescents in this sample compensated for being deprived of high-sugar foods
during the LSHF condition by consuming a mean of approximately 8 more grams of total sugar
than in the HSLF condition. The extra 8 grams of sugar is equivalent to 2 teaspoons, or about 32
calories. While this added caloric intake may appear to be small, it is not necessarily
insignificant. If compensatory sugar intake were to happen in a free-living setting, for example in
an attempt to improve dietary behaviors by reducing sugar intake, it could mean increased risk
for excess weight gain. Small, incremental increases in caloric intake over time can lead to
positive energy balance (i.e. greater energy intake than expenditure), which contributes to
gradual weight gain (Hill et al., 2009). This phenomenon has been explained as occurring
through a “ratchet effect”, whereby 1) small increases in caloric intake lead to incremental
positive energy balance, 2) increased energy balance leads to weight gain, 3) higher weight leads
to a higher energy requirement, and 4) higher energy requirement necessitates increased energy
intake to maintain energy balance homeostasis (Hill et al., 2009).
Indeed, studies on the association between excess energy intake and weight gain in youth
have shown that weight gain can be caused by intake of relatively few extra kilocalories per day
(Hill et al., 2009; Wang, Gortmaker, Sobol, & Kuntz, 2006). A longitudinal study examining
energy gaps and weight gain in youth found that small amounts of surplus energy intake, as few
as 22 kilocalories per day, were associated with excess weight gain over time and a higher risk
for overweight (Plachta ‐Danielzik et al., 2008). The authors of the aforementioned study
suggested that eliminating as few as 46 kilocalories per day of extra caloric intake could prevent
development of overweight in children (Plachta ‐Danielzik et al., 2008). It is possible that in a
free-living situation, a similar consumption pattern as that observed in the current study during
60

the LSHF snack period, extrapolated out over the period of a day, could lead to an even greater
positive energy balance and risk for weight gain.
In addition to causing excess caloric intake, a compensatory consumption response to
sugar restriction could also have implications for dietary carbohydrate quality. Sugar
consumption tends to displace intake of other, more healthful foods, such as fiber-rich vegetables
(Frary, Johnson, & Wang, 2004). Poor dietary carbohydrate quality has implications for greater
obesity, metabolic syndrome, cardiovascular disease, and type 2 diabetes risk (Gaesser, 2007;
Slyper, 2013). Compensatory intake of high-sugar foods following an attempt at modifying sugar
intake via dietary behavior changes could have adverse consequences for dietary behavior
change interventions. If overweight adolescents are likely to consume greater amounts of sugar
after attempting to mitigate sugar intake, this could have adverse consequences for adhering to
high fiber, low sugar dietary behavior change.
Contrary to the directional hypothesis, participants’ negative mood did not mediate the
relationship between meal condition and ad libitum total sugar intake in this sample.
Additionally, the amount of sugar that the participants reported habitually consuming in daily life
did not affect how much sugar they consumed during the ad libitum snack period in either meal
condition. This finding is surprising given that habitual sugar intake in this sample was high,
averaging 106.8 ± 50.3 grams (21.7 ± 10.2 teaspoons) of sugar per day. It is possible that more
stable, trait-like characteristics influence the compensatory sugar intake reaction to sugar
restriction in this sample of overweight adolescents. This notion is supported by the findings for
the moderating effect of impulsivity. In partial accordance with the hypothesis, greater
impulsivity was associated with greater ad libitum sugar intake in the LSHF meal condition.
However, higher impulsivity was also associated with greater ad libitum sugar intake in the
61

HSLF condition, and there were no significant differences on the impact of impulsivity between
meal conditions. This finding may suggest higher impulsivity predisposed participants to
consume more sugar when they had free access to sugar during the snack periods, regardless of
whether participants were provided access to sugar or restricted from sugar intake during the
breakfast and lunch tests meals. The notion that impulsivity may have been partially responsible
for driving higher ad libitum sugar intake in the LSHF and HSLF meal conditions is supported
by previous research. One study examining the association between impulsivity and food choice
in adults showed that people with higher impulsivity tended to choose highly palatable chocolate
cake over a more healthful salad compared to people with lower impulsivity (Sengupta & Zhou,
2007). The authors of the study found that this may have been due to more highly impulsive
people experiencing greater temptation toward the highly palatable food, rather than having
lower self-control (Sengupta & Zhou, 2007). Further research is necessary to understand the
mechanisms linking impulsivity to higher sugar intake in overweight adolescents, as well as
other characteristics that may have driven greater ad libitum sugar intake in the LSHF meal.
Finding from the current study have important potential implications for dietary
interventions. The evidence that participants ate on average more sugar when given a choice of
snacks after being deprived of sugar throughout the LSHF day indicates that steps should be
taken when encouraging dietary behavior change to prevent a rebound increase in sugar intake.
In particular, interventions should include components that address the needs of individuals with
impulsive tendencies. Designing interventions that address these issues could help to improve the
success of dietary behavior changes over time. Further research is necessary, however, to
understand additional reasons why a compensatory reaction to sugar restriction may occur in
overweight adolescents, and whether these can serve as targets in dietary interventions.
62

Strengths and Limitations
This study has several strengths. The controlled environment of the in-laboratory visits
allowed for objective measurement of ad libitum sugar intake and real-time measures of mood.
The crossover design of the study enabled a comparison of the influences of both HSLF and
LSHF meal consumption on ad libitum total sugar consumption in the same participants.
Focus groups of Latino and African American participants were used to determine the foods
included in the test meals, so the meals represented breakfast foods that were typical of these
populations.
This study also has limitations that should be noted. Impulsivity was measured using a
single subscale of the UPPS Impulsive Behavior scale, the negative urgency subscale. This
subscale measures the tendency for one to act rashly in response to negative feelings. The fact
that participants with higher impulsivity (negative urgency) consumed greater amounts of ad
libitum sugar suggests that they may have been impulsively responding to negative feelings
precipitated by sugar restriction with compensatory sugar consumption. However, negative mood
was not related to sugar consumption in this study. It is unclear why these findings do not
correspond. It is possible that a different measure of negative mood or a more comprehensive
measure of impulsivity are necessary to understand the nature of these relationships. There is a
possibility that other variables that were not tested could mediate the hypothesized relationships
between meal condition and ad libitum total sugar intake. For example, food cravings could have
partially explained the relationship between meal condition and ad libitum sugar intake, but this
variable was not measured in the current study. The rigorous design of the in-lab experiment is a
strength of this study. However, because the data are from an in-lab experiment, findings may
not be generalizable to free-living situations. For example, participants likely consumed
63

different, and possibly less healthful, foods with different macronutrient compositions in their
daily lives than what was provided during the breakfast and lunch meals in each experimental
condition. Participants were isolated during the experimental conditions. Consequently, the
laboratory environment lacked contextual factors that could impact mood and eating behaviors,
such as social interactions (Patel & Schlundt, 2001), the home environment (Campbell et al.,
2007), and the school environment (Story, Neumark-Sztainer, & French, 2002). Finally,
participants were aware that they were being observed, which may have influenced their eating
behaviors.
Conclusions
The findings from this study suggest that when overweight adolescents are deprived of
sugar, they compensate by eating more sugar when they subsequently have access to high-sugar
foods. Findings also suggest that overweight adolescents with higher impulsivity are predisposed
to consume greater amounts of sugar than those with lower impulsivity, regardless of previous
levels of sugar consumption. Attempts to reduce sugar intake in overweight adolescents may
have the unintended consequence of leading to greater sugar intake as a reaction to sugar
restriction. Findings from this study have implications for dietary interventions, notably that the
potential for compensatory intake should be considered when designing interventions to improve
dietary carbohydrate quality. Given that the food environment is swamped with high-sugar, low-
fiber food options, it is inevitable that an individual attempting to reduce sugar intake in his or
her diet will be exposed to sugar dense foods. Understanding how to reduce the risk of rebound
sugar consumption may be an important step for improving success of dietary interventions.
Future research should focus on understanding why a compensatory sugar intake reaction occurs
in response to sugar restriction, and design interventions to address this issue.
64

Chapter 3: Impact of negative mood on hunger ratings in response to high sugar/low fiber
versus low sugar/high fiber meals

Introduction
The quality of carbohydrates in the typical American diet is poor and skewed toward
excessive sugar intake and insufficient fiber intake. According to recommendations by the
American Heart Association, women should limit consumption of added sugar to 6 teaspoons
per day or less (equaling about 100 calories), and that men limit consumption to 9 teaspoons per
day or less (equaling about 150 calories) (Thompson & Veneman, 2005). However, intake of
added sugar far exceeds both of these recommendations; a recent estimate indicated that average
daily consumption of added sugar in the U.S. by adults is 22.2 teaspoons per day, constituting
355 calories (Johnson et al., 2009). This issue is prevalent across age groups. A recent study of
the latest available National Health and Nutrition Examination Survey (NHANES) data found
that the 16% of the average daily energy intake by adolescents in the U.S. comes from added
sugars (Zhang, Gillespie, Welsh, Hu, & Yang, 2015). This average exceeds the World Health
Organization’s recommendation for added sugars to make up less than 10% of average daily
caloric intake for this age group. The issue of high added sugar in the diet is particularly critical
because added sugar tends to displace other, more healthful dietary components, such as fiber-
rich fruits and vegetables (Frary et al., 2004).
In addition to consisting of high amounts of added sugar, the typical American diet also
consists of inadequate fiber intake. According to the American Heart Association, the daily fiber
intake recommendation for adults is 28 to 38 grams (King et al., 2012), and the recommendation
for adolescents is 26 to 38 grams (Slavin, 2005). However, average fiber intake in adults is
estimated to be 15.9 grams, less than half of the upper recommended amount (King et al., 2012).
65

A recent study found that less than 60% of a nationally representative sample of youth consumed
the recommended daily amount of fiber (Brauchla et al., 2012). Poor carbohydrate quality in the
diet is particularly pronounced in African American and Hispanic adolescents, who tend to
consume greater amounts of foods high in added sugar and lower amounts fiber-rich foods such
as fruits and vegetables than youth of other ethnicities (Ronette R. Briefel & Clifford L. Johnson,
2004; Choi et al., 2006; Mendoza et al., 2006; Reynolds & Spruijt-Metz, 2006).
Diets high in sugar and low in fiber are particularly detrimental to Hispanic and African
American youths because these populations are at higher risk of developing overweight and
obesity over the life course. A 2013 systematic review and meta-analysis of randomized
controlled trials and prospective cohort studies examined evidence of the impact of added sugars
on body weight (Te Morenga et al., 2013). Findings from the meta-analysis showed that
increasing intake of added sugar in free-living populations was associated with a 0.75 kg
(approximately 1.7 lbs) increase in body weight over intervention periods of up to eight weeks
(Te Morenga et al., 2013). A decrease in added sugar intake was associated with a 0.8 kg
(approximately 1.8 lbs) reduction in body weight over a period of ten weeks to eight months
(depending on the intervention period) (Te Morenga et al., 2013). The authors posited that the
association between increased added sugar intake and increased body weight is driven by an
overall increase in energy intake (Te Morenga et al., 2013). The studies analyzed in this meta-
analysis did not present findings on specific subcategories of sugar, so findings from the meta-
analysis did not speak to the impacts of specific types of sugar on body weight (Te Morenga et
al., 2013). Nonetheless, the findings show compelling evidence for a strong positive association
between intake of added sugars and weight gain.
66

In contrast to sugar intake, fiber intake may be protective against overweight and obesity
(Slavin, 2005). Large epidemiologic studies have shown that fiber intake is higher among
populations with lower prevalence of obesity compared to populations with higher prevalence of
obesity (Slavin, 2005). Evidence from large prospective studies also supports an inverse
association between fiber intake and adiposity. For example, findings from the Nurses Health
Study have shown that high fiber intake is inversely related to weight gain (Liu et al., 2003;
Slavin, 2005). An inverse association between fiber intake and body weight is further supported
by evidence from intervention studies. In one study by Howarth and colleagues, increased fiber
intake lead to an average weight loss of 1.9 kg (approximately 4.2 lbs) over a period of four
months (Howarth, Saltzman, & Roberts, 2001; Slavin, 2005). Fiber intake may lead to weight
loss and facilitate weight maintenance because fiber intake is associated with decreased energy
density in the diet and overall decreased energy intake (Rolls, Ello-Martin, & Tohill, 2004;
Slavin, 2005). Many knowledge gaps exist about what individual-level factors may drive people
to overconsume low fiber, high diets. Investigating possible individual-level mechanisms might
lead to high sugar intake and low fiber intake could provide useful insights for dietary
interventions.
One of the factors that may perpetuate this poor dietary profile is the impact of sugar on
mood. Sugar consumption has been shown to have a mood-elevating effect in some individuals.
Anecdotal evidence indicates that many people consume foods high in added sugar when they
experience feelings of low mood or low energy to achieve a short-term mood boost (Benton &
Nabb, 2003). Findings from experimental feeding studies also show that sugar intake can have at
least a short-term positive impact on mood. In one study, healthy young adult participants
consumed a glucose containing or a non-caloric breakfast drink (Benton, 2002). Higher blood
67

glucose levels in the participants who consumed the glucose containing drink were associated
with feeling less tense (Benton, 2002). The acute impacts of high-sugar foods on mood may
perpetuate sugar intake at the expense of fiber intake, especially in individuals who are
accustomed to habitually consuming high amounts of sugar. There is evidence that habitual
consumption of highly palatable foods (like those high in sugar) can have anxiolytic effects in
some individuals, and that removal of those foods from the diet my increase feelings of anxiety
and precipitate a return to eating those foods (Lowe & Butryn, 2007). It is possible that this
could also be driven by impacts of mood on hunger.
Evidence indicates that mood can influence feelings of hunger. Hunger has been found to
be more intense during periods of negative mood in some individuals (Macht & Simons, 2000).
According to the psychosomatic hypothesis, negative mood can influence hunger in individuals
who have difficulty differentiating negative moods from physical signals of hunger and satiety
(Bruch, 1964; Greeno & Wing, 1994). Negative mood may impact hunger in these individuals
because they learned to link these factors early in life (Greeno & Wing, 1994). If removal of
palatable foods such as those high in sugar from the diet can lead to negative mood in some
individuals (Lowe & Butryn, 2007), such a dietary change may also lead to increased feelings of
hunger. This may provide insight into why it is difficult for many individuals to maintain low-
sugar, high-fiber diets. However, to our knowledge the impact of replacing high-sugar foods with
high-fiber foods on mood and hunger in habitual lower vs. higher sugar consumers has not been
previously examined.
Overall, the literature shows that there are mixed findings on the effects of sugar and
fiber on feelings of hunger. In normal weight individuals, high sugar intake has been shown to
lead to a rapid return of hunger in some individuals (Brand-Miller et al., 2002; Rennie et al.,
68

2005; Roberts, 2000), while high fiber intake has been shown to sustain satiety and lead to a
slower return of hunger in some individuals (D. J. A. Jenkins & Jenkins, 1985; Roberts, 2000).
These relationships are supported by a study that examined the effects of snacks differing in fiber
and sugar content on short-term hunger in lean participants, which found that the high-fiber
snack suppressed hunger more than the high-sugar snack (Furchner-Evanson, Petrisko, Howarth,
Nemoseck, & Kern, 2010). However, another study found no difference between high glycemic
load (generally characteristic of high-sugar, low fiber foods) and low glycemic load (generally
characteristic of high-fiber, low sugar foods) foods on hunger (Alfenas & Mattes, 2005). These
different findings could be due to factors affected by sugar and fiber intake beyond physiological
responses, such as mood, the effect of mood on feelings of hunger, and the influence of habitual
sugar intake on the relationships between mood and hunger.
Habitual intake of palatable foods has been shown to affect mood, and previous studies in
animals and humans have shown that characteristics of habitual dietary intake can influence the
effects of experimental meals on hunger. An experimental feeding study in humans that
examined the impact of high-sugar compared to low-sugar lunches on short-term hunger found
that the effects of the lunches differed by habitual consumption of sweet low-calorie drinks (Low
et al., 2014). In low consumers of sweet low-calorie drinks, consumption of a high-sugar lunch
lead to a rapid return in short-term hunger ratings, whereas no effect of a high-sugar lunch on
hunger was found in habitual high consumers of sweet low-calorie drinks (Low et al., 2014). In a
subsample of participants from the current proposed study, our research group found that
compared to participants who habitually consumed high amounts (≥3 servings per day) of sugar
sweetened beverages (SSBs), those who consumed low amounts (≤1 serving per day) reported
lower hunger ratings during an ad libitum snack period that occurred subsequent to high-sugar
69

low fiber (HSLF) and low-sugar high fiber (LSHF) test meals (Shearrer, 2015). This evidence
indicates that certain characteristics of habitual dietary intake can influence the effects of high-
sugar and high-fiber experimental meals. However, it is unknown whether overall habitual total
sugar intake influenced hunger ratings in these participants, or whether hunger ratings were
influenced by negative mood.
Understanding how replacing high sugar carbohydrates with high fiber carbohydrates can
impact mood and hunger in habitual sugar consumers may provide insight into why dietary
behavior changes – particularly replacing high-sugar carbohydrates with more healthful high-
fiber carbohydrates – are difficult to maintain in some individuals. Testing these relationships in
the context of a controlled experimental feeding study gives the opportunity to examine multiple
repeated mood and hunger ratings to examine how they may be associated over time. The study
used data from a randomized experimental crossover design feeding study in a sample of
overweight Hispanic and African American adolescents to examine how consumption of high
sugar/low fiber (HSLF) vs. low sugar high fiber (LSHF) test meals a) affected how negative
mood influenced hunger over the course of in-lab observation periods and b) whether habitual
total sugar intake moderated the relationship between negative mood and hunger.
Specific Aims and Hypotheses
The specific aims of this study were:
Aim 1: To examine the influence of negative mood ratings on initial hunger ratings and change
in hunger ratings throughout the day in HSLF (i.e. sugar exposure) vs. LSHF (i.e. sugar
restriction) test meal conditions in overweight minority adolescents, and whether the relationship
between negative mood and hunger was moderated by habitual total sugar intake.
70

Hypothesis 1a: During both the HSLF (sugar exposure) and LSHF (sugar restriction) test
meal conditions, higher initial negative mood ratings will be associated with higher
fasting hunger ratings, and this relationship will be moderated by habitual sugar intake
(i.e. there will be a significant positive interaction between initial negative mood and
habitual total sugar intake on fasting hunger ratings). This is hypothesized to occur in
both the HSLF and LSHF conditions because initial negative mood and fasting hunger
ratings were assessed at time 0, prior to exposure to the test meal conditions.
Hypothesis 1b: Higher negative mood will be associated with an increase over time in
hunger ratings in the LSHF (sugar restriction) condition but not in the HSLF condition.
This relationship will be moderated by habitual sugar intake, such that in the LSHF
condition, higher negative mood ratings will be associated with an increase over time in
hunger ratings for participants with higher habitual sugar intake.
Methods
Study Sample
The sample was composed of participants from the Food Adolescence Mood and
Exercise (FAME) study, a randomized experimental crossover design feeding study conducted at
the University of Southern California. The FAME study investigated the acute effects of meals
that differed in sugar and fiber content on physiology, mood, and behavior in overweight and
obese African American and Hispanic adolescents. The total sample included 87 participants.
Participant recruitment
Participants were recruited from clinics, hospitals, churches, schools, health fairs, and
other gatherings in the Los Angeles area from 2007–2010 by an ethnically diverse staff. To be
included in the study, participants were required to meet the following inclusion criteria: a)
71

African American or Hispanic/Latino ethnicity (all four grandparents reported to be African
American or Hispanic/Latino), b) male or female between 14 and 17 years old, and c) body mass
index BMI ≥ 85
th
percentile for age and sex. Potential participants were excluded if they: a) had a
diagnosis of diabetes, b) were participating in a weight loss or exercise program, c) used
medications that influenced body weight or insulin sensitivity, or d) had a diagnosis of a
syndrome that influences body composition. Prior to study procedures, informed written parental
consent and participant assent were obtained. The study was approved by the Institutional
Review Board of the University of Southern California. Data were collected between 2008 and
2010. The characteristics of the study sample are presented in Table 3-1.
Table 3-1. FAME study sample characteristics (n=87)

Variable Mean (SD)
Age (years) 16.3 (1.2)
BMI z-score 2.02 (0.52)
Weight status
Overweight 5.7% (5)
Obese 94.3% (83)
Insulin Sensitivity (SI) 1.6 (1.1)
Sex
1

Male 48.9% (43)
Ethnicity
1

African American 43.2% (38)
Latino 56.8% (50)

Mean and (SD) reported, unless otherwise
indicated
1
: Frequencies (N)
SD = standard deviation; BMI = body mass index

72

Experimental Design
The controlled experimental meal conditions employed in this study consisted of a high
sugar/low fiber (HSLF) condition and a low sugar/high fiber condition (LSHF), during which
participants ate condition-specific breakfast and lunch meals. The meal conditions were
developed using data from focus groups conducted at the University of Southern California with
the purpose of identifying foods typically eaten by the study population. Focus group participants
were 12 African American (5 male, 7 female) and 19 Hispanic/Latino (9 male, 10 female)
adolescents. The contents of the HSLF and LSHF meals were determined based on focus group
participant feedback. Foods included in each meal condition are listed in Table 3-2. The meal
conditions were isocaloric and matched for macronutrients except for the experimental
components (sugar and fiber content), and were determined using the Nutrient Data System for
Research (NDS-R 2010, University of Minnesota, Minneapolis, MN). Portions for the test meals
were determined for each participant based on 20% of the participant’s daily caloric needs,
which were calculated using sex, age, height, and body weight using the Dietary Reference
Intake Guidelines Estimation of Energy Expenditure for overweight children ages 3-18 (Hellwig
et al., 2006). An unblinded Registered Dietician designed and prepared or supervised the
preparation of all test meals. Meal macronutrients are presented in Table 3-3. Each participant
underwent both meal conditions in separate 8-hour in-lab observation visits that were conducted
2 to 4 weeks apart. Participants were randomly assigned to a HSLF/LSHF or a LSHF/HSLF visit
order using a stratified block design randomization procedure. The study visits were identical
except for the experimental meal conditions.

73

Table 3-2. HSLF and LSHF test meal contents
HSLF meals LSHF meals

Regular Pop-tart
(Kellogg NA Co., Battle Creek, MI)

Thomps o n ’ s 100% Whole-wheat bagel
(Bimbo Bakeries, Horsham, PA)
Calcium-enriched string cheese
(Sargento Mootown Light String Cheese, Sargento Food Inc.,
Plymouth, WI)
Margarine
(I Ca n’t Believe It ’s Not Butter Light, Unilever PLC/Unilever
N.V., Englewood Cliffs, NJ)

Tampico juice
(Tampico Beverages, Chicago, IL)
Water treated with soluble fiber powder
(Benefiber Powder, Novartis Consumer Health Inc., Parsippany,
NJ)

Table 3-3. HSLF and LSHF test meal nutrient compositions

Macronutrient
g (% kcal)
HSLF Meal LSHF Meal
54.0g Poptart 61.0g whole wheat bagel
42.0 string cheese 14.0g margarine
247.0g juice 10.5g Benefiber supplement
Fat 11.0 (24%) 9.5 (24%)
Carbohydrate 64.0 (61%) 61.0 (68%)
Protein 14.0 (13%) 10.0 (11%)
Sugar 41.0 (39%) 7.0 (8%)
Fiber 1.0 (1%) 16.0 (18%)

In-lab Procedures
Study procedures took place at the Spruijt-Metz observation laboratory at the University
of Southern California’s Health Science Campus. Participants began each study visit after a 10-
hour overnight fast. During each visit, breakfast and lunch meal times were standardized to 15
minutes, and participants were instructed to eat the entire meal for both breakfast and lunch.
After finishing the breakfast meal, participants were instructed to choose from activities
available in the laboratory for an 8-hour observation period. At arrival (prior to breakfast) and
74

every 30 minutes after the start of the breakfast meal, participants completed self-report
questionnaires that assessed appetite and mood.
Measures
Outcome
Hunger rating. Hunger was measured via a single-item visual analogue scale (VAS).
Participants were asked to rate how much hunger they felt at the time of measurement by
indicating with a mark on a 100 mm line that had endpoints labeled “no hunger” (corresponding
to 0) and “the greatest imaginable hunger” (corresponding to 100). The item yielded a single
score between 0 and 100 (corresponding to the number of millimeters from 0, or “no hunger”, to
where the participant placed the mark on the line). The VAS for hunger is depicted in Figure 3-
1. Participants self-reported their hunger rating every 30 minutes from the start to the end of the
visit. A total of 18 hunger rating measures were conducted over each in-lab visit. This variable
was treated as a continuous, repeated measures variable in the analyses.

Figure 3-1. Visual analogue scale (VAS) for hunger

Predictors
Negative mood. Negative mood was measured via a four-item visual analogue scale
(VAS) adapted from the Profiles of Mood States (POMS) (Terry et al., 1999). The scale was
composed of one item for each of the following feelings: nervous, worried, anxious, and panicky.
Participants were asked to rate how nervous, worried, anxious, and panicky they felt at the time
of measurement by indicating with a mark on the corresponding 100 mm line that had endpoints
No
hunger
Greatest
imaginable
hunger
75

labeled “not at all” (corresponding to 0) and “extremely” (corresponding to 100). Each item
yielded a single score between 0 and 100 (corresponding to the number of millimeters from 0, or
“not at all”, to where the participant placed the mark on each line). The VAS for negative mood
is depicted in Figure 3-2. Participants self-reported their levels of nervousness, worry,
anxiousness, and panic every 30 minutes from the start to the end of the visit. A composite
negative mood score was obtained by averaging the nervous, worry, anxious, and panic scores at
each time point. A total of 18 negative mood rating measures were conducted over each in-lab
visit. This variable was treated as a time-varying, continuous variable in the analyses.
Figure 3-2. Visual analogue scale (VAS) for negative mood

Moderator
Habitual total sugar intake. Habitual total sugar intake was measured using a series of
three multiple pass 24 hour dietary recalls prior to the first in-lab visit. The dietary recalls were
conducted via phone call by trained interviewers using the Nutrient Data System for Research
(NDS-R 2010, University of Minnesota, Minneapolis, MN). The number of grams of dietary
sugar that participants reported habitually consuming was calculated using the NDSR system.
Habitual total sugar intake was calculated by averaging the reported number of grams of total
sugar intake from each dietary recall sessions. Habitual total sugar intake was split into “low”,
“medium”, and “high” tertiles and treated as an ordinal variable in the analyses.
76

Participant demographics. Prior to the two in-lab feeding visits, participants completed
an inpatient visit at the Clinical Trials Unit at the USC University Hospital where body weight,
height, and demographic data were collected. Age (years) and sex were self-reported by
participants and collected from the protocol flow sheets at the in-patient Clinical Trials Unit
visit. Height (cm) and weight (kg) were measured in triplicate by a Registered Nurse. Body mass
index (BMI) z-score was calculated using the height and weight measurements based on the
CDC age- and sex- specific growth charts (Kuczmarski et al., 2002).
Statistical Analyses
Analyses were conducted using SAS version 9.4 (Cary, NC). Significance for all
statistical tests was set at p < 0.05. Prior to the analyses, the distribution of the outcome variable,
self-reported hunger, was checked for normality and skew. The application of transformations
did not significantly improve the normality of the variable, so hunger was used in the analyses in
a non-transformed state.
To test hypotheses 1a and 1b, repeated measures analysis via SAS PROC MIXED was
used to examine the relationship between initial negative mood and fasting hunger ratings and
between negative mood and hunger ratings throughout the day during the HSLF and LSHF meal
conditions. The analysis also examined whether these relationships were moderated by habitual
total sugar intake (low intake vs. high intake). Two separate models were used to examine these
relationships, one for the HSLF meal condition and one for the LSHF meal condition. Each
model included 1) a 2-way interaction between negative mood and habitual total sugar intake
(negative mood*habitual total sugar intake) to examine whether the relationship between initial
negative mood and fasting hunger was moderated by habitual total sugar intake, and 2) a 3-way
interaction between time, negative mood, and habitual total sugar intake (time*negative mood*
77

habitual total sugar intake) to examine whether the association between time-varying negative
mood and change in hunger ratings throughout the day was moderated by habitual total sugar
intake.
The negative mood variable was a time-varying predictor that varied both between- and
within-person. An advantage of repeated-measures analysis is the ability to differentiate whether
the effects of a time-varying predictor on a time-varying outcome comes from differences
between people in the time-varying predictor or differences within person in the time-varying
predictor. In order to differentiate the effects of the between- and within- person variance of
negative mood on hunger, the negative mood variable was disaggregated into its between-person
(between-person) and within-person (within-person) effects (Curran & Bauer, 2011). These
between- and within-person effects were included as separate variables in each model. A priori
time-invariant covariates included ethnicity, sex, body mass index (BMI) z-score, and
randomization order. All continuous predictors were centered to reduce multicollinearity of the
interaction terms with their main effects and to aid in interpretation of the outcomes. The
between-person negative mood variable, habitual total sugar intake, and BMI z-score variable
were centered on their respective grand (sample) means, and the within-person negative mood
variable was centered on each person’s individual mean. For all models, the variance
components were estimated using the restricted maximum likelihood (REML) estimation
method. An autoregressive structure was specified for the residuals to specify that measures of
the hunger outcome variable that were reported closer together in time were more highly
correlated than measures that were reported farther apart. Random effects for intercept and time
were included in both the HSLF and LSHF models to allow fasting hunger ratings and hunger
rating trajectories over the course of the day to differ between participants.
78

Prior to conducting the full PROC MIXED analyses, mixed models were conducted to
calculate the intraclass correlation coefficients (ICCs) and examine the growth trajectories of the
dependent hunger variable in the HSLF and LSHF meal conditions. The ICCs and growth
trajectories were analyzed separately for each meal condition. These steps were taken to ensure
that repeated measures modeling was an appropriate analytic approach for the data. First,
unconditional means models were run to investigate the within-person and between-person
variance components of the dependent hunger variable. These variance components were used to
calculate the ICCs, which showed the proportion of total variance in the variable that was
attributable to differences between versus within participants. This information revealed whether
there was within-person variance in the hunger variable that could be modeled using a repeated
measures analysis (Singer & Willett, 2003). Second, the shape of the growth trajectories of the
hunger variable in the HSLF and LSHF conditions was examined. This was done to check
whether there was systematic change in the dependent variable and to inform the slope trajectory
(i.e. time trend) for the dependent variable in the full models used to test hypotheses 1a and 1b.
The results from the mixed models examining the growth trajectories of hunger in the
HSLF and LSHF conditions are shown in Tables 3-4 and 3-5. The ICC results showed that 81%
of the variance in hunger during the LSHF meal condition and 78% of the variance in hunger
during the HSLF condition came from between-person differences. This indicates that the
majority of the variance in the outcome was due to between-person factors, but there was also
some variance in the outcome that could be explained by within-person factors. The results also
indicated that there was systematic change in hunger over time during both meal conditions.
Examination of the hunger trajectory during both meal conditions showed that this variable had
significant fixed effects for a linear time trend. Models including quadratic and cubic time trends
79

showed that fixed effects for these higher order time trends were not significant. Comparison of
the fit statistics between the 1) unconditional growth model and 2) model that included a linear
time trend showed that the model including the linear trend (model “b” in Tables 4 and 5) had
the best fit. The significant linear time trend fixed effect and the superior model fit of the linear
trend model indicated that the full mixed models used to examine hypotheses 1a and 1b should
include a linear time trend to describe the growth trajectories of the dependent hunger variable.
Table 3-4. Results from preliminary analyses of ICC and growth trajectory of hunger
during LSHF meal condition

Parameter Model A (no change) Model B (linear change)
Fixed effects
Composite model

Intercept γ
00
35.75 (3.37)*** 44.08 (2.83)***
Time (linear term) γ
10
-0.82 (0.15)***
Variance components
Level 1 Within-person σ
ε
2
226.00 (8.39)*** 275.86 (24.71)***
Level 2 In intercept σ
0
2
977.16 (150.10)*** 519.83 (113.78)***
In linear term σ
1
2
0.50 (0.31)
Goodness of fit
Deviance

13088.1 12459.6
AIC 13094.1 12469.6
BIC 13101.5 12482.0
ICC = 0.81 (81% of variation is attributable to differences between participants, 19% to differences within participant)

*p<0.05; **p<0.01; ***p<0.0001

Table 3-5. Results from preliminary analyses of ICC and growth trajectory of hunger
during HSLF meal condition

Parameter Model A (no change) Model B (linear change)
Fixed effects
Composite model Intercept γ00
35.35
(3.15)***
42.77 (2.76)***
Time (linear term) γ10 -0.74 (0.16)***
Variance components
Level 1 Within-person σ ε
2
241.01 (8.98)*** 261.50 (21.95)***
Level 2 In initial status σ0
2
851.19 (131.11)*** 492.21 (104.81)***
In rate of change (time) σ1
2
0.98 (0.37)**
Goodness of fit
Deviance 13069.1 12350.3
AIC 13075.1 12360.3
BIC 13082.6 12372.7
ICC = 0.78 (78% of variation attributable to between-person differences, 22% to within-person differences)

*p<0.05; **p<0.01; ***p<0.0001
The analytic sample for the analyses testing the effect of negative mood on hunger for
low versus high habitual sugar consumers was N = 67 for both meal conditions. A total of 20
80

participants were excluded from the analyses due to missing or unusable 24 hour dietary recall
data. Of the 20 participants who were excluded due to 24 hour recall data, 6 participants did not
have 24 hour dietary recall data, 9 participants had only 1 day of dietary recall (valid recall data
had to include at least 2 days of dietary recall data), 4 participants had greater than 1 month
between each 24 hour dietary recall data collection (valid recall data had to have less than 1
month between each recall date), and 1 participant was excluded due to implausible 24 hour
dietary recall energy intake. In order to screen for implausible energy intake, standardized
residuals were obtained from a regression of mean kilocalories on BMI. A cut-off of ± 3 SD was
used to determine implausible energy intake. The participant who was excluded was excluded
due to implausibly low (greater than 3 SD below the mean) reported energy intake based on the
cut-off.
Results
Sample characteristics for negative mood at each time point, hunger at each time point,
and habitual total sugar intake are displayed in Table 3-6.

81

Table 3-6. Average negative mood scores, hunger scores, and grams of habitual total sugar
intake
Variable Mean (SD) [range]
Habitual total sugar intake (grams) 106.8 (50.3) [14.9-258.7]
Tertile 1 (“low” intake) (grams) [14.9-82.6]
Tertile 2 (“medium” intake) (grams) [87.6-122.3]
Tertile 3 (“high” intake) (grams) [123.0-258.7]
Negative mood
a
LSHF condition

HSLF condition

Time 0 (fasting, pre-breakfast) 7.3 (13.1) [0-75.3] 13.1 (16.1) [0-62.5]
Time 30 6.8 (13.2) [0-76.5] 10.4 (13.3) [0-52.0]
Time 60 6.3 (13.6) [0-79.8] 9.2 (9.9) [0-37.8]
Time 90 7.8 (17.1) [0-80.5] 7.7 (7.8) [0-27.3]
Time 120 9.5 (19.9) [0-87.8] 7.1 (8.2) [0-26.0]
Time 150 7.6 (17.4) [0-81.5] 6.5 (7.7) [0-27.3]
Time 180 7.2 (15.5) [0-81.8] 7.2 (9.6) [0-41.5]
Time 210 6.9 (15.0) [0-83] 5.8 (7.4) [0-27.8]
Time 240 (lunch meal) 8.3 (17.5) [0-81.3] 5.9 (6.6) [0-26.0]
Time 270 7.6 (13.6) [0-63.5] 6.5 (8.4) [0-30.5]
Time 300 (snack tray introduced) 6.5 (14.5) [0-72.5] 6.3 (8.7) [0-33.0]
Time 330 6.0 (12.3) [0-68.8] 5.7 (6.7) [0-24.5]
Time 360 5.0 (9.2) [0-52.0] 4.8 (5.6) [0-23.5]
Time 390 6.5 (12.8) [0-67.5] 5.2 (9.7) [0-55.3]
Time 420 5.5 (8.2) [0-33.8] 4.8 (7.5) [0-40.5]
Time 450 5.4 (9.5) [0-50.8] 5.6 (8.6) [0-40.3]
Time 480 6.8 (12.9) 4.5 (7.5)
Hunger
a

Time 0 (fasting, pre-breakfast) 50.0 (23.8) [1-100] 51.7 (23.0) [2 -99 ]
Time 30 35.9 (36.2) [0-100] 32.9 (33.1) [0-100]
Time 60 34.5 (36.3) [0-100] 33.1 (32.8) [0-100]
Time 90 34.4 (36.5) [0-100] 33.3 (33.7) [0-100]
Time 120 36.1 (34.9) [0-100] 35.2 (32.6) [0-100]
Time 150 37.3 (33.0) [0-100] 36.1 (30.9) [0-100]
Time 180 40.9 (32.2) [0-100] 38.5 (29.8) [0-100]
Time 210 44.3 (31.0) [0-100] 40.0 (27.9) [0-100]
Time 240 (lunch meal) 43.5 (28.6) [0-100] 43.3 (27.1) [0-100]
Time 270 34.1 (35.8) [0-100] 44.6 (26.9) [0-99]
Time 300 (snack tray introduced) 35.2 (33.9) [0-100] 34.1 (35.5) [0-100]
Time 330 33.6 (32.0) [0-100] 36.0 (33.9) [0-100]
Time 360 32.0 (36.4) [0-100] 34.3 (33.6) [0-100]
Time 300 31.1 (37.2) [0-100] 30.2 (34.0) [0-100]
Time 420 28.5 (37.6) [0-100] 30.1 (36.2) [0-100]
Time 450 28.4 (37.5) [0-100] 29.3 (37.7) [0-100]
Time 480 31.4 (38.4) [0-100] 28.3 (37.4) [0-100]
SD: standard deviation; HSLF: high sugar/low fiber meal condition; LSHF: low sugar/high fiber meal condition
82

a: time points reported as number of minutes after the condition-specific breakfast meal (e.g.: time 30 was the
measurement time point 30 minutes after the breakfast meal)

Results from the mixed models testing hypotheses 1a and 1b are presented in Tables 3-7
and 3-8. The results show that fasting hunger ratings were significantly greater than 0 during
both meal conditions (HSLF: b=36.51, SE=7.32, p<0.0001; LSHF: b=41.01, SE=7.67,
p<0.0001), and hunger decreased over the course of the day in both meal conditions (HSLF: b =
-0.76, SE=0.32, p<0.05; LSHF: b = - 0.99, SE = 0.29, p<0.01). In partial accordance with
hypothesis 1a, higher within-person negative mood was associated with greater fasting hunger
during both the HSLF and LSHF meal conditions (HSLF b=0.50, SE=0.21, p<0.05; LSHF
b=0.58, SE=0.25, p<0.05). This indicates that a 1 point increase in a participant’s negative mood
above his or her average negative mood rating for that day was associated with a 0.5 and 0.6
point increase, respectively, in fasting perceived hunger ratings during the HSLF and LSHF
conditions. However, there was no significant association between between-person negative
mood and initial hunger ratings. This indicates that variance in the hunger outcome at time 0
(fasting) was explained by within-person differences in negative mood, but not by between-
person differences. Combined, these negative mood findings indicate that how participants
differed from their average negative mood rating during the day was associated with initial
hunger ratings, but how participants differed from each other in their negative mood was not
associated with initial hunger ratings.
The lack of a significant interaction between negative mood and habitual total sugar
intake indicates that the association between negative mood and fasting hunger was not
moderated by level of habitual total sugar intake in either meal condition. However, findings did
indicate that there was a significant main effect of habitual total sugar intake on hunger at the
beginning of the day during the LSHF meal condition. Participants in the lowest tertile of
83

habitual total sugar intake reported lower fasting hunger ratings (b= -17.40, SE=7.99, p<0.05)
compared to participants in the highest tertile of intake.
In partial accordance with the hypothesis 1b, there was a significant association between
within-person negative mood and time in the LSHF condition (negative mood
within-person
* Time
b=-0.07, SE=0.03, p<0.05). This indicates that variation in negative mood within participants
was associated with change in hunger ratings over the course of the day in the LSHF condition.
However, the relationship between negative mood and change in hunger was not in the
hypothesized direction and was not moderated by habitual total sugar intake. We hypothesized
that greater negative mood would be associated with an increase in hunger during the LSHF
condition for participants with higher habitual sugar intake, but the results indicate that greater
negative mood was associated with a more pronounced (or faster) decline in hunger throughout
the day. The lack of a significant negative mood by time by habitual total sugar intake interaction
(negative mood * time * habitual total sugar intake) indicates that contrary to the hypothesis, the
relationship between negative mood and change in hunger throughout the day in the LSHF
condition was not moderated by habitual total sugar intake.

84

Table 3-7. Results from mixed model HSLF condition
a
Parameter Estimate (SE)
Intercept 36.51 (7.32)***
Negative mood
between-person
-0.71 (0.82)
Negative mood
within-person
0.50 (0.21)*
Habitual total sugar intake (low) -13.29 (7.59)
Habitual total sugar intake (medium) -0.21 (7.58)
Time -0.76 (0.32)*
Negative mood
between-person
* Time -0.01 (0.05)
Negative mood
within-person
* Time -0.02 (0.02)
Habitual total sugar intake (low) * Time 0.41 (0.46)
Habitual total sugar intake (medium)* Time -0.22 (0.47)
Negative mood
between-person
* Habitual total sugar intake (low) 1.02 (0.97)
Negative mood
between-person
* Habitual total sugar intake
(medium)
1.51 (1.1)
Negative mood
within-person
* Habitual total sugar intake (low) 0.21 (0.31)
Negative mood
within-person
* Habitual total sugar intake
(medium)
-0.02 (0.36)
Negative mood
between-person
* Time * Habitual total sugar
intake (low)
0.05 (0.06)
Negative mood
between-person
* Time * Habitual total sugar
intake (medium)
-0.02 (0.07)
Negative mood
within-person
* Time * Habitual total sugar
intake (low)
-0.03 (0.03)
Negative mood
within-person
* Time * Habitual total sugar
intake (medium)
-0.03 (0.04)
a: controlled for a priori covariates: ethnicity, randomization order, sex, BMI z-score
HSLF: high sugar/low fiber meal condition; SE: standard error;
***p<0.0001, **p<0.01, *p<0.05

85

Table 3-8. Results from mixed model LSHF condition
a
Parameter Estimate (SE)
Intercept 41.01 (7.67)***
Negative mood
between-person
0.16 (0.80)
Negative mood
within-person
0.58 (0.25)*
Habitual total sugar intake (low) -17.40 (7.99)*
Habitual total sugar intake (medium) -6.61 (7.90)
Time -0.99 (0.29)**
Negative mood
between-person
* Time -0.05 (0.04)
Negative mood
within-person
* Time -0.07 (0.03)*
Habitual total sugar intake (low) * Time 0.45 (0.40)
Habitual total sugar intake (medium)* Time 0.23 (0.41)
Negative mood
between-person
* Habitual total sugar intake (low) 0.54 (0.90)
Negative mood
between-person
* Habitual total sugar intake
(medium)
0.48 (1.09)
Negative mood
within-person
* Habitual total sugar intake (low) -0.59 (0.43)
Negative mood
within-person
* Habitual total sugar intake
(medium)
-0.54 (0.41)
Negative mood
between-person
* Time * Habitual total sugar
intake (low)
0.06 (0.05)
Negative mood
between-person
* Time * Habitual total sugar
intake (medium)
0.08 (0.05)
Negative mood
within-person
* Time * Habitual total sugar
intake (low)
0.06 (0.04)
Negative mood
within-person
* Time * Habitual total sugar
intake (medium)
0.05 (0.05)
a: controlled for a priori covariates: ethnicity, randomization order, sex, BMI z-score
HSLF: high sugar/low fiber meal condition; SE: standard error
***p<0.0001, **p<0.01, *p<0.05

Discussion
Findings from this experimental cross-over design feeding study in overweight
adolescents demonstrate that negative mood was associated with hunger and this association
varied when participants were fasting versus when they were exposed to test meals differing in
sugar and fiber content. As expected, higher negative mood was associated with higher fasting
hunger at the beginning of both the HSLF and LSHF meal condition days. The positive
association between negative mood and fasting hunger ratings was not expected to differ
between meal conditions because participants reported these ratings at time 0, prior to being
86

exposed to the test meals. Our finding is supported by prior research examining the effects of
mood on appetite. In a mood manipulation study of normal weight women who were exposed to
an experimental condition that induced negative mood or a condition that induced neutral mood,
participants in the negative mood condition reported an increase in appetite after being exposed
to the negative mood manipulation (Hepworth, Mogg, Brignell, & Bradley, 2010). The fact that
negative mood was induced suggests that there was a causal relationship between experiencing
negative mood and experiencing increased appetite. This lends support to our directional
hypothesis and finding that suggests that greater negative mood precipitated greater fasting
hunger at the beginning of both meal condition days. Findings from the current study extend
evidence from previous research because, to our knowledge, this is the first study to demonstrate
that experiencing negative mood prior to breakfast may cause overweight adolescents to
experience greater perceived hunger.
In partial accordance with the hypotheses, within-person differences in negative mood
were also associated with the trajectory (i.e. change over time) of participant hunger ratings
throughout the day during the LSHF condition and not during the HSLF condition. We
hypothesized that exposure to the LSHF condition would lead to higher negative mood ratings,
which would in turn be associated with an increase in hunger ratings. However, we found that
hunger decreased throughout the day during both the HSLF and LSHF conditions, and that an
increase in an individual’s negative mood above his or her average negative mood was
associated with a more pronounced decrease in hunger ratings in the LSHF condition. While this
finding was unexpected, there is evidence to support that negative mood can also lead to reduced
appetite. For example, while individuals who have restrained eating styles (characterized by
dieting and restricting food intake) tend to have increased appetite in response to negative
87

affective states, individuals with non-restrained eating styles tend to have reduced appetite in
response to experiencing negative affective states (Macht, 2008). Given the fact that the majority
of participants in this sample were obese, it is unlikely that they engaged in eating behaviors
typical of a restrained eating style, so they were likely non-restrained eaters. An experimental
study that examined the effect of induced negative mood on eating behaviors in women with
binge eating disorder who were either depressed or not depressed found that those who were not
depressed ate less after a negative mood manipulation compared to those who were depressed
(Dingemans, Martijn, van Furth, & Jansen, 2009). Combined, this evidence suggests that
negative mood may have a suppressant effect on appetite in individuals with certain
characteristics, under certain circumstances.
Findings from the current study may indicate that not only individual characteristics, such
as eating style and psychological conditions, but also food environment could influence the
effect of mood on hunger. Given that we only observed an association between within-person
negative mood and change in hunger in the LSHF condition, it is possible that consumption of
foods that were low in sugar and high in fiber influenced the association between negative mood
and perceived hunger in this sample. Consumption of foods high in fiber has been shown to lead
to greater and more prolonged reductions in hunger compared to foods high in other types of
carbohydrates (Burton-Freeman, 2000; Pasman, Saris, Wauters, & Westerterp-Plantenga, 1997;
Slavin, 2005), and as previously mentioned, experiencing negative mood appears to reduce
appetite in some individuals. Given the effects that consuming fiber and experiencing negative
mood can independently have on hunger, it is possible that our findings were driven by a
combined effect of negative mood and fiber intake on the trajectory of perceived hunger in the
LSHF meal condition. To our knowledge, this is the first study to show that increased negative
88

mood can be associated with more pronounced declines in hunger in overweight adolescents
exposed to an experimental LSHF food environment. However, it is not clear what mechanisms
would have driven a combined effect of negative mood and LSHF test meal consumption on
hunger trajectories. Future studies that manipulate both mood and carbohydrate quality of test
meals would be necessary to elucidate the mechanisms of this unexpected finding.
Contrary to the hypotheses, the associations between negative mood and fasting hunger
and between negative mood and change in hunger over time were not moderated by how much
total sugar participants reported habitually consuming in daily life. However, habitual sugar
intake did have a seemingly independent effect on fasting hunger ratings at the beginning of the
LSHF meal condition day. Compared to individuals in the lowest habitual total sugar intake
tertile (those who reported consuming an average of approximately 15 to 83 grams of total sugar)
per day, those in the highest tertile (those who reported consuming an average of approximately
123 to 259 grams of total sugar per day) reported higher hunger ratings prior to breakfast in the
LSHF meal condition. This finding is supported by evidence that other characteristics of habitual
dietary intake, such as dietary fat content, can influence appetite measured in experimental
studies. Studies examining hunger and satiety in habitual consumers of low-fat versus high-fat
diets have demonstrated that habitual high-fat consumers have higher fasting hunger compared to
habitual low-fat consumers (Blundell et al., 2005; Cooling & Blundell, 1998; French, Murray,
Rumsey, Fadzlin, & Read, 1995). Surprisingly, the association between higher habitual total
sugar intake and greater fasting hunger was not observed prior to breakfast during the HSLF
meal condition. Although it is unlikely, one potential explanation for the fact that we only
observed this association prior to the LSHF breakfast is that participants may have inadvertently
been cued to the fact that they were going to be restricted from consuming sugar-rich foods for
89

the test meal. Anticipation of a LSHF breakfast could have caused participants in the highest
sugar intake tertile to experience sugar cravings. Evidence suggests that habitual consumers of
high-sugar diets tend to experience sugar cravings when deprived of sugar intake (Reid &
Hammersley, 1998). Although cravings are distinct from hunger, cravings can be associated with
and possibly conflated with perceived hunger (Cepeda-Benito, Gleaves, Williams, & Erath,
2001; A. J. Hill, C. F. Weaver, & J. E. Blundell, 1991; Martin, O'neil, & Pawlow, 2006; White,
Whisenhunt, Williamson, Greenway, & Netemeyer, 2002), so habitual high sugar consumers
may have misconstrued experiencing cravings with feeling hungry. Although this is speculation,
it is a potential explanation that provides a possible line of inquiry for future work to identify the
mechanism for our observation.
Findings from the current study provide important insights that extend previous research
and have potential implications for future research. Much of the previous research examining the
impact of mood on hunger has been conducted predominantly in samples of adult women,
particularly women with binge eating disorder (Chua, Touyz, & Hill, 2004; Dingemans et al.,
2009; Munsch, Michael, Biedert, Meyer, & Margraf, 2008; Telch & Agras, 1996). Additionally,
there has been lack of research examining how habitual intake of all types of sugar (i.e. total
dietary sugar) impacts hunger, especially in the context of simulated food environments that
reflect real-world dietary patterns, such as the HSLF and LSHF experimental conditions in the
current study. It is important to investigate how mood and habitual dietary intake impact hunger
in overweight adolescents in these types of simulated food environments because this
demographic is in need of effective dietary interventions to promote weight loss and sustainable
healthful eating behavior changes. Results from these types of studies have the potential to
provide insights that improve the outcomes of dietary interventions. Our finding that higher
90

negative mood was associated with higher hunger in the fasting state could indicate that
experiencing negative mood in a fasting state may be a trigger for overconsumption or poor food
choice by overweight adolescents. The finding that higher negative mood was associated with a
decrease in hunger in the LSHF meal condition may suggest that a fiber-rich diet that is low in
sugar could be protective against negative mood causing greater hunger. Our observation that
higher habitual sugar consumption was associated with greater hunger prior to the LSHF
breakfast meal may have implications for dietary interventions in overweight adolescents.
Dietary interventions for this population aimed at increasing fiber intake and decreasing sugar
intake may need to include components that address the impact of high habitual sugar intake on
perceived hunger. Successful efforts are needed to shift the quality of carbohydrates consumed
by overweight adolescents toward high-fiber foods and away from high-sugar foods. Insights
from the current study indicate that future research focusing on negative mood, habitual sugar
intake, and hunger in overweight adolescents could provide important direction for dietary
interventions in this population.
This study has potential limitations that should be noted. While the laboratory setting of
the study provided advantages such as a controlled feeding and observation environment, the
results from the study may not be generalizable outside of the laboratory setting. For example,
the negative mood reported by participants during the laboratory visits may not correspond to
what they would have normally experienced, as they were isolated from factors that may
influence negative mood in the free-living setting, such as interactions with family members and
peers or events that occur at school. Participants were aware that they were being observed
which could also have impacted their mood. Additionally, participants were fed at scheduled
times, so their hunger may not have developed as it usually does throughout the course of the day
91

in a free-living setting. It is possible that the relationships found between negative mood and
hunger could be influenced by a third variable that was not examined. For example, recent
research in animal models has shown that levels of ghrelin, a gut peptide that is involved in
physiological hunger signaling, may mediate the relationship between stress and hunger in rats
(Schellekens et al., 2012). It is possible that levels of ghrelin may mediate the relationships
between negative mood and hunger in humans. However, this relationship could not be tested in
the current sample because while ghrelin data was collected in this study, it was only available
for 23 of the participants. The complexity of the growth curve model used in this analysis and the
relatively small sample size with available ghrelin data would preclude conducting such an
analysis in this sample. Finally, a single in-lab visit for each meal condition may not be sufficient
to reveal how meals differing in sugar and fiber content can influence the relationship between
negative mood and hunger in a free-living context. Longer studies in free-living settings may be
necessary to understand how all of these factors interrelate in real life.

Conclusions
The findings from this study indicate that negative mood and habitual sugar intake may
be important contributors to greater perceived hunger when individuals are in a fasting state,
which could have implications for the amount of food overweight adolescents consume at
breakfast and for their breakfast food choices. However, our findings did not demonstrate that
negative mood precipitates greater hunger in overweight adolescents when they are restricted
from sugar intake. This suggests that negative mood may not be an important determinant of
whether overweight adolescents are able to sustain dietary patterns that are low in sugar and high
in fiber in the midst of a sugar-rich food environment. Thus, while we postulated that sugar
restriction-dependent negative mood may play a role in poor quality carbohydrate diets in this
92

population, that may not be the case. Further research is necessary to determine other individual-
level factors that may perpetuate high sugar, low fiber dietary patterns in this population.

93

Chapter 4: Self-tracking of weight-related behaviors: a promising intervention approach
for sustaining weight-related behavior change
Introduction
Few interventions have been successful in helping people to adopt low-sugar, high fiber
diets and maintain those dietary changes over the long-term. Self-tracking, which consists of
paying deliberate attention to an aspect of one’s behavior and recording details about that
behavior (Lora E. Burke, Jing Wang, & Mary Ann Sevick, 2011), may represent a viable
approach to the challenge of helping people make and maintain dietary changes. Self-tracking
enhances cognitive processes that aid in dietary behavior change, such as one’s awareness of
eating behaviors and foods consumed (Burke et al., 2005; Yasmin Mossavar-Rahmani et al.,
2004). According to theories of self-regulation, awareness of and attention to behaviors are
necessary for one to be able to engage in behavior change and achieve behavior change goals
(Baker & Kirschenbaum, 1998; Baumeister, Heatherton, & Tice, 1994; Lora E. Burke et al.,
2011; Carver & Scheier, 1990). This suggests that awareness of and attention to weight-related
behaviors promoted by self-tracking may support behavior change. Self-tracking is also thought
to support behavior change by aiding recall of goals and intentions (Latner & Wilson, 2002), and
has been shown to increase self-efficacy for dietary behavior change (Yasmin Mossavar-
Rahmani et al., 2004). Evidence from intervention studies demonstrates that self-tracking of
weight-related behaviors results in greater weight loss and success in weight loss maintenance
(L. E. Burke et al., 2011), as well as adherence to weight-related behavior change (Lora E. Burke
et al., 2011; Burke et al., 2005; Yasmin Mossavar-Rahmani et al., 2004).
The utility of self-tracking for aiding dietary behavior change is supported by the
ubiquity of tools such as smartphone apps and on-body sensors that can aid in self-tracking of
94

health-related behaviors (O’Reilly & Spruijt-Metz, 2013). According to research by the Pew
Internet & American Life Project, 64% of U.S. adults own a smartphone smartphones (Smith,
2015). Self-tracking using smartphones can be done using a variety of health apps. A recent
study of health application use in a representative sample of US adults found that 58% of
smartphone users had downloaded at least one health application, with applications for nutrition
and fitness being the most prevalent. Estimates for ownership of on-body devices that are used
for self-tracking health behaviors and related metrics (such as FitBit or Jawbone UP devices) are
not available; however, the pervasiveness of smartphone ownership and app usage demonstrates
that self- tracking using devices could be a feasible intervention approach to support dietary
behavior change.
Although self-tracking has been shown to be a cornerstone of weight-related behavior
change, especially for dietary behavior (Lora E. Burke et al., 2011), evidence from observational
studies and interventions indicates that adherence to self-tracking tends to be low and decline
over time (Burke et al., 2012; Krebs & Duncan, 2015; Y. Mossavar-Rahmani et al., 2004; Tate et
al., 2001; Yon et al., 2007). A recent study found that approximately 46% of mobile health
application users have stopped using an application they formerly used due to loss of interest or
high burden of data entry (Krebs & Duncan, 2015). Given that self-tracking of weight-related
behaviors is such an important component of weight-related behavior change, it is important to
understand how to help individuals seeking to lose weight, maintain weight, or maintain weight-
related behavior changes adhere to long-term self-tracking. Little is known about how to improve
adherence to long-term self-tracking. While some research has focused on understanding barriers
to self-tracking, no previous research has focused on understanding why and how those who do
engage in long-term self-tracking are able to adhere to this practice. Individuals who use self-
95

tracking to aid in behavior change, such as those who are part of online fitness communities and
the “Quantified Self” (QS) movement, represent untapped sources of information about
volitional, long-term self- self-tracking. The QS movement is a community of individuals who
use a variety of methods to self- self-track personal metrics, often health-related metrics
(http://quantifiedself.com/about/). Many individuals who participate in online fitness
communities and the QS movement apply the data they collect from self-tracking to better
understand aspects of their health and wellness.
Examining what motivates individuals within these communities to volitionally self-track
their health behaviors and metrics may reveal common attributes among them that could
potentially be targeted in other populations, such as overweight minority adolescents, to improve
self-tracking adherence in future studies. For example, since self-tracking requires deliberately
paying attention to one’s behavior, it is possible that those who volitionally self-track tend to
have a high level of dispositional mindfulness, and thus an enhanced attention to and awareness
of the present moment. While mindfulness is considered to be an inherent quality, it can be
cultivated through systematic training (Daubenmier et al., 2011). Mindfulness has been
successfully targeted in many interventions to improve a number of different health outcomes as
well as to enhance self-regulation (Baer, 2003; Dalen et al., 2010; Kristeller & Wolever, 2010;
Shapiro, Carlson, Astin, & Freedman, 2006). If mindfulness is indeed an attribute of those who
volitionally self-track their weight-related behaviors, this could be a promising target of change
for future interventions that aim to help adolescents improve their self-tracking behaviors
through enhancing self-regulation. Additionally, individuals who volitionally self-track may
have a high internal health locus of control. Health locus of control describes the extent to which
one believes one’s health is controlled by one’s own actions (defining an internal locus of
96

control) or by external factors, such as chance or powerful others like medical professionals
(defining an external locus of control) (B. S. Wallston, Wallston, Kaplan, & Maides, 1976). An
internal locus of control has been related to positive health behaviors and health outcomes,
including adherence to medical regimens (Lewis, Morisky, & Flynn, 1978) and weight loss
(Balch & Ross, 1975). It may be informative for future interventions to determine if individuals
who successfully adhere to self-tracking weight-related behaviors have high internal locus of
control.
Self Determination Theory (SDT) provides a context for examining other potentially
modifiable attributes that that may be common among individuals who volitionally self-track.
SDT is a theory of motivation that is concerned with explaining intrinsic tendencies to behave in
effective and healthful ways (Deci & Ryan, 2011). According to SDT, intrinsic motivation and
competence to engage in a behavior are necessary for the behavior to occur and persist (G. C.
Williams, Grow, Freedman, Ryan, & Deci, 1996). Given that individuals in online fitness
communities and the QS movement self-track health-related metrics and behaviors of their own
volition, it is possible that these individuals are intrinsically motivated and have high competence
for self-tracking and behavior change. There is evidence that motivation orientation can
potentially be modified toward intrinsic (i.e. behavior change supportive) motivation (Friedman,
Deci, Elliot, Moller, & Aarts, 2010; Isen & Reeve, 2005; Laran & Janiszewski, 2011), and that
perceived competence to engage in a behavior can be improved (Silva et al., 2010). Thus,
motivation orientation and competence may be promising targets for future interventions that
aim to improve self-tracking adherence.
To our knowledge, to date no formal research has been done with fitness and QS
communities to understand 1) what attributes are common among those who volitionally self-
97

track weight-related behaviors, 2) what motivates those who self-track weight-related behaviors
to begin self-tracking, 3) what motivates them to adhere to self-tracking, and 4) how self-
tracking helps them attain personal health- and weight-related goals. The objective of the
proposed study is to use a mixed qualitative/quantitative methods approach to understand
characteristics that may contribute to developing a practice of self-tracking weight-related
behaviors in a group of participants who self-track, and to compare some of these characteristics
to participants who do not self-track .
Specific Aims and Hypotheses
This study addresses the following specific aims:

Aim 1: To examine attributes that may be common among individuals who self-track weight-
related behaviors by comparing them to individuals who do not self-track weight-related
behaviors on the following domains: health locus of control, dispositional mindfulness levels,
motivation orientation for healthful diet and physical activity behaviors, and competence for
making changes toward healthful behaviors. This aim was achieved through questionnaires.
Hypothesis 1a: Compared to people who do not self-track weight-related behaviors,
individuals who do self-track weight-related behaviors will have a greater tendency
toward an internal health locus of control.
Hypothesis 1b: Compared to people who do not self-track, people who do will have
higher dispositional mindfulness.
Hypothesis 1c: Compared to people who do not self-track, people who do will have a
greater orientation toward autonomous motivation for engaging in healthful dietary and
physical activity behaviors.
98

Hypothesis 1d: Compared to people who do not self-track, people who do will have
higher perceived competence for making changes toward healthful behaviors.
Aim2: To investigate factors that may contribute to initiation and adherence to self-tracking of
weight-related behaviors, as well as how self-tracking aids health-related goal attainment. This
was achieved through in-depth qualitative interviews.
Hypothesis 2: The in-depth interviews functioned to elicit descriptive information about
what motivates individuals to begin to self-track weight-related behaviors and factors that
help them to adhere to self-tracking.

Methods
Study Design
This project was an exploratory descriptive study meant as a hypothesis generating
exercise for future studies. In order to accomplish the specific aims, this study used a mixed
methods design consisting of a qualitative component and a quantitative component. The study
sample included a self-tracking group and a comparison group of individuals who do not self-
track. Participants in both the self-tracking and comparison groups completed the quantitative
component. A subsample of participants in the self-tracking group also completed the qualitative
component. Details of the study methods are presented below:
Quantitative component. For the quantitative component of the study, participants
completed written questionnaires that were administered via a secure online Qualtrics form. The
purpose of the quantitative study component was to examine characteristics potentially common
among individuals who self-track their weight-related behaviors and compare these
characteristics to those of individuals who do not self-track. The outcome variables for the
quantitative component of the study included: health locus of control, mindfulness, and Self-
99

Determination Theory constructs including motivation orientation for engaging in healthful
dietary and physical activity behaviors and competence for making changes toward healthful
dietary and physical activity behaviors. The measures that were used are detailed in the
Measurements section and the questionnaire is included in Appendix B.
Qualitative component. For the qualitative component of the study, a subsample of self-
tracking participants completed one-on-one in-depth interviews. The in-depth interviews used a
semi-structured format guided by an interview protocol with pre-defined key questions that was
developed based on standard in-depth interview protocol guidelines (Berg & Lune, 2004; Boyce
& Neale, 2006). The in-depth interviews consisted of open-ended questions and used a
discovery-oriented approach to elicit information from participants about why they self-track
weight-related behaviors, what types of methods or tools they use for self-tracking, what
motivates their adherence to self-tracking, and how self-tracking of weight-related behaviors
supports their personal health goals. The interview protocol is included in Appendix C.
Interviews were conducted in English and were tape-recorded with participant permission. After
interviews were completed, they were transcribed verbatim.
Informed consent. This study was approved by the University of Southern California
Institutional Review Board as an exempt study that did not require written informed consent.
Prior to participating in study procedures, participants were provided with a research information
sheet and provided with answers to questions about the study, but were not required to provide
written informed consent. Participants were provided with compensation for their time ($15.00
for participants who completed the interview and questionnaire, $5.00 for participants who
completed the questionnaire only).
100

Study Sample
Participant recruitment and screening. A total of 116 participants were recruited for
this study. Twenty-five self-tracking participants were recruited for the qualitative component
from fitness communities via the website Reddit (n= 19) and the Quantified Self community
through Quantified Self meetup groups and the Quantified Self online forums (n = 6) (these 25
participants also completed the quantitative component of the study). A sample size of 25
interview participants was chosen based on sample sizes in previous qualitative studies on self-
tracking of health-related metrics and behaviors including blood pressure (n=23) (Jones et al.,
2012), blood glucose (n=13) (Hortensius et al., 2012), and dietary intake (n=15) (Burke, Swigart,
Turk, Derro, & Ewing, 2009). Recruitment for self-tracking participants was conducted digitally
through online posts on Reddit fitness and health forums and in Quantified Self online forums, as
well as presentations given at Quantified Self meetups.
A power analysis was conducted to determine the sample size needed to detect
differences in the outcomes of interest between individuals who self-track and individuals who
do not self-track. The power analysis was conducted using G*Power version 3.1 (Erdfelder,
Faul, & Buchner, 1996) for one-tailed independent group t-tests with α=0.05 and power = 0.8.
To detect a moderate effect size (Cohen’s d=0.5), both the self-tracking and comparison groups
needed to include 51 participants. To reach this sample size for the self-tracking group, an
additional 38 self-tracking participants were recruited from online Reddit fitness communities
and through the USC Psychology department student listserv to complete the questionnaires. The
total sample size self-tracking participants who completed the questionnaires was n=63. For the
comparison group of individuals who do not self-track, a sample of 53 participants was recruited
through the USC Psychology department study listserv. Administrators of the USC Psychology
101

department listserv granted permission for participant recruitment. All individuals who were
interested in participating in this study were provided with information about the significance
and purpose of the study as well as a description of what the study involved. Given that the study
was granted exempt status by the University of Southern California Institutional Review Board,
written informed consent was not required for participants who completed the quantitative
component of the study.
Inclusion and exclusion criteria. Participants were eligible for inclusion in the study if
they met the following inclusion criteria. Self-tracking group: a) self-track dietary intake,
physical activity, stress levels, sleep patterns, and/or weight for at least the past 6 months, b) ≥
18 years of age. Comparison non-self-tracking group: a) University of Southern California
students who do not self-track dietary intake, physical activity, stress levels, sleep patterns,
weight, or any other health behavior or metric, b) ≥ 18 years of age.
Measure
Participant demographics. Participant demographics were collected via the quantitative
questionnaire form that participants completed via the secure web-based Qualtrics form.
Demographic data included age, sex, and ethnicity.
Quantitative domains. The quantitative component of this study examined constructs
based on Self-Determination Theory framework to examine motivation, as well as constructs
related to dispositional mindfulness and health locus of controls that are hypothesized to be
related to propensity for self-tracking.
Treatment Self-Regulation. Motivation for changing to and maintaining healthful
dietary and exercise habits was measured using the Diet and Exercise subscales of the
Treatment Self-Regulation Questionnaire (TSRQ) (G. C. Williams, Ryan, & Deci). These
102

15 item subscales assess the degree to which an individual’s motivation for engaging in
healthful dietary and physical activity behaviors is relatively autonomous versus
controlled. Autonomous motivation refers to motivation driven by intrinsic, or personal,
interests and values (Koestner, Otis, Powers, Pelletier, & Gagnon, 2008). Controlled
motivation refers to motivation that is driven by factors external to personal values, such
as reward or punishment (Koestner et al., 2008). Greater autonomous motivation has
been found to be associated with goal progress, while controlled motivation has been
found to not be associated with goal progress (Koestner et al., 2008). The TSRQ
exercise and diet subscales had good internal consistencies (exercise: Cronbach’s α
= 0.87, diet: Cronbach’s α = 0.89)
Multidimensional Health Locus of Control. Health locus of control was measured
using the Multidimensional Health Locus of Control (MHLC) scale (K. A. Wallston,
Wallston, & DeVellis, 1978). This 36-item scale measures the extent to which one
believes one’s general health status is governed by oneself (internally) or by outside
entities (externally). The internal consistencies of the MHLC subscales were acceptable
(internal: Cronbach’s α = 0.72; chance: Cronbach’s α = 0.71; powerful others:
Cronbach’s α = 0.78).
Five Facet Mindfulness Questionnaire. Mindfulness was measured using the Five Facet
Mindfulness Questionnaire (FFMQ) (Baer, Smith, Hopkins, Krietemeyer, & Toney,
2006). This 39-item scale assesses five facets of general tendency to be mindful in daily
life (observing, describing, acting with awareness, nonreactivity to inner experience, and
nonjudging of inner experience). The five facets of the FFMQ had acceptable to good
internal consistency (observe: Cronbach’s α = 0.72; describe: Cronbach’s α = 0.85; acting
103

with awareness Cronbach’s α = 0.83; nonreactivity to inner experience: Cronbach’s α
= 0.79; nonjudging of inner experience: Cronbach’s α = 0.87).
Perceived Competence for Diet. Feelings about engaging in healthful dietary intake
were measured using the Perceived Competence Scale (PCS) for Diet
(G. C. Williams, Ryan, & Deci). This 4-item questionnaire assesses the degree to which
an individual feels confident about being able to make or maintain a change toward
healthful dietary behaviors. The PCS for diet had excellent internal consistency
(Cronbach’s α = 0.95).
Perceived Competence for Exercise. Feelings about engaging in exercise were
measured using the Perceived Competence Scale (PCS) for Exercise (G. C. Williams et
al.). This 4-item questionnaire assesses the degree to which an individual feels confident about
being able to make or maintain a change toward healthful exercise behaviors. The PCS for
exercise had excellent internal consistency (Cronbach’s α = 0.97).
Qualitative domains. In the in-depth interviews, participants were asked what weight-
related behaviors they track, why they self-track (what prompted them to initiate self-tracking),
what factors have contributed to long-term self-tracking adherence, what methods or tools they
use to self-track, what insights they have gained through self-tracking, and how self-tracking has
helped them with health-related goal attainment.
Data Analysis
Quantitative data analysis. The first step in analyzing the data from the quantitative
component of the study was to conduct descriptive exploratory analyses to check for outliers.
Given that this study has a relatively small sample size and is not powered for complex analyses,
the quantitative data was used to calculate summary statistics with the purpose of describing
104

characteristics of individuals who self-track weight-related behaviors versus those who do not
self-track weight-related behaviors. Independent sample t-tests were used to compare the mean
outcome variable scores of individuals who self-track versus individuals who do not self-track, in
order to compare attributes between these two groups.
Qualitative data analysis. Interviews audio files were transcribed by G.O. and trained
undergraduate research assistants. Coding and analysis of the interview transcriptions were
conducted using the qualitative analysis software ATLAS.ti by G.O. Interview data were
analyzed using a directed content analysis approach (Hsieh & Shannon, 2005). Directed content
analysis is a deductive analytic method that uses theory or relevant research findings to guide
development of a set of a priori codes (Hsieh & Shannon, 2005). These a priori codes are used to
analyze data with a structured process (Hsieh & Shannon, 2005). In the current study, a coding
scheme was created using concepts of motivation and perceived competence for behavior change
from Self-Determination Theory, as well as questions from the in-depth interview protocol that
related to facilitators of and barriers to behavioral adherence. During the transcription coding
process, the initial set of a priori codes was applied to the interview transcriptions, and the code
list was updated and edited based on common themes that emerged from the interview
conversations. The first step in analyzing data from the interview transcripts was a review of the
interview transcriptions. Interview data analysis then proceeded with the use of line-by-line
coding using the set of a priori codes. Transcript coding was completed in an iterative process to
identify and refine emerging conceptual categories in the data (Bradley, Curry, & Devers, 2007;
Corbin & Strauss, 2014). This process was completed until each transcript was analyzed using
the complete final set of a priori and inductive codes. The list of a priori and inductive codes
used for transcript analysis are included in Appendix D.
105

Results
Sample Characteristics
Sample characteristics for self-tracking and comparison group participants are presented
in Table 4-1.
Table 4-1. Sample characteristics
a
Characteristic
Self-trackers
(qualitative + quantitative)
(n=25)
b

Self-trackers
(quantitative only)
(n= 38)
b

Non-trackers (n=53)
b

Age

25.8 (8.6) [19-49] 20.0 (1.9) [18-29] 20.2 (1.8) [18-27]
Sex (female) 68% 65.8 58.5%
Hispanic/Latino 0 7.9% 5.7%
African American 0 10.5% 11.3%
White 78.3% 47.4% 39.6%
Asian 8.7% 26.3% 24.5%
Multiethnic 4.3% 5.3% 15.1%
Other 8.7% 2.6% 3.8%
Behaviors that were
tracked
(self-trackers only)

Physical activity 100% 71.1% -
Diet 88% 68.4% -
Sleep 52% 34.2% -
Stress 20% 7.9% -
Weight 60% 39.5% -
Multiple 96% 71.1% -
a: characteristics for the three samples included in this study: self-trackers who completed the qualitative and quantitative
components, self-trackers who completed the quantitative-only component, and participants who do not self-track
b: percentage, except for age which is reported as mean (SD) [minimum-maximum]

Quantitative Results: Differences Between Self-Trackers and Non-Trackers
Results from the independent-sample t-tests showed that the self-tracking and comparison
(non-tracking) groups differed on some attributes but not others. There were no differences
between the self-tracking group and the non-tracking group on health locus of control (HLOC)
(HLOC internal: t = 1.28, p = 0.20; HLOC chance: t = -0.79, p = 0.43; HLOC powerful others: t
= 0.41, p = 0.68). There were no differences between the two groups on mindfulness facets
106

measured by the FFMQ (observing: t = 1.74, p = 0.09; describing: t = 1.45, p = 0.15; acting with
awareness: t = 1.63, p = 0.11; nonjudging of inner experience: t = -0.48, p = 0.63; non-reactivity
to inner experience: t = 0.68, p = 0.50). There was also no difference between the groups in total
mindfulness score (t = 1.48, p = 0.14). However, there were differences between the groups on
motivation orientation and perceived competence for healthful behaviors. Compared to the non-
tracking group, the self-tracking group had higher autonomous motivation for engaging in
healthful exercise habits (t = 2.5, p = 0.013) and for engaging in healthful dietary habits (t = 4.1,
p<0.0001). Compared to the non-tracking group, the self-tracking group also had a higher
perceived competence for making or maintaining a change toward healthful dietary behaviors (t
= 3.22, p = 0.002) and a higher perceived competence for making or maintaining a change
toward healthful exercise behaviors (t = 3.18, p = 0.002).
Qualitative Results
Behaviors and metrics tracked. The majority of the participants in the self-tracking
group who participated in the in-depth interviews reported tracking more than one weight-related
behavior or metric. Most commonly, participants reported concurrently tracking dietary intake,
exercise, sleep, and weight. Few participants reported simultaneously self-tracking all of the
behaviors or metrics of interest to this study (diet, exercise, sleep, stress, and weight), and only
one participant reported self-tracking a single behavior (exercise). The least tracked metric was
stress. One possible reason for this is a lack of suitable methods for tracking stress with mobile
applications, web platforms, or wearable devices. This is exemplified by one participant who
explained why he does not track stress despite being interested in doing so, “I’m actually curious
about that kind of data but I haven’t found a good streamlined way to do it yet”. All of the
participants who did report tracking stress said that they made notes of how they felt in paper in
107

notebooks. The most commonly used methods for self-tracking in this sample were mobile
applications (particularly My Fitness Pal, used by 55% of participants who track dietary intake
and/or exercise) and wearable devices (particularly the Fitbit, used by 64% of participants who
track dietary intake, exercise, and/or sleep). Several participants reported using spreadsheets for
self-tracking. Although use of paper notebooks was not common in this sample, it was a
preferred method for some, particularly those participants who followed personalized weight
lifting routines that are not often trackable using other methods. Very few participants mentioned
using tools that were only available via websites to track their behaviors or metrics (such as
TrendWeight, a website used for tracking weight trends over time), but participants who reported
using mobile applications commonly reported also using the websites that accompany the
applications. The majority of participants use more than one method, most commonly a
combination of wearable devices and mobile applications. Most participants agreed that they
tend to stick with their self-tracking methods rather than frequently trying new methods.
Motivations for self-tracking. Participants reported initiating a habit of self-tracking for
a number of reasons, the majority of which were grounded in autonomous motivations, as they
reflected personal values or interests rather than external forces. Very few individuals mentioned
reasons that were grounded in controlled motivations. Frequently cited reasons included a
general interest in understanding personal health or to aid progress toward fitness goals. The
most common reason participants reported starting self-tracking was to support weight loss
efforts. The prevalent feeling was that self-tracking is a data-driven method for aiding adherence
to weight-related behavior changes in order to achieve weight loss goals. This is demonstrated by
one participant who stated, “Initially I actually started [self-tracking] to lose weight. I actually
used to be quite a bit heavier, like another half of myself heavier. So I started just like learning
108

about nutrition and realized that there’s only so far that just being mindful can get you. So being
a little bit more scientific and relying on data other than, oh I feel fat today, but what does the
data say? So to me, it gives me an objectivity that I otherwise wouldn’t have. I really rely on
those numbers”. The way in which some participants described using self-tracking for weight
loss suggests that they turned to this approach in order to aid weight loss efforts after previous
unsuccessful weight loss attempts. One participant stated, “I have always wanted to lose weight.
It’s been pretty much a constant concern, so it’s something I’ve been doing for a while. I tried
Weight Watchers a long time ago, I’ve tried a lot of things, and they always say if you track
things, if you track what you’re eating you’ll be more mindful of it, and you’re more likely to use
naturally eat less. And so I just wanted to be more consistent about it”. While weight loss was the
impetus for self-tracking for the majority of participants (60%), it was not the concern for all
participants who started self-tracking to support weight change goals. One participant explained
that she started self-tracking to gain weight, “Seven years ago when I started, it was to put on
weight. And I didn’t have a clear understanding at the time of how much calories were in
different food, and so I wasn’t putting together meals that were like fulfilling my caloric needs
for the day, because when I just felt full I would stop eating, but it wasn’t enough. So then I
started counting calories just to make sure that I got the minimum”. While it was clear from the
discussions that the most prevalent reason for starting a practice of self-tracking was to aid in
weight loss, this anecdote highlights that self-tracking is a flexible behavior change approach that
can be tailored to support other weight-related goals.
During discussions of what motivated participants to begin self-tracking, they often
explained that their reasons for self-tracking have evolved over time. Many explained that they
were motivated to continue self-tracking to support updated goals after achieving their initial
109

goals. For example, one participant initially tracked her dietary intake in order to lose weight, but
after reaching her weight goal, she began to use self-tracking to coach herself for power lifting
and body building competitions, “I realized all the different things that I could manipulate when I
started tracking. Like oh what happens if I take my carbs up? What happens if I take my fat up?
What happens if I increase my protein this week? And then recording just that anecdotal stuff,
like I mentioned, oh how did I feel? How did I do at the gym? What was my energy like? How
did I sleep? Um that kind of helped me really learn how I could manipulate things. And um, it
kind of became a fun science experiment. And then competing was kind of born out of that. Just
kind of another challenge. What can I do? I totally treat my body like a science experiment in
terms of nutrition and tracking things”. Anecdotes such as this suggest that self-tracking can help
individuals gain important insight into how to change their behaviors, and this insight can lead to
increased perceived competence for additional goal attainment.
Factors that facilitate self-tracking adherence. When participants were asked what has
helped them develop a habit of self-tracking, one of the most common answers was the ability to
get immediate feedback that reinforces their goal-related behaviors. One participant stated, “It’s
always really reinforcing when I look at the Fitbit at the end of the workout and it tells me how
many calories I burned, or during the workout I can see how high my heart rate is. So it helps me
to like monitor my workout during, and afterward it’s like a nice bit of validation. Because it
always makes me feel good when I see wow, I burned this many calories during the workout”.
Another participant explained that immediate feedback can provide motivation to continue with
behavior modifications when it takes time for benefits from behavior change to occur,
“Obviously anything like weight loss or strength gains, those are the kind of things that are
generally slower over time, so to have these nice little immediate things [feedback] along the
110

way is definitely motivating”. This suggests that immediate feedback can support development
of a self-tracking habit because it is validates one’s efforts and provides insights into progress for
goals that tend to take time to achieve, such as weight loss.
The most commonly cited factor that helps participants adhere to a long-term habit of
self-tracking is the ability to reflect on past data and see goal-related progress over time. When
explaining how reflecting on her self-tracking data helps her maintain a practice of self-tracking,
one participant said, “It’s nice to just kind of like look back at everything I’ve done and have it
all right there, just to kind of motivate me to keep going with my habits and just feel proud of
how much I’ve done in the past year”. Another participant explained that seeing her progress
through her data made her enjoy self-tracking, “Once I started seeing it [progress] I started liking
it [tracking], and when we did the measuring inches, like waist, and the bust and all those
measurements as well, and actually watch them go down after the course of a few months was
amazing”. Another participant explained that using data to see her progress to motivates her to
continue self-tracking to see how far she can push herself, “I think just having like a visual
representation of how far I’ve come or how much better I am now or how much I’ve improved is
a good way of motivating myself, to see, I’ve come this far, so let’s see how much further I can
go. If it’s working let’s see how much, you know, you can improve on top of that”. Rather than
being driven to self-track by seeing progress, some participants expressed that they are driven to
self-track because they find collecting their data to be satisfying, “I like quantifying things. I
don’t know how to say it any better than that. It’s part of my personality that you know I like
collecting things, I like lining things up and seeing visual feedback”. These responses reveal that
for this sample, self-tracking adherence is either aided by evidence of success in behavior change
and progress toward goals, or simply by an interest in data.
111

A prevalent topic that emerged when discussing what helps participants adhere to self-
tracking was ease and convenience. One participant stated, “Things just need to be sort of easy. I
don’t have a lot of time at the moment with the kids and getting to work, and stuff like that. So
convenience is very important with the process”. Participants in this sample frequently stated that
they use the food tracking mobile application My Fitness Pal and/or the wearable device Fitbit. A
common reason why participants reported that they used these tools was because they have
features that give them low friction for utilization. For example, several participants mentioned
that they use the My Fitness Pal application for tracking their dietary intake because it has a large
food database and a food item barcode scanning system. Both of these features enable them to
track their dietary intake without having to manually enter the data for each food item they
consume. When explaining what she liked about using the Fitbit, one participant stated, “I really
like just how effortless it is. Like once it’s charged I just put it on my wrist and it does
everything for me. Except for like when I have to track a workout. But even then it’s just like
hitting a button at the beginning and the end. So I really like that it can give me such
comprehensive data without me really having to do anything for it”. Another participant said of
using the Fitbit, “A lot of times I just want to run on auto pilot. You know the thing that I like
about tracking all this data is I don’t like to have to do a lot of manual work. I like the fact that
you can wear a Fitbit and it just logs for you, and when you want to pay attention you can pay
attention”. These comments indicate that specific features and designs of tracking tools can play
an important role in facilitating self-tracking adherence.
While some features of self-tracking tools are helpful for facilitating adherence, others
have variable utility. One theme that emerged when discussing motivations for self-tracking was
attitudes toward added features that are integrated into many self-tracking tools to motivate
112

adherence, such as social networks, virtual rewards, and games. Comments about these features
revealed that they are appealing to some participants but annoying to others. For example, one
participant explained, “MapMyRun [mobile application] has a whole badges component. So if
you run faster than the top three people who do a certain section of your run that they have in
their system, and you do it, you know, third fastest compared to everyone in their community,
they give you some kind of [virtual] badge. And I know some people, like really like that and
care about it, and to me that’s irrelevant. The whole gamification and stuff like really, I think it’s
silly”. This reveals that a mobile application feature that is intended to be motivating does not
serve that purpose for everyone. However, another participant had an opposing perspective on
gamification in self-tracking tools, “It [Fitbit] did have games on it where you could friend
people, and there’s one that’s like a step challenge and you see who takes the most steps, and
whether you want to do a day or a week or whatever, and then you can send little encouragement
things to friends, and it’s just like a little smiley face thing that says you’re doing great, and stuff
like that. Um, and there’s two other games, I can’t really remember what they were, but that one
was really fun and it kind of helps you stay motivated, too”. While this demonstrates that social
components of self-tracking tools are useful for some, other participants described them as
problematic. One participant explained that she had a negative experience with the social
component of a dietary intake tracking tool that she uses, “I worked with a trainer earlier this
summer, and I allowed them to have access to my My Fitness Pal [mobile application] diary.
And I actually hated that. I was like, I did not like the idea that somebody could look at what I’m
eating. Because I just, would get worried like, they’re gonna think like, ‘Oh why is she eating
that? Why is she eating so much?’ And so it kind of actually was a bit of a, it had a bit of a
sabotaging effect. Or like, I experienced some reactance from it where I didn’t want to be honest
113

in my online diary”. This comment suggests that perhaps specific tracking data, such as dietary
intake, can be sensitive for some individuals to share through social features in tracking tools.
Overall, the difference in sentiments communicated by participants may suggest that the
utilization of features in tracking tools (for example, as fun and engaging components versus
ways to share data with health care professionals) could be an important determinant of whether
they are perceived as helpful or hindering.
Some participants described using simple strategies that help them adhere to self-tracking
their weight-related behaviors. One of the common strategies mentioned was reviewing past data
as a motivator to continue self-tracking. As one participant described, “I can look through my
spreadsheet and I have so many rows of data, and so I just want to stick with it”. Another
participant stated, “The thought of maybe like finishing the notebook I’m using would be kind of
cool. Like I just look at all the pages that I’ve filled because I’ve been tracking so much, so that’s
kind of rewarding, so it kind of keeps… I like seeing that, so then I keep going”. Another
participant described using a visual cue of her self-tracking behavior to start and continue her
self-tracking habit, “I read somewhere that you should have a visual reminder of how many days
you’ve been doing something….the idea was to take like a jar full of really small objects, like a
jar of paper clips…and every day that you do something you’re supposed to do you put a paper
clip into a different jar. And so if you ever feel like skipping a day, you see how many is there in
the second jar and you don’t want to break the chain. So I started, I printed out a calendar, that is
like a three month calendar, and every day I tracked I drew a little blue line over that day. And so
as soon as I started doing that, I never stopped. I did it every day after that”. These tactics
resonated as a simple but possibly very effective method to motivate one to continue self-
114

tracking. Insights gathered from participants about what facilitates a habit of self-tracking could
be used to help others self-track.
Self-tracking adherence barriers. Although the purpose of this study was to understand
what helps the participants adhere to a practice of self-tracking, much of the responses from
participants about adherence included references to actual or potential barriers to long-term self-
tracking. Even among the diligent self-trackers that comprised this study sample, some expressed
that they have experienced that long-term adherence to self-tracking can be difficult. The most
common discussion about barriers to self-tracking centered on dietary intake. Forty-four percent
of the interviewed sample mentioned that they had experienced barriers to self-tracking dietary
intake. One aspect that participants mentioned as a barrier to tracking dietary intake is perceived
social stigma, especially when one is in a social meal setting. One participant explained that this
was an issue while he was using a picture-taking application to track his dietary intake, “It’s
intrusive. Like people think you’re one of those like weird, like you know Instagram foodie
people if you’re like whipping out your phone taking a picture of everything you eat. You gotta
explain it like, ‘Oh I’m trying to track this, you know, just for the data sciences kind of
perspective and to get a better understanding of my health’. It’s just like, there’s this weird like,
there’s this cultural barrier to overcome there”. Comments such as this indicate that tracking
dietary intake in a social setting can make people feel self-conscious, which discourages
adherence.
Another barrier that was frequently mentioned is the tedious nature of dietary tracking.
Several participants commented that having to input food data into dietary tracking applications,
especially when having to estimate what is in food that is not included in application databases,
causes fatigue with dietary tracking. One participant explained, “It’s really an annoyance in daily
115

life to track so much. Because every time I eat something I will weigh it, or if I can’t I will take a
picture of it and I will try and kind of guestimate at a later time what I ate…It steals a lot of time
and a lot of focus”. Another participant explained that the annoyances of tracking her diet caused
her to stop tracking it, “I’ve tried to do diet tracking, but I can’t keep up with that as well as I do
the fitness workout stuff. I just, I really want it to be accurate. So like it’s kind of annoying to
have to like scan all the barcodes, or if you have something that doesn’t have a barcode on it and
like putting that in there, and you’re like hoping it’s accurate. Also like weighing stuff, and all
that.” Another participant said, “One thing I don’t like, and this will go for any calorie tracking
app, is that just having to like guess how many calories I’m eating all the time gets really tiring.
Especially because I really can’t tell, like, I eat a lot of foods that are not as common in My
Fitness Pal, and sometimes it just gets so troublesome to try and guess how many calories I’m
eating that I’d rather just eat like a prepared salad from Trader Joe’s because it has the calories
on the sticker on the back. But that’s probably less healthy, right? So I feel like it might
encourage some people to eat processed food, because it’s easier to track that way”. It was
apparent from these conversations that the limitations of dietary tracking tools, the frequency of
tracking required to monitor dietary intake, and the need for accurate data to effectively track
diet make adhering to long-term dietary intake tracking very difficult for many people.
Another barrier to self-tracking adherence that was cited by several participants is the fact
that it can be unclear what data means and how one should use data. One participant summarized
this issue by stating, “You can track all of this data, but what does it really mean? At the end of
the day you end up with a ‘big data’ problem that doesn’t actually add a whole lot of value”.
Other participants speculated that people tend to get fatigued with self-tracking quickly because
of the issue of not knowing how to use self-tracking data to inform behavior change. One
116

participant speculated that this issue may be a cause for the high self-tracking attrition that tends
to happen in the general population, “A dashboard or say a mobile app is not going to change
behavior, especially if the learning curve is really high…For somebody who says, ‘OK, I’m 15
lbs overweight’, and they don’t even know where to start at what to do, there aren’t a lot of tools
to give guidance for that”. Another participant stated, “I’ve found these tools to be great for
tracking, but if you don’t know what you’re doing, then it makes it a lot more difficult”. Several
participants explained that not knowing what to do with data was a particular barrier to adhering
to self-tracking sleep. As one participant described, “I didn’t really know how to improve the
quality. Like with steps you can do something, you can walk. But with sleep it’s not clear to me
how I can improve sleep”. Comments such as these demonstrate that in order for people to
adhere to a practice of self-tracking, it is likely important for them to gain insights from the data
they track and know what to do with those insights to support behavior change and goal
attainment.
How self-tracking supports behavior change and goal attainment. When participants
were asked how self-tracking helps them maintain behavior change, three topics were repeatedly
mentioned: it helps with awareness of behaviors, it helps with accountability to goals, and it
helps prevent discouragement. The discussion of increased awareness occurred most commonly
in reference to eating behaviors. Several participants stated that self-tracking their dietary intake
helped them pay more attention to their dietary intake and avoid mindless eating. One participant
explained that tracking her dietary intake using a photo food diary mobile application helps her
to take a step back and evaluate whether or not she is in fact hungry when she wants to eat, “In
order to snap a picture, and upload it, you’ve almost got an extra step that you’ve gotta go
through. So it does make you think more…It’s an extra process you have to go through. Like,
117

actually am I really hungry or am I just in a really bad mood? So I guess it’s almost like a
buffer”. Participants also suggested that self-tracking is particularly important for managing how
much food they are eating because it can be easy to underestimate portion size. One participant
explained that after taking a break from tracking her diet, she realized tracking her diet was
important for understanding how much food she was consuming, “I thought I knew how much I
was eating because I had been tracking and I had been weighing stuff. I thought I could eyeball
everything. But soon you forget how things weigh, and I thought maybe some of the portions
were maybe a bit smaller than they were in fact. So I started gaining a little bit of weight again,
so I thought, ‘OK, I need to track my food”. Several participants also commented on how self-
tracking helps them be more aware of sedentary behavior. As one participant explained, “When
people get busy they don’t pay attention and they think they’re more active than they are. And
I’ve had weeks where I’ve found myself going through a work crunch, and I do pull that data and
realize, ‘Oh I haven’t moved much in the last three days. So it is helpful for that. Because, you
know, unconscious patterns where you think things are OK, but you look back, you look at the
data and you’re like, ‘Well, I think I’ve been doing OK’. I really thought that I did. I really
thought that I was doing good, but I wasn’t”. Such comments from participants about how self-
tracking has supported increased awareness of their eating habits and sedentary behaviors
highlights the importance of dietary tracking for not only initiating but also maintaining dietary
and physical activity behavior changes.
Participants also mentioned that self-tracking helps them maintain accountability to their
health goals. On participant explained greater accountability through self-tracking has helped her
with her exercise goals, “Instead of just being like, ‘Oh, I can do this later’, writing it down has
really helped me stick with it. Because I don’t want to see all those gaps in dates when I write it
118

down and stuff like that. It’s just, writing it down really solidifies, man, this happened’. Another
participant explained how self-tracking helps her make food choices that are more in line with
her weight goals, “It helps me make smarter choices overall. You know, when you go to track a
Dunkaccino, and you see in front of you just how many carbs and how many calories and how
much sugar is in it, you know, now if I go to Dunkin’ Donuts, if I want to treat myself to
something I’ll get a non-fat latte”. These experiences highlight that self-tracking can support
goal-oriented behavior by aiding individuals in making healthful decisions about physical
activity and dietary intake.
Participants commonly reported that being able to reflect on past data and seeing their
overall progress helps them from becoming discouraged about reaching their goals when it
seems like they struggling to make progress or have setbacks. One participant explained how
reflecting on her running data keeps her from getting discouraged when she has trouble with her
running goals, “Tracking gives you like a baseline to measure yourself, because it’s really hard
when you’re stuck in it every day. You know, like maybe I’ll have a run next week that’s like
super slow and it’s ‘Ah, why am I so slow? Why haven’t I improved at all?’ But I can still look
back and say, ‘No, I have improved, like maybe this day was bad, but I’m still doing a lot better
overall.’ So it’s like, I can look at the trend. One blip on the radar, you know, that’s all it is”.
Another participant stated that reflecting on her weight data keeps her from getting discouraged
when she loses perspective on her weight loss progress, “I really like the apps that show you
your change over time in the big picture, because it’s so hard to be, like for example, I’ve lost 30
pounds this year, and at times it just feels like I haven’t changed at all. Or I’ll just like kind of
discount the progress that I’ve made. So it’s helpful to just be like, geez, I’ve come really far
since last year even though it’s been very slow going”. This aspect of self-tracking seems to
119

increase competence for achieving health goals. For example, one participant stated, “Once you
go, ‘Oh wow, I actually can change this, I’m not just like, stuck here’, it makes it easier to make
other changes.” Reflection on self-tracking data can also provide insight into problem solving
approaches in the face of issues with goal attainment. One participant explained, “If maybe I’m
eating more than I think I am, and then I’m gaining weight, I can at least look back at the
pictures [of food consumed] and be like oh, OK, so maybe I was just thinking that I was eating
the same for some reason but objectively I can see that I’m not”. Another participant expressed a
similar sentiment about reflecting on data for goal attainment problem solving, “I could look
back and kind of do some trouble shooting if I didn’t reach my goal. Like why didn’t I lose
weight this weekend? Well, because you ate over your maintenance or at your maintenance
every day”. Using past self-tracking data to reflect on progress appears to be an important as
easily implementable tactic for long-term adherence to self-tracking and for goal achievement.
Negative consequences of self-tracking. While most of the participants expressed that
self-tracking was helpful for goal attainment, some participants did mention negative
repercussions from self-tracking, specifically of dietary intake. Specifically, closely tracking
dietary intake precipitated problematic eating behaviors for some individuals. One participant
explained, “ It got to a point where I was just sick of it because it felt like tracking calories was
getting a bit weird, and I didn’t want to do it anymore. Kind of, I feel like it got me into some
bad eating habits, and kind of an unhealthy relationship with food…I felt like it was encouraging
me to continue to lose weight. And at that time, that wouldn’t have been healthy for me”. For
this participant closely tracking caloric intake lead to unhealthy restrictive eating. For another
participant, consuming over a caloric limit lead to a reactance effect, “Once I went over the
calorie allotment I was supposed to have, my head would just go like ‘let’s eat everything’. So I
120

try to stay away from numbers now. Like weight and calories and thing like that, they tend to
make me a bit neurotic”. These anecdotes indicate that reactions to dietary self-tracking can
inhibit adherence and that caution may be necessary when using self-tracking in dietary behavior
change interventions.
Utility of long-term self-tracking. An interesting theme that emerged from the
interviews was the questionable utility of continued long-term self-tracking. Most participants
explained that they anticipate continuing to self-track because it has been helpful for them to
reach their goals, and they see utility in self-tracking as a means of behavior change maintenance
or to achieve new goals. However, some participants mentioned that they consider self-tracking
to be a temporary “crutch” and hope they do not always need to self-track. One participant
referred to self-tracking as a temporary method to aid behavior change, “I kind of feel like self-
tracking is like training wheels almost”. Another participant stated, “If it can ingrain habits into
me, like if I get used to eating a certain amount when I’ve done a certain amount of exercise, or
you know, having a certain size of meal, rather than a huge mal, I guess that’s the hope, that
eventually it’ll kind of, I won’t have to track anymore because it’ll kind of just be natural”. This
demonstrates that not all individuals who diligently self-track consider self-tracking to be a
necessary component of behavior change maintenance. Rather, some seem to hope that they will
gain insights or that their behavior change will become habit, and that they will not need to
continue to diligently self-track as a lifestyle in order to maintain behavior change.
Recommendations to improve self-tracking. While participants spoke about many
ways self-tracking is helpful for them, they commonly also spoke about what could be improved
about self-tracking to optimize its utility. Of particular note were comments about how
participants wished they had a means to integrate data from different devices they use to track
121

different behaviors and metrics into one tool that would allow them to see how their behaviors
interrelate. One participant stated, “It would be nice to be able to integrate the data, match it up
based on like time stamps or whatever. That way you could see how they were all connected or
not… I’m definitely interested in a more integrated platform, like that integrates physical and
health stuff with psychological stuff and cognitive stuff. I think that would be most useful to
people. I think there’s just a limit with what you can learn when it’s just purely about one thing
or another. And I guess I’m just, I’m not really that, I’m excited about the future of self-tracking
and stuff, and it’s really useful to see what people have done, like the tools that they’re making,
but I don’t think they’re up to the level where people can really get down to deep problems yet.
Kind of curiosities. But I think that’ll get better and better over time”
Another point of note was that participants would like if self-tracking platforms would
have the built in capacity to use their data to make recommendations for how to make
improvements in their behaviors or metrics, or to suggest actions to take to reach their goals. One
participant summarized this sentiment by stating, “I do think that there are levels of usability in
all this information. One is, the lowest level is just giving feedback and quantifying the stuff, uh
the second is trying to correlate things or analyzing it and correlating one thing to another, and
then the third level is kind of recommendations, right, like using that information and saying ‘ok
this is correlated to this’, and I suggest you do this, this, and this. So you know, where these
things are all mostly falling short it’s in the analytics and recommendations. If you look at all
these things that I’ve just been talking about and quantifying, a lot of them give you the metrics
and the feedback and the data, but they don’t really recommend much”. Another participant
expressed, “I think the companies need to figure out really how to help people use the data and
provide more tools to help them answer their questions and what not, rather than just giving them
122

counts and things like that”. Recommendations such as these for improving self-tracking utility
may provide greater reach for self-tracking to greater array of individuals who could benefit from
self-tracking to support weight-related behavior change.
Discussion
This study used a mixed methods approach to describe key attributes of a sample of
adherent self-trackers, understand what helps them adhere to a practice of self-tracking weight-
related behaviors, and discover how self-tracking helps them achieve their health-related goals.
To our knowledge, this is the first study to formally explore factors that facilitate self-tracking
adherence in individuals who are volitional trackers of weight-related behaviors and metrics.
Findings from the quantitative and qualitative components of this study provide important
insights into factors that have the potential to be cultivated to improve self-tracking adherence in
future research. In particular, the results from the quantitative component of this study indicate
that individuals who self-track weight-related behaviors differ from individuals who do not self-
track on potentially modifiable characteristics that may influence adherence to a practice of self-
tracking. In accordance with the hypotheses, self-trackers had greater autonomous motivation for
engaging in healthful dietary and exercise habits, as well as greater perceived competence for
making and maintaining dietary and exercise behavior changes. According to Self-Determination
Theory, these two factors are essential for supporting intrinsic motivation, which is important for
behavior change and behavior change maintenance (G. C. Williams et al., 1996). These findings
suggest that motivation orientation and competence for behavior change could be targets for
interventions to improve adherence to self-tracking of weight-related behaviors.
Findings from the in-depth interviews provide support for how self-tracking may work to
help goal attainment according to self-regulation theories and findings from previous research. A
123

predominant theme from discussions of how self-tracking helps with goal attainment was that it
enhances awareness of monitored behaviors. Theories of self-regulation posit that awareness of
and attention to behavior are necessary for behavior change and behavior change maintenance
(Baker & Kirschenbaum, 1998; Baumeister et al., 1994; Lora E. Burke et al., 2011; Carver &
Scheier, 1990). Many participants stated that self-tracking helps them by making them more
mindful or aware of the behaviors or metrics that they track, which provides support for the
relationship between increased awareness and behavior change posited by self-regulation
theories. Comments from participants also demonstrate that the act of self-tracking helps support
goal-directed behaviors and enhances their perceived competence for making and maintaining
behavior changes. These findings are supported by previous research that suggest self-tracking is
thought to support behavior change by facilitating recall of goals and intentions (Latner &
Wilson, 2002) and by enhancing self-efficacy for behavior change (Yasmin Mossavar-Rahmani
et al., 2004).
Insights from the qualitative component of this study point to strategies that could be
engaged to help people adhere to a practice of self-tracking. The conversations with avid self-
trackers in this study revealed that paying attention to immediate feedback from tracking tools
can help establish a habit of self-tracking because it provides instantaneous reinforcement of and
validation for positive behaviors. Another data-driven strategy for supporting adherence that was
derived from the in-depth interviews is reflecting on past self-tracking data to gain perspective
on goal-related progress. Based on comments from several participants, revisiting past data has
the ability to reduce discouragement about goal attainment because it allows one to reflect on
one’s progress and enhances perceived competence for continued progress. This tactic seems to
be especially important for supporting progress toward goals that take time to manifest, such as
124

strength gains and weight loss. Coaching individuals who are just starting a practice of self-
tracking to pay attention to their data and feedback from their tracking tools could help them
establish an understanding of the utility of self-tracking, thus enhancing motivation for continued
self-tracking. Teaching self-trackers to reflect on past data could be an important approach for
preventing return to unhealthy behavior or weight regain.
One characteristic that is common among the self-tracking participants who completed
the in-depth interviews is that they track several different behaviors or metrics. This provides
them with different markers for their progress toward goal attainment. This tactic could be useful
for helping people stay motivated to progress toward health-related goals that are difficult and
take time to achieve. Having more than one metric for progress could be useful for seeing an
overall picture of progress. For example, if weight loss is stalling but fitness or body
measurements are improving, this could be an indication that one is still heading in the right
direction for achieving a weight-related goal. Interventions aimed at promoting weight-related
behavior change, weight loss, or other health improvements could integrate tracking of different
behaviors and metrics to facilitate different perspectives and insights on progress.
Another strategy for supporting adherence to self-tracking that emerged from the
interviews is setting updated goals after completion of initial goals. Many participants explained
that self-tracking had helped them to achieve an initial goal, and after achieving that initial goal
they set new, often more challenging goals. The new goals compelled them to continue self-
tracking. This finding suggests that self-tracking can have long-term positive effects on behavior
change, such that people can use it to continually push themselves to improve their health-related
behaviors. Updating goals to support self-tracking adherence could help individuals maintain
their positive behavior changes rather than reverting to previous, less healthful behaviors. For
125

instance, if one consistently achieves a daily step count goal aided by self-tracking, rather than
ceasing to self-track, one could create a new goal to run a certain number of miles each week.
According to insights from the self-tracking participants, this updated goal would likely motivate
continued self-tracking, which would also support continued physical activity. Such an approach
could be utilized in interventions for weight loss and in post-intervention periods to support
continued behavior change and health improvement.
An important point that many participants made was that people need to understand what
their self-tracking data means and how to use it in order to benefit from self-tracking. These
participants were self-motivated to derive insights from their self-tracking data, but not all
individuals would necessarily be so self-motivated or have the background knowledge to
understand their data. This suggests that interventions aimed at improving self-tracking
adherence could incorporate educational components aimed at helping individuals understand
what their tracking data means and how to use it for problem solving, progress evaluation, and
continued behavior change. This also suggests that self-tracking tools need to be more
sophisticated in order to support the long-term needs of users. Discussions on this topic that
emerged during the in-depth interviews suggest that one way self-trackers could be better
designed to support user needs is by integrating self-tracking data from different implements in
order to provide an overall picture of patterns and correlations between different behaviors
and/or metrics. This type of design would allow self-trackers to make more informed decisions
about how to progress with behavior change and reach their health-related goals.
Although this study was aimed at examining factors that facilitate self-tracking
adherence, it was evident from the conversations with many of the participants that self-tracking
can be a difficult practice to maintain. This suggests that even the most avid self-trackers can
126

experience difficulty with self-tracking adherence, particularly because of limitations with
dietary self-tracking tools and a general lack of actionable direction from self-tracking tools to
guide user behavior. However, it is evident that long-term adherence to self-tracking is possible,
and that it can be supported by specific tactics and strategies. Future research is necessary to
formally test whether these strategies can be taught and utilized to improve weight-related
behavior self-tracking adherence and outcomes of weight-related behavior interventions.
Study Limitations
This study has limitations that should be noted. It is likely that the study sample consisted
predominantly or entirely of health-conscious subset of the population, since participants were
individuals who self-select into health-promoting activities such as dietary and physical activity
tracking. This may limit the generalizability of the findings for individuals who are not health-
conscious. Given the financial cost of some of the tools people use to self-track, it is possible that
individuals who self-track with technologies are from a higher socioeconomic background. This
may limit generalizability of the findings for individuals from lower socioeconomic
backgrounds. Participants self-reported information about their behaviors through interviews and
questionnaires, so data may be subject to reporting biases. The small sample size of this study
precluded complex analyses of quantitative psychosocial data and of qualitative interview data.
However, this study serves as a hypothesis-generating exercise, so future research can expand
upon the preliminary findings.

Conclusions
This exploratory mixed methods study used questionnaires and in-depth interviews to
gain insights into factors that facilitate long-term self-tracking adherence among individuals who
127

track weight-related behaviors or metrics (including diet, exercise, sleep, stress, or weight).
Findings from this study revealed that there are common tactics and strategies that help volitional
self-trackers initiate and maintain a practice of self-tracking. Many of the factors that facilitate
self-tracking adherence among this sample of participants have the potential to be targeted in
future interventions to improve behavior self-tracking adherence in individuals attempting to
change weight-related behaviors. Future research can use the groundwork laid by this
exploratory study to harness the potential of self-tracking in improving weight-related health
outcomes.

128

Chapter 5: Summary and Conclusions
Summary of Aims and Findings
This dissertation had two overarching aims: 1) to examine psychological, physiological,
and behavioral effects of high sugar test meal consumption compared to high fiber test meal
consumption in overweight minority youth, and 2) to explore factors that facilitate adherence to
self-tracking, a key strategy to aid dietary behavior change. These aims were fulfilled via three
studies. The overall objective of study 1 was to investigate the acute impact of high sugar/low
fiber (HSLF) versus low sugar/high fiber (LSHF) test meals consumed in a laboratory setting on
subsequent ad libitum sugar intake in a sample of overweight minority adolescents. The overall
objective of study 2 was to examine the impact of negative mood on hunger ratings during the
HSLF and LSHF test meal conditions in the same sample of adolescents. The overall objective of
study 3 was to use a mixed methods approach to explore factors that facilitate self-tracking
adherence in a sample of individuals who self-track weight-related behaviors.
The results from study 1 demonstrated that consumption of HSLF and LSHF test meals
had different effects on ad libitum total sugar intake by overweight minority adolescents in the
laboratory setting. Participants consumed more total sugar during the snack period, when they
could choose which foods to consume, in the LSHF condition than in the HSLF condition.
Negative mood experienced during the snack period was expected to mediate the association
between meal condition and ad libitum sugar intake during the LSHF condition, but no evidence
of mediation was found. Another individual-level factor expected to influence the effect of meal
condition on ad libitum sugar intake, habitual sugar intake, was not found to influence the
relationship. However, trait impulsivity was found to moderate the relationship between meal
129

condition and ad libitum sugar intake, such that higher levels of impulsivity were associated with
greater ad libitum sugar intake in both meal conditions.
Findings from study 2 showed that there was a relationship between negative mood and
perceived hunger during both the HSLF and LSHF meal conditions, and that the relationship
differed based on whether it was captured in a cross-sectional or repeated measures manner. At
the baseline time points of both the HSLF and LSHF test meals, higher negative mood was
associated with higher perceived hunger. However, the positive association between negative
mood and hunger found at the start of the day did not continue throughout the remainder of the
meal condition time periods. During both meal conditions, perceived hunger declined throughout
the day, and in the LSHF condition only, an increase in one’s negative mood compared to one’s
usual negative mood was associated with a more pronounced decrease in hunger. Contrary to
what was expected, the relationships between negative mood and hunger were not influenced by
habitual sugar intake. However, habitual sugar intake had an independent effect on hunger
ratings at the beginning of the day during the LSHF condition, whereby greater habitual sugar
intake was associated with greater perceived hunger ratings.
Study 3 findings provided insights into how self-trackers differ from people who do not
self-track on key characteristics. This study is also one of the first to explore approaches that
help individuals who self-track adhere to self-tracking practices over the long-term. Compared to
participants who do not self-track, those who do had higher autonomous motivation for engaging
in healthful exercise and dietary habits and higher perceived competence for making or
maintaining a change toward healthful dietary and exercise behaviors. There were no differences
between the groups in health locus of control orientation or dispositional mindfulness. Given that
the quantitative component of this study was cross-sectional in design, we could not determine
130

whether higher autonomous motivation and competence facilitates self-tracking, or whether self-
tracking increases autonomous motivation and competence. Future longitudinal studies would be
necessary to determine causality in these associations. The exploratory qualitative component of
the study revealed several strategies that help people adhere to a habit of self-tracking. One
approach that appears to assist in adherence is setting new goals after achieving an initial goal.
This practice encourages continued self-tracking to support goal attainment and facilitates
maintenance of goal-oriented health behaviors. Paying attention to feedback from tracking
devices and reviewing both short-term and long-term data also appear to aid adherence to self-
tracking. Insights from the in-depth interviews revealed that immediate feedback reinforces the
utility of self-tracking because it provides instantaneous reinforcement of and validation for
positive behaviors. Data review can itself be a motivator for continued self-tracking; participants
revealed that data review promotes awareness of progress toward goals that take time to achieve,
so self-tracking is viewed by many of the participants interviewed as a method to prevent
discouragement. Further, understanding how to interpret and utilize self-tracking data was
revealed as an important determinant of self-tracking adherence. This first exploratory study of
factors that facilitate self-tracking adherence revealed many insights that remain to be tested in
future research.
The findings from these three studies provide new insights into factors that may drive
high-sugar, low-fiber dietary patterns, as well as insights into adherence to self-tracking that
might be used to improve future interventions for dietary behavior change. Overall, findings
from study 1 support that dietary sugar intake can acutely impact food choices made by
overweight adolescents. More specifically, restricting sugar-rich food intake by overweight
adolescents appears to lead to compensatory sugar intake when sugar-laden foods are available.
131

Findings from this study also suggested that higher trait impulsivity predisposed participants to
consume greater quantities of ad libitum sugar regardless of sugar content in condition-specific
meals. This suggests that impulsive tendency may make some people more inclined to reach for
sugar-laden foods regardless of prior sugar intake during the day. This is the first study, to our
knowledge, to examine and reveal a phenomenon of compensatory sugar intake in humans.
However, this finding is supported by and extends previous research that has demonstrated
compensatory intake of specific preferred foods after restriction of those foods, which has been
studied in young children (Birch et al., 2001b; Jennifer Orlet Fisher & Leann Lipps Birch, 1999).
Based on findings from previous research (Major, 2015) and affect regulation theories,
specifically mood management theory and the mood maintenance hypothesis (Andrade, 2005;
Clark & Isen, 1982; Zillmann, 1988), we expected that there would be a psychological
mechanism driving greater sugar intake subsequent to sugar restriction in the high fiber, low
sugar test meals. Specifically, we postulated that compensatory sugar intake would be driven be
an increase in negative mood in reaction to sugar restriction, and a subsequent attempt to assuage
negative mood by ad libitum consumption of high-sugar foods. However, the fact that this
hypothesis was not supported by the findings reveals that there may be other factors that could
drive compensatory sugar intake in overweight adolescents.
Factors that could potentially drive compensatory sugar intake that were not tested by
study 1 include other psychological factors or physiological responses to sugar deprivation. It is
possible that simply cravings for sugar drove the participants to consume more sugar during the
ad libitum snack period in the LSHF meal condition. Food cravings were not assessed by this
study, so it was not possible to test whether cravings were involved in the observed
compensatory sugar intake. One potential physiological explanation for compensatory sugar
132

intake involves neural reward response. There is evidence that overweight and obese individuals
have reduced dopamine receptor availability in regions of the brain involved in reward response
(Blum, Thanos, & Gold, 2014). Low dopamine receptor availability leads to under-stimulated
reward circuits (Blum et al., 2014). This has been hypothesized by researchers to drive
consumption of highly palatable foods, such as those high in sugar, as a temporary way to
compensate for an under-stimulated neural reward circuitry (Blum et al., 2014; Thanos,
Michaelides, Piyis, Wang, & Volkow, 2008). It follows that under-stimulated reward circuitry
may explain the compulsion by overweight adolescents deprived of sugar intake to eat higher
amounts of sugar when they are faced with a free-feeding situation. This potential mechanism
could be an important line of inquiry for future studies aimed at explaining compensatory sugar
intake in this population.
Study 2 was intended to build upon study 1 by revealing additional factors that could
perpetuate high sugar, low fiber dietary patterns in overweight adolescents. We anticipated that
results would reveal a cascade of effects from sugar restriction in the LSHF test meal, such that
sugar restriction would lead to increased negative mood, which would lead to increased hunger
throughout the day. This cascade of effects was not found. While the observed findings were not
anticipated, the finding that higher within-person negative mood was associated with a more
pronounced decrease in hunger during the LSHF test meal condition is supported by research
that has shown negative mood can suppress appetite (Dingemans et al., 2009; Macht, 2008). The
current study extends previous research with the unexpected finding that the relationship
between negative mood and perceived hunger differed based on whether it was assessed at
baseline (essentially as a cross-sectional relationship) or through repeated measures (change
throughout the day). Had the variance from the repeated measures negative mood variable not
133

been disaggregated into between-person and within-person effects in the study 2 analysis, the
nuances of the relationship between negative mood and hunger would not have been observed.
Such a difference in findings between cross-sectional and repeated measures relationships
highlights the importance of disaggregating the variance of repeated measures variables in
multilevel modeling. Overall, the findings from this study suggest that in the fasting state, higher
negative mood was associated with higher hunger, but a combination of the effects of high fiber
and the effects of negative mood may have led to decreased hunger during the LSHF condition.
Future research, in particular experimental studies that manipulate both dietary intake and mood,
would need to be conducted to confirm this postulation.
The potential implications of studies 1 and 2 led us to conduct study 3 in order to gain
insights that have the potential to improve dietary behavior change maintenance. To our
knowledge, study 3 is the first qualitative study to explore how people who self-track weight-
related behaviors manage to adhere to the practice over the long-term. Previous research has
demonstrated that self-tracking is an important strategy for maintaining dietary behavior change,
as well as for supporting weight loss and weight maintenance (Lora E. Burke et al., 2011; Burke
et al., 2005). The study of adherent self-trackers is relatively new and little research has been
done to tap into the wealth of insights that can be gleaned from individuals who manage to
adhere to long-term self-tracking. The single previous study we could find in the literature that
was conducted to gain knowledge about adherence in long-term self-trackers focused exclusively
on what factors create barriers to self-tracking (Choe, Lee, Lee, Pratt, & Kientz, 2014). Findings
from study 3 extend this previous research by taking a novel approach and investigating what
can instead facilitate self-tracking adherence. Findings from study 3 point to directions for future
134

research on strategies to improve self-tracking adherence that can be utilized in dietary behavior
interventions.
Overall, from the findings for studies 1 and 2, we conclude that test meals differing in
sugar and fiber content have different effects on ad libitum food choice but not on negative mood
or perceived hunger levels in overweight adolescents. Results from these studies support that
attempts at long-term reductions in sugar intake and increases fiber intake in this population may
be obstructed by a rebound effect of increased sugar intake. From the findings for study 3, we
conclude that self-tracking is a viable intervention strategy for aiding dietary behavior change,
and that there are indeed modifiable characteristics and strategies common among people who
successfully adhere to self-tracking of weight-related behaviors that can be harnessed and
utilized in future research to improve self-tracking adherence and to aid success of dietary
behavior interventions. Findings from study 3 also provided useful knowledge that can guide
development of future self-tracking tools to improve their utility for supporting weight-related
behavior change.
Implications and Future Research
Given that overweight and obesity are significant health issues among African American
and Hispanic adolescents, it is critical to elucidate behavior change strategies that will most help
these populations. Dietary sugar and fiber intake are relevant intervention targets for weight
reduction, given their respective negative and positive impacts on weight, and that African
American and Hispanic adolescents largely do not meet recommendations for these dietary
components. Understanding how the effects of dietary sugar and dietary fiber consumption
interrelate can provide insight into what reinforces unhealthy dietary patterns that are low in fiber
and high in sugar. The results from studies 1 and 2 provide insights that can inform dietary
135

interventions in these groups, and insights from study 3 can inform approaches that can improve
the long-term outcomes of such interventions. Nonetheless, it is clear from the findings of this
dissertation that further research be conducted to understand the impacts of high sugar compared
to high fiber intake in this population and how self-tracking adherence strategies can be
harnessed for interventions to improve these dietary behaviors.
The fact that adolescents were found to react to sugar restriction in the controlled
laboratory setting by engaging in compensatory sugar intake has important implications for
dietary interventions in this population. Dietary interventions aimed at reducing sugar intake
should consider and address the possibility that individuals could relapse to high sugar intake
during or after the intervention period. This is particularly important because given the
obesogenic food environment in the U.S., it is likely that most adolescents will be exposed on a
daily basis to an abundance of opportunities to consume high sugar foods. Consequently,
interventions may need to include components that can aid prevention of dietary relapse to
improve long-term dietary behavior change maintenance. It is clear from study 1 findings,
however, that further research is necessary to understand why adolescents respond to sugar
restriction this way. The findings indicate that a negative mood response to sugar restriction is
not likely to be a factor that can be targeted to prevent the observed rebounding behavior. To rule
out negative mood as a potential target, future work should include both a test meal and mood
manipulation component. Future studies should also aim to examine other factors that could
drive a compensatory sugar intake response. Research on this issue should be continued in both
laboratory and free-living settings. Studies conducted in a laboratory setting would provide a
controlled environment to examine acute relationships between sugar restriction, compensatory
sugar intake, and potential mechanisms. Studies conducted in real-living situations would
136

complement laboratory-based studies because they would allow for testing of whether
relationships found in a laboratory setting hold up in a naturalistic setting.
Findings from study 2 indicate that the relationship between negative mood and
perceived hunger in overweight adolescents is complex. During a fasting state, negative mood
seems to be related to greater perceived hunger in this group. This may have implications for
breakfast food choice and intake quantity. However, even though this study used repeated
measures analyses, the relationship between negative mood and hunger in the fasting state (at the
baseline time point for each meal condition day) is essentially cross-sectional. Consequently, it is
not possible to rule out that higher hunger actually causes negative mood. The nature of this
relationship could be better understood using a different study design, for example, a laboratory
study that manipulates mood and measures effects of mood manipulation on hunger. Another
potential study design that could a free-living study using Ecological Momentary Assessment
methods with event-based sampling to examine how self-reported changes in mood affect
subsequent perceived hunger.
Given that self-tracking can help with competence for behavior change and remind one of
goal-oriented behaviors, it is possible that teaching adolescents to self-track and use their self-
tracking data could lessen or prevent the rebounding effect of sugar restriction observed in study
1. Although study 3 was conducted in a sample of adults who were predominantly white, self-
tracking has the potential to be applied to minority adolescent populations. The utility of tracking
for aiding dietary behavior change in overweight minority adolescents is supported by the
ubiquity of tools such as smartphone apps and on-body sensors that can aid in self-tracking of
health-related behaviors (O’Reilly & Spruijt-Metz, 2013). According to research by the Pew
Internet & American Life Project, 73% of adolescents aged 13-17 years old own or have access
137

to smartphones (Lenhart, 2015). Smartphone ownership and access among minority adolescents
is comparable to if not greater than the average for U.S. adolescents; an estimated 71% of
Hispanic adolescents and 85% of African American adolescents own or have access to a
smartphone (Lenhart, 2015). Self-tracking using smartphones can be done using a variety of
health apps. While health app use among adolescents is not currently known, evidence indicates
that app use in general is becoming popular among this age group. A recent study found that
53% of adolescents have downloaded an app onto their smartphone or tablet computer (Fox,
2013). Estimates for ownership among adolescents of on-body devices that are used for self-
tracking health behaviors and related metrics (such as FitBit or Jawbone UP devices) are not
available; however, the pervasiveness of smartphone ownership and app usage among
adolescents demonstrates that self- tracking using smartphones could be a feasible intervention
approach to support dietary behavior change in adolescent populations.
Through examining effects of test meals differing in sugar and fiber content on
overweight adolescents and exploring factors that can facilitate weight-related behavior self-
tracking, this dissertation focused on factors that may perpetuate sugar-rich, fiber-poor diets and
suggest how self-tracking can be used as strategy in interventions to change unhealthful dietary
behaviors. Overall, findings from this dissertation demonstrate that sugar-rich, fiber-poor meals
can have negative effects in this population, and that self-tracking adherence is modifiable, so it
had the potential to be taught to adolescents and implemented in dietary interventions.
Limitations
This dissertation has limitations that should be noted. The controlled laboratory based
setting for studies 1 and 2 provided a way to examine the acute impacts of test meals differing in
sugar and fiber content on various outcomes in a controlled manner. While this study design was
138

a major strength, it is also a limitation because it does not allow for examination of how the
dietary intake, physiological, psychological, and behavioral factors studied interrelate in a free-
living setting. However, it is possible that the environment set up by the laboratory-based study
design is somewhat close to a free-living school setting. Participants were placed in an
environment where they were provided with set meal times and specific food choices. This is
likely to mimic the structured meal times during the typical school day as well as the limited
choices provided by free school lunches. It is possible that the relationships found in the
laboratory-based settings in studies 1 and 2 would reflect what happens during a school day. For
example, since participants were provided with set meal times during the laboratory visits, their
perceived hunger may have developed like it would during a school day during which
adolescents are allowed only to eat at set snack and lunch periods. Consequently, the findings
from study 2 regarding the relationships between condition-specific meal sugar and fiber content,
negative mood, and hunger may actually be close to those relationships that would occur during
a free-living school day. However, it is important to continue research on these topics in free-
living settings to understand how these factors interrelate in real life.
Other limitations that should be noted include the fact that negative mood was not
manipulated in the laboratory study. This may explain the lack of findings for relationships
between negative mood and food choice and between negative mood and hunger. It was not
possible to recruit a sample of adherent self-trackers from African American and Latino
adolescent populations. Instead, the sample for study 3 consisted predominantly of White adult
self-trackers. This sample was also likely predominantly made up of a health-conscious subset of
the populations since participants will be individuals who self-select into health-promoting
activities such as dietary and physical activity monitoring. These factors may limit the
139

generalizability of the study 3 findings for overweight minority adolescents. The small sample
size of study 3 precluded complex analyses of quantitative psychosocial data. However, this
study serve as a hypothesis-generating exercise, so future research can expand upon these
preliminary findings. Despite the limitations of this dissertation, the findings provide useful
insights and directions for future dietary interventions in overweight adolescents.
Contribution to the Literature
The findings from this dissertation make important contributions to the body of research
on the effects of dietary intake in overweight adolescents and on the emerging self-tracking
literature. To our knowledge, this is the first study that has used a laboratory-based cross-over
study design to examine acute effects of meals differing in sugar and fiber content psychological,
physiological, and behavioral outcomes in overweight minority adolescents. Most notably,
findings from this dissertation are the first to our knowledge to demonstrate that restricting sugar
intake in this population can have a negative impact on subsequent food choices by leading to a
compensatory sugar intake response. Additionally, this dissertation includes the first formal
study of individuals who are adherent to self-tracking weight-related behaviors, and has provided
important insights into strategies that can improve self-tracking adherence for weight-related
behaviors. The findings from this dissertation reveal several avenues for future research into
effects of dietary sugar intake in overweight adolescents and the potential for use of self-tracking
in dietary interventions.

140

Alphabetic Bibliography
Adam, T. C., & Epel, E. S. (2007). Stress, eating and the reward system. Physiology & behavior,
91(4), 449-458.
Alfenas, R. C. G., & Mattes, R. D. (2005). Influence of glycemic index/load on glycemic
response, appetite, and food intake in healthy humans. Diabetes care, 28(9), 2123-2129.
Anderson, J. W., Baird, P., Davis Jr, R. H., Ferreri, S., Knudtson, M., Koraym, A., . . . Williams,
C. L. (2009). Health benefits of dietary fiber. Nutrition reviews, 67(4), 188-205.
Anderson, J. W., Smith, B. M., & Gustafson, N. J. (1994). Health benefits and practical aspects
of high-fiber diets. The American journal of clinical nutrition, 59(5), 1242S-1247S.
Andrade, E. B. (2005). Behavioral consequences of affect: Combining evaluative and regulatory
mechanisms. Journal of Consumer Research, 32(3), 355-362.
Angulo, P. (2007). Obesity and nonalcoholic fatty liver disease. Nutrition reviews, 65(suppl 1),
S57-S63.
Asakawa, A., Inui, A., Kaga, T., Yuzuriha, H., Nagata, T., Fujimiya, M., . . . Kasuga, M. (2001).
A role of ghrelin in neuroendocrine and behavioral responses to stress in mice.
Neuroendocrinology, 74(3), 143-147.
Avena, N. M., Rada, P., & Hoebel, B. G. (2008). Evidence for sugar addiction: behavioral and
neurochemical effects of intermittent, excessive sugar intake. Neuroscience &
Biobehavioral Reviews, 32(1), 20-39.
Baer, R. A. (2003). Mindfulness training as a clinical intervention: A conceptual and empirical
review. Clinical psychology: Science and practice, 10(2), 125-143.
Baer, R. A., Smith, G. T., Hopkins, J., Krietemeyer, J., & Toney, L. (2006). Using self-report
assessment methods to explore facets of mindfulness. Assessment, 13(1), 27-45.
Baker, R. C., & Kirschenbaum, D. S. (1998). Weight control during the holidays: highly
consistent self-monitoring as a potentially useful coping mechanism. Health Psychology,
17(4), 367.
Balch, P., & Ross, A. W. (1975). Predicting success in weight reduction as a function of locus of
control: A unidimensional and multidimensional approach. Journal of consulting and
clinical psychology, 43(1), 119.
Ball, K., & Crawford, D. (2006). Socio-economic factors in obesity: a case of slim chance in a
fat world? Asia Pacific journal of clinical nutrition, 15(Supplement), 15-20.
Baranowski, T., Mendlein, J., Resnicow, K., Frank, E., Cullen, K. W., & Baranowski, J. (2000).
Physical activity and nutrition in children and youth: an overview of obesity prevention.
Preventive Medicine, 31(2), S1-S10.
Bassett, R., Chapman, G. E., & Beagan, B. L. (2008). Autonomy and control: The co-
construction of adolescent food choice. Appetite, 50(2), 325-332.
Baumeister, R. F., Heatherton, T. F., & Tice, D. M. (1994). Losing control: How and why people
fail at self-regulation: Academic Press.
Bennett, G. G., Wolin, K. Y., & Duncan, D. T. (2008). Social determinants of obesity. Obesity
Epidemiology: Methods and Applications, 342-376.
Benton, D. (2002). Carbohydrate ingestion, blood glucose and mood. Neuroscience &
Biobehavioral Reviews, 26(3), 293-308.
Benton, D., & Nabb, S. (2003). Carbohydrate, memory, and mood. Nutrition reviews, 61(suppl
5), S61-S67.
141

Berg, B. L., & Lune, H. (2004). Qualitative research methods for the social sciences (Vol. 5):
Pearson Boston.
Berridge, K. C. (2003). Pleasures of the brain. Brain and cognition, 52(1), 106-128.
Beuther, D. A. (2009). Obesity and asthma. Clinics in chest medicine, 30(3), 479-488.
Beuther, D. A., & Sutherland, E. R. (2007). Overweight, obesity, and incident asthma: a meta-
analysis of prospective epidemiologic studies. American journal of respiratory and
critical care medicine, 175(7), 661-666.
Birch, L. L., Fisher, J., Grimm-Thomas, K., Markey, C., Sawyer, R., & Johnson, S. L. (2001a).
Confirmatory factor analysis of the Child Feeding Questionnaire: a measure of parental
attitudes, beliefs and practices about child feeding and obesity proneness. Appetite, 36(3),
201-210.
Birch, L. L., Fisher, J. O., Grimm-Thomas, K., Markey, C., Sawyer, R., & Johnson, S. L.
(2001b). Confirmatory factor analysis of the Child Feeding Questionnaire: a measure of
parental attitudes, beliefs and practices about child feeding and obesity proneness.
Appetite, 36(3), 201-210.
Blum, K., Thanos, P. K., & Gold, M. S. (2014). Dopamine and glucose, obesity, and reward
deficiency syndrome. Front. Psychol, 5(919), 10.3389.
Blundell, J. E., Stubbs, R., Golding, C., Croden, F., Alam, R., Whybrow, S., . . . Lawton, C.
(2005). Resistance and susceptibility to weight gain: individual variability in response to
a high-fat diet. Physiology & behavior, 86(5), 614-622.
Boggiano, M. M., Chandler, P. C., Viana, J. B., Oswald, K. D., Maldonado, C. R., & Wauford,
P. K. (2005). Combined dieting and stress evoke exaggerated responses to opioids in
binge-eating rats. Behavioral neuroscience, 119(5), 1207.
Bolger, N., & Shrout, P. E. (2007). Accounting for statistical dependency in longitudinal data on
dyads. Modeling contextual effects in longitudinal studies, 285-298.
Booth, S. L., Sallis, J. F., Ritenbaugh, C., Hill, J. O., Birch, L. L., Frank, L. D., . . . Popkin, B. M.
(2001). Environmental and societal factors affect food choice and physical activity:
rationale, influences, and leverage points. Nutrition reviews, 59(3), S21-S36.
Bouchard, C. (2008). Gene–environment interactions in the etiology of obesity: defining the
fundamentals. Obesity, 16(S3), S5-S10.
Boyce, C., & Neale, P. (2006). Conducting in-depth interviews: A guide for designing and
conducting in-depth interviews for evaluation input: Pathfinder International Watertown,
MA.
Bradley, E. H., Curry, L. A., & Devers, K. J. (2007). Qualitative data analysis for health services
research: developing taxonomy, themes, and theory. Health services research, 42(4),
1758-1772.
Brand-Miller, J. C., Holt, S. H. A., Pawlak, D. B., & McMillan, J. (2002). Glycemic index and
obesity. The American journal of clinical nutrition, 76(1), 281S-285S.
Brauchla, M., Juan, W., Story, J., & Kranz, S. (2012). Sources of dietary fiber and the
association of fiber intake with childhood obesity risk (in 2–18 year olds) and diabetes
risk of adolescents 12–18 year olds: NHANES 2003–2006. Journal of nutrition and
metabolism, 2012.
Bray, G. A., & Popkin, B. M. (1998). Dietary fat intake does affect obesity! The American
journal of clinical nutrition, 68(6), 1157-1173.
Briefel, R. R., & Johnson, C. L. (2004). Secular trends in dietary intake in the United States.
Annu Rev Nutr, 24, 401-431. doi: 10.1146/annurev.nutr.23.011702.073349
142

Briefel, R. R., & Johnson, C. L. (2004). Secular trends in dietary intake in the United States*.
Annu. Rev. Nutr., 24, 401-431.
Bruch, H. (1964). Psychological aspects of overeating and obesity. Psychosomatics, 5(5), 269-
274.
Burke, L. E., Styn, M. A., Sereika, S. M., Conroy, M. B., Ye, L., Glanz, K., . . . Ewing, L. J.
(2012). Using mHealth technology to enhance self-monitoring for weight loss: a
randomized trial. American journal of preventive medicine, 43(1), 20-26.
Burke, L. E., Swigart, V., Turk, M. W., Derro, N., & Ewing, L. J. (2009). Experiences of self-
monitoring: successes and struggles during treatment for weight loss. Qualitative health
research.
Burke, L. E., Wang, J., & Sevick, M. A. (2011). Self-monitoring in weight loss: a systematic
review of the literature. Journal of the American dietetic association, 111(1), 92-102.
Burke, L. E., Wang, J., & Sevick, M. A. (2011). Self-monitoring in weight loss: a systematic
review of the literature. J Am Diet Assoc, 111(1), 92-102. doi: S0002-8223(10)01644-5
[pii]
10.1016/j.jada.2010.10.008
Burke, L. E., Warziski, M., Starrett, T., Choo, J., Music, E., Sereika, S., . . . Sevick, M. A.
(2005). Self-monitoring dietary intake: current and future practices. Journal of Renal
Nutrition, 15(3), 281-290.
Burton-Freeman, B. (2000). Dietary fiber and energy regulation. The Journal of nutrition,
130(2), 272S-275S.
Calle, E. E., Teras, L. R., & Thun, M. J. (2005). Obesity and mortality. New England Journal of
Medicine, 353(20), 2197-2199.
Campbell, K. J., Crawford, D. A., Salmon, J., Carver, A., Garnett, S. P., & Baur, L. A. (2007).
Associations Between the Home Food Environment and Obesity ‐promoting Eating
Behaviors in Adolescence. Obesity, 15(3), 719-730.
Carver, C. S., & Scheier, M. (1990). Principles of self-regulation: Action and emotion: Guilford
Press.
Centers for Disease, C., & Prevention. (2010). Vital signs: state-specific obesity prevalence
among adults---United States, 2009. MMWR. Morbidity and mortality weekly report,
59(30), 951.
Centers for Disease, C., Prevention, Centers for Disease, C., & Prevention. (2011). National
diabetes fact sheet: national estimates and general information on diabetes and
prediabetes in the United States, 2011. Atlanta, GA: US Department of Health and
Human Services, Centers for Disease Control and Prevention, 201.
Cepeda-Benito, A., Gleaves, D. H., Williams, T. L., & Erath, S. A. (2001). The development and
validation of the state and trait food-cravings questionnaires. Behavior Therapy, 31(1),
151-173.
Choe, E. K., Lee, N. B., Lee, B., Pratt, W., & Kientz, J. A. (2014). Understanding quantified-
selfers' practices in collecting and exploring personal data. Paper presented at the
Proceedings of the 32nd annual ACM conference on Human factors in computing
systems.
Choi, H., Meininger, J. C., & Roberts, R. E. (2006). Ethnic differences in adolescents' mental
distress, social stress, and resources. Adolescence.
Christakis, N. A., & Fowler, J. H. (2007). The spread of obesity in a large social network over 32
years. New England Journal of Medicine, 357(4), 370-379.
143

Chua, J. L., Touyz, S., & Hill, A. (2004). Negative mood-induced overeating in obese binge
eaters: an experimental study. International journal of obesity, 28(4), 606-610.
Chuang, J.-C., & Zigman, J. M. (2010). Ghrelin's roles in stress, mood, and anxiety regulation.
International journal of peptides, 2010.
Clark, M. S., & Isen, A. M. (1982). Toward understanding the relationship between feeling states
and social behavior. Cognitive social psychology, 73, 108.
Cohen-Cole, E., & Fletcher, J. M. (2008). Is obesity contagious? Social networks vs.
environmental factors in the obesity epidemic. Journal of health economics, 27(5), 1382-
1387.
Cooling, J., & Blundell, J. (1998). Are high-fat and low-fat consumers distinct phenotypes?
Differences in the subjective and behavioural response to energy and nutrient challenges.
European journal of clinical nutrition, 52(3), 193-201.
Corbin, J., & Strauss, A. (2014). Basics of qualitative research: Techniques and procedures for
developing grounded theory: Sage publications.
Cottone, P., Sabino, V., Roberto, M., Bajo, M., Pockros, L., Frihauf, J. B., . . . Grigoriadis, D. E.
(2009). CRF system recruitment mediates dark side of compulsive eating. Proceedings of
the National Academy of Sciences, 106(47), 20016-20020.
Cottone, P., Sabino, V., Steardo, L., & Zorrilla, E. P. (2009). Consummatory, anxiety-related and
metabolic adaptations in female rats with alternating access to preferred food.
Psychoneuroendocrinology, 34(1), 38-49.
Cruz, M. L., Weigensberg, M. J., Huang, T. T. K., Ball, G., Shaibi, G. Q., & Goran, M. I. (2004).
The metabolic syndrome in overweight Hispanic youth and the role of insulin sensitivity.
The Journal of Clinical Endocrinology & Metabolism, 89(1), 108-113.
Curran, P. J., & Bauer, D. J. (2011). The disaggregation of within-person and between-person
effects in longitudinal models of change. Annual review of psychology, 62, 583.
Dabelea, D., Bell, R. A., D'Agostino Jr, R. B., Imperatore, G., Johansen, J. M., Linder, B., . . .
Mayer-Davis, E. J. (2007). Incidence of diabetes in youth in the United States. JAMA: the
journal of the American Medical Association, 297(24), 2716-2724.
Dalen, J., Smith, B. W., Shelley, B. M., Sloan, A. L., Leahigh, L., & Begay, D. (2010). Pilot
study: Mindful Eating and Living (MEAL): weight, eating behavior, and psychological
outcomes associated with a mindfulness-based intervention for people with obesity.
Complementary therapies in medicine, 18(6), 260-264.
Dallman, M. F., Pecoraro, N., Akana, S. F., La Fleur, S. E., Gomez, F., Houshyar, H., . . .
Manalo, S. (2003). Chronic stress and obesity: a new view of “comfort food”.
Proceedings of the National Academy of Sciences, 100(20), 11696-11701.
Dallman, M. F., Pecoraro, N. C., & la Fleur, S. E. (2005). Chronic stress and comfort foods: self-
medication and abdominal obesity. Brain, behavior, and immunity, 19(4), 275-280.
Daubenmier, J., Kristeller, J., Hecht, F. M., Maninger, N., Kuwata, M., Jhaveri, K., . . . Epel, E.
(2011). Mindfulness intervention for stress eating to reduce cortisol and abdominal fat
among overweight and obese women: an exploratory randomized controlled study.
Journal of obesity, 2011.
Davis, C., & Carter, J. C. (2009). Compulsive overeating as an addiction disorder. A review of
theory and evidence. Appetite, 53(1), 1-8.
Davis, J. N., Alexander, K. E., Ventura, E. E., Kelly, L. A., Lane, C. J., Byrd-Williams, C. E., . .
. Goran, M. I. (2007). Associations of dietary sugar and glycemic index with adiposity
144

and insulin dynamics in overweight Latino youth. Am J Clin Nutr, 86(5), 1331-1338. doi:
86/5/1331 [pii]
Davis, J. N., Lê, K.-A., Walker, R. W., Vikman, S., Spruijt-Metz, D., Weigensberg, M. J., . . .
Goran, M. I. (2010). Increased hepatic fat in overweight Hispanic youth influenced by
interaction between genetic variation in PNPLA3 and high dietary carbohydrate and
sugar consumption. The American journal of clinical nutrition, 92(6), 1522-1527.
Davis, J. N., Le, K. A., Walker, R. W., Vikman, S., Spruijt-Metz, D., Weigensberg, M. J., . . .
Goran, M. I. (2010). Increased hepatic fat in overweight Hispanic youth influenced by
interaction between genetic variation in PNPLA3 and high dietary carbohydrate and
sugar consumption. Am J Clin Nutr, 92(6), 1522-1527. doi: ajcn.2010.30185 [pii]
10.3945/ajcn.2010.30185
Deci, E., & Ryan, R. M. (2011). Self-determination theory. Handbook of theories of social
psychology, 1, 416-433.
Dingemans, A. E., Martijn, C., van Furth, E. F., & Jansen, A. T. (2009). Expectations, mood, and
eating behavior in binge eating disorder. Beware of the bright side. Appetite, 53(2), 166-
173.
Doyle, S., Kelly-Schwartz, A., Schlossberg, M., & Stockard, J. (2006). Active community
environments and health: the relationship of walkable and safe communities to individual
health. Journal of the American Planning Association, 72(1), 19-31.
Eaton, D. K., Kann, L., Kinchen, S., Shanklin, S., Ross, J., Hawkins, J., . . . Chyen, D. (2010).
Youth risk behavior surveillance-United States, 2009. Morbidity and mortality weekly
report. Surveillance summaries (Washington, DC: 2002), 59(5), 1-142.
Eisenberg, M. E., Neumark-Sztainer, D., Story, M., & Perry, C. (2005). The role of social norms
and friends’ influences on unhealthy weight-control behaviors among adolescent girls.
Social Science & Medicine, 60(6), 1165-1173.
Ellaway, A., Macintyre, S., & Bonnefoy, X. (2005). Graffiti, greenery, and obesity in adults:
secondary analysis of European cross sectional survey. Bmj, 331(7517), 611-612.
Fabbrini, E., Sullivan, S., & Klein, S. (2010). Obesity and nonalcoholic fatty liver disease:
biochemical, metabolic, and clinical implications. Hepatology, 51(2), 679-689.
Faloia, E., Giacchetti, G., & Mantero, F. (2000). Obesity and hypertension. Journal of
endocrinological investigation, 23(1), 54-62.
Finkelstein, E. A., Khavjou, O. A., Thompson, H., Trogdon, J. G., Pan, L., Sherry, B., & Dietz,
W. (2012). Obesity and severe obesity forecasts through 2030. American journal of
preventive medicine, 42(6), 563-570.
Fisher, J. O., & Birch, L. L. (1999). Restricting access to foods and children's eating. Appetite,
32(3), 405-419.
Fisher, J. O., & Birch, L. L. (1999). Restricting access to palatable foods affects children's
behavioral response, food selection, and intake. The American journal of clinical
nutrition, 69(6), 1264-1272.
Flegal, K. M., Carroll, M. D., Kit, B. K., & Ogden, C. L. (2012). Prevalence of obesity and
trends in the distribution of body mass index among US adults, 1999-2010. JAMA,
307(5), 491-497.
Flegal, K. M., Carroll, M. D., Ogden, C. L., & Curtin, L. R. (2010). Prevalence and trends in
obesity among US adults, 1999-2008. JAMA, 303(3), 235-241.
Folsom, A. R., Yatsuya, H., Nettleton, J. A., Lutsey, P. L., Cushman, M., & Rosamond, W. D.
(2011). Community prevalence of ideal cardiovascular health, by the American Heart
145

Association definition, and relationship with cardiovascular disease incidence. Journal of
the American College of Cardiology, 57(16), 1690-1696.
Ford, E. S., Moriarty, D. G., Zack, M. M., Mokdad, A. H., & Chapman, D. P. (2001). Self ‐
Reported Body Mass Index and Health ‐Related Quality of Life: Findings from the
Behavioral Risk Factor Surveillance System. Obesity Research, 9(1), 21-31.
Fox, S. (2013). Mobile health in context: how information is woven into our lives.
Frary, C. D., Johnson, R. K., & Wang, M. Q. (2004). Children and adolescents' choices of foods
and beverages high in added sugars are associated with intakes of key nutrients and food
groups. Journal of Adolescent Health, 34(1), 56-63.
French, S., Murray, B., Rumsey, R., Fadzlin, R., & Read, N. (1995). Adaptation to high-fat diets:
effects on eating behaviour and plasma cholecystokinin. British Journal of Nutrition,
73(02), 179-189.
Friedman, R., Deci, E. L., Elliot, A. J., Moller, A. C., & Aarts, H. (2010). Motivational
synchronicity: Priming motivational orientations with observations of others‚Äô
behaviors. Motivation and Emotion, 34(1), 34-38.
Fryar, C. D., Carroll, M. D., & Ogden, C. L. (2014). Prevalence of overweight and obesity
among children and adolescents: United States, 1963-1965 through 2011-2012. Atlanta,
GA: National Center for Health Statistics.
Fryar, C. D., Carroll, M. D., & Ogden, C. L. (2015). Prevalence of Overweight, Obesity, and
Extreme Obesity Among Adults: United States, 1960–1962 Through 2011–2012.
Furchner-Evanson, A., Petrisko, Y., Howarth, L., Nemoseck, T., & Kern, M. (2010). Type of
snack influences satiety responses in adult women. Appetite, 54(3), 564-569.
Gaesser, G. A. (2007). Carbohydrate quantity and quality in relation to body mass index. Journal
of the American dietetic association, 107(10), 1768-1780.
Garg, N., Wansink, B., & Inman, J. J. (2007). The influence of incidental affect on consumers'
food intake. Journal of Marketing, 71(1), 194-206.
Giles-Corti, B., & Donovan, R. J. (2002). The relative influence of individual, social and
physical environment determinants of physical activity. Social Science & Medicine,
54(12), 1793-1812.
Goran, M. I. (2001). Metabolic precursors and effects of obesity in children: a decade of
progress, 1990-1999. Am J Clin Nutr, 73(2), 158-171.
Goran, M. I., Ball, G. D. C., & Cruz, M. L. (2003). Obesity and risk of type 2 diabetes and
cardiovascular disease in children and adolescents. The Journal of Clinical
Endocrinology & Metabolism, 88(4), 1417-1427.
Goran, M. I., & Ventura, E. E. (2012). Genetic predisposition and increasing dietary fructose
exposure: The perfect storm for fatty liver disease in Hispanics in the US. Digestive and
Liver Disease, 44(9), 711-713.
Gortmaker, S. L., Must, A., Perrin, J. M., Sobol, A. M., & Dietz, W. H. (1993). Social and
economic consequences of overweight in adolescence and young adulthood. New
England Journal of Medicine, 329(14), 1008-1012.
Greeno, C. G., & Wing, R. R. (1994). Stress-induced eating. Psychological bulletin, 115(3), 444.
Guerrieri, R., Nederkoorn, C., & Jansen, A. (2007). How impulsiveness and variety influence
food intake in a sample of healthy women. Appetite, 48(1), 119-122.
Guerrieri, R., Nederkoorn, C., & Jansen, A. (2008). The interaction between impulsivity and a
varied food environment: its influence on food intake and overweight. International
journal of obesity, 32(4), 708-714.
146

Gunnell, D. J., Frankel, S. J., Nanchahal, K., Peters, T. J., & Smith, G. D. (1998). Childhood
obesity and adult cardiovascular mortality: a 57-y follow-up study based on the Boyd Orr
cohort. The American journal of clinical nutrition, 67(6), 1111-1118.
Handy, S. L., Boarnet, M. G., Ewing, R., & Killingsworth, R. E. (2002). How the built
environment affects physical activity: views from urban planning. American journal of
preventive medicine, 23(2), 64-73.
Hasson, R. E., Adam, T. C., Davis, J. N., Kelly, L. A., Ventura, E. E., Byrd ‐Williams, C. E., . . .
Azen, S. P. (2012). Randomized Controlled Trial to Improve Adiposity, Inflammation,
and Insulin Resistance in Obese African ‐American and Latino Youth. Obesity, 20(4),
811-818.
Hasson, R. E., Adam, T. C., Pearson, J., Davis, J. N., Spruijt-Metz, D., & Goran, M. I. (2013).
Sociocultural and socioeconomic influences on type 2 diabetes risk in overweight/obese
African-American and Latino-American children and adolescents. Journal of obesity,
2013.
Hayes, A. F. (2009). Beyond Baron and Kenny: Statistical mediation analysis in the new
millennium. Communication monographs, 76(4), 408-420.
Hellwig, J. P., Otten, J. J., & Meyers, L. D. (2006). Dietary Reference Intakes:: The Essential
Guide to Nutrient Requirements: National Academies Press.
Hepworth, R., Mogg, K., Brignell, C., & Bradley, B. P. (2010). Negative mood increases
selective attention to food cues and subjective appetite. Appetite, 54(1), 134-142.
Herbert, A., Gerry, N. P., McQueen, M. B., Heid, I. M., Pfeufer, A., Illig, T., . . . Hu, F. B.
(2006). A common genetic variant is associated with adult and childhood obesity.
Science, 312(5771), 279-283.
Herva, A., Laitinen, J., Miettunen, J., Veijola, J., Karvonen, J. T., Läksy, K., & Joukamaa, M.
(2006). Obesity and depression: results from the longitudinal Northern Finland 1966
Birth Cohort Study. International journal of obesity, 30(3), 520-527.
Hill, A. J., Weaver, C. F., & Blundell, J. E. (1991). Food craving, dietary restraint and mood.
Appetite, 17(3), 187-197.
Hill, A. J., Weaver, C. F. L., & Blundell, J. E. (1991). Food craving, dietary restraint and mood.
Appetite, 17(3), 187-197.
Hill, J. O., & Melanson, E. L. (1999). Overview of the determinants of overweight and obesity:
current evidence and research issues. Medicine and science in sports and exercise, 31(11
Suppl), S515-521.
Hill, J. O., Peters, J. C., & Wyatt, H. R. (2009). Using the energy gap to address obesity: a
commentary. Journal of the American dietetic association, 109(11), 1848-1853.
Hinkle, A., & Wu, E. (2003). Communities of color issue briefing paper: addressing the obesity
epidemic—public policies for healthy eating and physical activity environments,
California Adolescent Nutrition and Fitness Program & California Pan Ethnic Health
Network. California Adolescent Nutrition and Fitness Program & California Pan Ethnic
Health Network.
Hortensius, J., Kars, M. C., Wierenga, W. S., Kleefstra, N., Bilo, H. J. G., & van der Bijl, J. J.
(2012). Perspectives of patients with type 1 or insulin-treated type 2 diabetes on self-
monitoring of blood glucose: a qualitative study. BMC Public Health, 12(1), 167.
Howarth, N. C., Saltzman, E., & Roberts, S. B. (2001). Dietary fiber and weight regulation.
Nutrition reviews, 59(5), 129-139.
147

Hsieh, H.-F., & Shannon, S. E. (2005). Three approaches to qualitative content analysis.
Qualitative health research, 15(9), 1277-1288.
Ifland, J. R., Preuss, H. G., Marcus, M. T., Rourke, K. M., Taylor, W. C., Burau, K., . . . Manso,
G. (2009). Refined food addiction: a classic substance use disorder. Medical hypotheses,
72(5), 518-526.
Isen, A. M., & Reeve, J. (2005). The influence of positive affect on intrinsic and extrinsic
motivation: Facilitating enjoyment of play, responsible work behavior, and self-control.
Motivation and Emotion, 29(4), 295-323.
Jebb, S. A. (2007). Dietary determinants of obesity. Obesity reviews, 8(s1), 93-97.
Jenkins, D. J., Wolever, T. M., Collier, G. R., Ocana, A., Rao, A. V., Buckley, G., . . .
Thompson, L. U. (1987). Metabolic effects of a low-glycemic-index diet. The American
journal of clinical nutrition, 46(6), 968-975.
Jenkins, D. J. A., & Jenkins, A. L. (1985). Dietary fiber and the glycemic response. Experimental
Biology and Medicine, 180(3), 422-431.
Johannsen, D. L., Johannsen, N. M., & Specker, B. L. (2006). Influence of parents’ eating
behaviors and child feeding practices on children's weight status. Obesity, 14(3), 431-
439.
Johnson, R. K., Appel, L. J., Brands, M., Howard, B. V., Lefevre, M., Lustig, R. H., . . . Wylie-
Rosett, J. (2009). Dietary sugars intake and cardiovascular health a scientific statement
from the american heart association. Circulation, 120(11), 1011-1020.
Jones, M. I., Greenfield, S. M., Bray, E. P., Baral-Grant, S., Hobbs, F. D. R., Holder, R., . . .
Williams, B. (2012). Patients’ experiences of self-monitoring blood pressure and self-
titration of medication: the TASMINH2 trial qualitative study. British Journal of General
Practice, 62(595), e135-e142.
Kahn, S. E., Hull, R. L., & Utzschneider, K. M. (2006). Mechanisms linking obesity to insulin
resistance and type 2 diabetes. Nature, 444(7121), 840-846.
Kant, A. K., Graubard, B. I., Schatzkin, A., & Ballard-Barbash, R. (1995). Proportion of energy
intake from fat and subsequent weight change in the NHANES I Epidemiologic Follow-
up Study. The American journal of clinical nutrition, 61(1), 11-17.
Karnehed, N., Rasmussen, F., Hemmingsson, T., & Tynelius, P. (2006). Obesity and attained
education: cohort study of more than 700,000 Swedish men. Obesity, 14(8), 1421-1428.
Kimm, S. Y. S. (1995). The role of dietary fiber in the development and treatment of childhood
obesity. Pediatrics, 96(5), 1010-1014.
King, D. E., Mainous, A. G., & Lambourne, C. A. (2012). Trends in dietary fiber intake in the
United States, 1999-2008. Journal of the Academy of Nutrition and Dietetics, 112(5),
642-648.
Kirk, S. F. L., Penney, T. L., & McHugh, T. L. (2010). Characterizing the obesogenic
environment: the state of the evidence with directions for future research. Obesity
reviews, 11(2), 109-117.
Koestner, R., Otis, N., Powers, T. A., Pelletier, L., & Gagnon, H. (2008). Autonomous
motivation, controlled motivation, and goal progress. Journal of personality, 76(5), 1201-
1230.
Krebs, P., & Duncan, D. T. (2015). Health App Use Among US Mobile Phone Owners: A
National Survey. JMIR mHealth and uHealth, 3(4).
Kristeller, J. L., & Wolever, R. Q. (2010). Mindfulness-based eating awareness training for
treating binge eating disorder: the conceptual foundation. Eating Disorders, 19(1), 49-61.
148

Kristenssson, E., Sundqvist, M., Astin, M., Kjerling, M., Mattsson, H., de la Cour, C. D., . . .
Lindström, E. (2006). Acute psychological stress raises plasma ghrelin in the rat.
Regulatory peptides, 134(2), 114-117.
Kuczmarski, R. J., Ogden, C. L., Guo, S. S., Grummer-Strawn, L. M., Flegal, K. M., Mei, Z., . . .
Johnson, C. L. (2002). 2000 CDC Growth Charts for the United States: methods and
development. Vital and health statistics. Series 11, Data from the national health
survey(246), 1-190.
Lafay, L., Thomas, F., Mennen, L., Charles, M. A., Eschwege, E., & Borys, J. M. (2001). Gender
differences in the relation between food cravings and mood in an adult community:
Results from the Fleurbaix Laventie Ville Sante study. International Journal of Eating
Disorders, 29(2), 195-204.
Laran, J., & Janiszewski, C. (2011). Work or fun? How task construal and completion influence
regulatory behavior. Journal of Consumer Research, 37(6), 967-983.
Larson, D. E., Hunter, G. R., Williams, M. J., Kekes-Szabo, T., Nyikos, I., & Goran, M. I.
(1996). Dietary fat in relation to body fat and intraabdominal adipose tissue: a cross-
sectional analysis. The American journal of clinical nutrition, 64(5), 677-684.
Latner, J. D., & Wilson, G. T. (2002). Self-monitoring and the assessment of binge eating.
Behavior Therapy, 33(3), 465-477.
Lenhart, A. (2015). Teens, Social Media, and Technology Overview 2015: Pew Research Center.
Lewis, F. M., Morisky, D. E., & Flynn, B. S. (1978). A test of the construct validity of health
locus of control: effects on self-reported compliance for hypertensive patients. Health
Education & Behavior, 6(1), 138-148.
Li, J., & Wang, Y. (2008). Tracking of dietary intake patterns is associated with baseline
characteristics of urban low-income African-American adolescents. The Journal of
nutrition, 138(1), 94-100.
Linder, B., & Imperatore, G. (2013). Research Updates on Type 2 Diabetes in Children. NASN
School Nurse.
Lissner, L., Lindroos, A. K., & Sjöström, L. (1998). Swedish obese subjects (SOS): an obesity
intervention study with a nutritional perspective. European journal of clinical nutrition,
52(5), 316-322.
Liu, S., Willett, W. C., Manson, J. E., Hu, F. B., Rosner, B., & Colditz, G. (2003). Relation
between changes in intakes of dietary fiber and grain products and changes in weight and
development of obesity among middle-aged women. The American journal of clinical
nutrition, 78(5), 920-927.
Logan, G. D., Schachar, R. J., & Tannock, R. (1997). Impulsivity and inhibitory control.
Psychological science, 8(1), 60-64.
Lovasi, G. S., Hutson, M. A., Guerra, M., & Neckerman, K. M. (2009). Built environments and
obesity in disadvantaged populations. Epidemiologic reviews, mxp005.
Low, Y. Q., Lacy, K., & Keast, R. (2014). The Role of Sweet Taste in Satiation and Satiety.
Nutrients, 6(9), 3431-3450.
Lowe, M. R., & Butryn, M. L. (2007). Hedonic hunger: a new dimension of appetite? Physiology
& behavior, 91(4), 432-439.
Ludwig, D. S. (2000). Dietary glycemic index and obesity. The Journal of nutrition, 130(2),
280S-283S.
Lyons, K. S., & Sayer, A. G. (2005). Longitudinal dyad models in family research. Journal of
Marriage and Family, 67(4), 1048-1060.
149

Macht, M. (2008). How emotions affect eating: a five-way model. Appetite, 50(1), 1-11.
Macht, M., & Simons, G. (2000). Emotions and eating in everyday life. Appetite, 35(1), 65-71.
MacKinnon, D. P., Fairchild, A. J., & Fritz, M. S. (2007). Mediation analysis. Annual review of
psychology, 58, 593.
MacKinnon, D. P., Lockwood, C. M., & Williams, J. (2004). Confidence limits for the indirect
effect: Distribution of the product and resampling methods. Multivariate behavioral
research, 39(1), 99-128.
Maddock, J. (2004). The relationship between obesity and the prevalence of fast food
restaurants: state-level analysis. American journal of health promotion, 19(2), 137-143.
Major, O. (2015). The effect of a low refined and added sugar diet on mood and fatigue levels.
Cardiff Metropolitan University.
Martin, C. K., O'neil, P. M., & Pawlow, L. (2006). Changes in food cravings during low ‐calorie
and very ‐low ‐calorie diets. Obesity, 14(1), 115-121.
Mendoza, J. A., Drewnowski, A., Cheadle, A., & Christakis, D. A. (2006). Dietary energy
density is associated with selected predictors of obesity in US Children. The Journal of
nutrition, 136(5), 1318-1322.
Monasta, L., Batty, G. D., Cattaneo, A., Lutje, V., Ronfani, L., Van Lenthe, F. J., & Brug, J.
(2010). Early ‐life determinants of overweight and obesity: a review of systematic
reviews. Obesity reviews, 11(10), 695-708.
Montonen, J., Knekt, P., Järvinen, R., Aromaa, A., & Reunanen, A. (2003). Whole-grain and
fiber intake and the incidence of type 2 diabetes. The American journal of clinical
nutrition, 77(3), 622-629.
Montoya, A. K., & Hayes, A. F. (2015). Two condition within-participant statistical mediation
analysis: A path-analytic framework. Manuscript submitted for publication.
Morland, K., Wing, S., & Roux, A. D. (2002). The contextual effect of the local food
environment on residents' diets: the atherosclerosis risk in communities study. American
journal of public health, 92(11), 1761-1768.
Morland, K., Wing, S., Roux, A. D., & Poole, C. (2002). Neighborhood characteristics
associated with the location of food stores and food service places. American journal of
preventive medicine, 22(1), 23-29.
Mossavar-Rahmani, Y., Henry, H., Rodabough, R., Bragg, C., Brewer, A., Freed, T., . . .
Vosburg, S. (2004). Additional self-monitoring tools in the dietary modification
component of The Women's Health Initiative. J Am Diet Assoc, 104(1), 76-85. doi:
10.1016/j.jada.2003.10.017
S0002822303014482 [pii]
Mossavar-Rahmani, Y., Henry, H., Rodabough, R., Bragg, C., Brewer, A., Freed, T., . . .
Vosburg, S. (2004). Additional self-monitoring tools in the dietary modification
component of the Women’s Health Initiative. Journal of the American dietetic
association, 104(1), 76-85.
Munsch, S., Michael, T., Biedert, E., Meyer, A., & Margraf, J. (2008). Negative mood induction
and unbalanced nutrition style as possible triggers of binges in binge eating disorder
(BED). Eating and Weight Disorders-Studies on Anorexia, Bulimia and Obesity, 13(1),
22-29.
Muraven, M., & Baumeister, R. F. (2000). Self-regulation and depletion of limited resources:
Does self-control resemble a muscle? Psychological bulletin, 126(2), 247.
150

Murphy, C. M., Stojek, M. K., & MacKillop, J. (2014). Interrelationships among impulsive
personality traits, food addiction, and body mass index. Appetite, 73, 45-50.
Mustillo, S., Worthman, C., Erkanli, A., Keeler, G., Angold, A., & Costello, E. J. (2003).
Obesity and psychiatric disorder: developmental trajectories. Pediatrics, 111(4), 851-859.
National Cancer Institute. Learn More about Energy Adjustment. Dietary Assessment Primer.
2015, from http://dietassessmentprimer.cancer.gov/learn/adjustment.html
Nesbitt, S. D., Ashaye, M. O., Stettler, N., Sorof, J. M., Goran, M. I., Parekh, R., & Falkner, B.
E. (2003). Overweight as a risk factor in children: a focus on ethnicity. Ethnicity &
disease, 14(1), 94-110.
O'Reilly, G. A., Belcher, B. R., Davis, J. N., Martinez, L. T., Huh, J., Antunez ‐Castillo, L., . . .
Spruijt ‐Metz, D. (2015). Effects of high ‐sugar and high ‐fiber meals on physical activity
behaviors in Latino and African American adolescents. Obesity, 23(9), 1886-1894.
O’Reilly, G. A., & Spruijt-Metz, D. (2013). Current mHealth technologies for physical activity
assessment and promotion. American journal of preventive medicine, 45(4), 501-507.
Ogden, C. L., Carroll, M. D., Kit, B. K., & Flegal, K. M. (2012). Prevalence of obesity and
trends in body mass index among US children and adolescents, 1999-2010. JAMA,
307(5), 483-490.
Ogden, C. L., Carroll, M. D., Kit, B. K., & Flegal, K. M. (2014). Prevalence of childhood and
adult obesity in the United States, 2011-2012. JAMA, 311(8), 806-814. doi: 1832542 [pii]
10.1001/jama.2014.732
Oliver, G., Wardle, J., & Gibson, E. L. (2000). Stress and food choice: a laboratory study.
Psychosomatic medicine, 62(6), 853-865.
Onyike, C. U., Crum, R. M., Lee, H. B., Lyketsos, C. G., & Eaton, W. W. (2003). Is obesity
associated with major depression? Results from the Third National Health and Nutrition
Examination Survey. American journal of epidemiology, 158(12), 1139-1147.
Ouwens, M. A., van Strien, T., & van Leeuwe, J. F. J. (2009). Possible pathways between
depression, emotional and external eating. A structural equation model. Appetite, 53(2),
245-248.
Pan, L., Blanck, H. M., Sherry, B., Dalenius, K., & Grummer-Strawn, L. M. (2012). Trends in
the prevalence of extreme obesity among US preschool-aged children living in low-
income families, 1998-2010. JAMA, 308(24), 2563-2565.
Parkes, A., & Kearns, A. (2006). The multi-dimensional neighbourhood and health: a cross-
sectional analysis of the Scottish Household Survey, 2001. Health & Place, 12(1), 1-18.
Parylak, S. L., Koob, G. F., & Zorrilla, E. P. (2011). The dark side of food addiction. Physiology
& behavior, 104(1), 149-156.
Pasman, W., Saris, W., Wauters, M., & Westerterp-Plantenga, M. (1997). Effect of one week of
fibre supplementation on hunger and satiety ratings and energy intake. Appetite, 29(1),
77-87.
Patel, K. A., & Schlundt, D. G. (2001). Impact of moods and social context on eating behavior.
Appetite, 36(2), 111-118.
Pereira, M. A., Jacobs, D. R., Pins, J. J., Raatz, S. K., Gross, M. D., Slavin, J. L., & Seaquist, E.
R. (2002). Effect of whole grains on insulin sensitivity in overweight hyperinsulinemic
adults. The American journal of clinical nutrition, 75(5), 848-855.
Petry, N. M., Barry, D., Pietrzak, R. H., & Wagner, J. A. (2008). Overweight and obesity are
associated with psychiatric disorders: results from the National Epidemiologic Survey on
Alcohol and Related Conditions. Psychosomatic medicine, 70(3), 288-297.
151

Plachta ‐Danielzik, S., Landsberg, B., Bosy ‐Westphal, A., Johannsen, M., Lange, D., & Müller,
M. J. (2008). Energy Gain and Energy Gap in Normal ‐weight Children: Longitudinal
Data of the KOPS. Obesity, 16(4), 777-783.
Poirier, P., Giles, T. D., Bray, G. A., Hong, Y., Stern, J. S., Pi-Sunyer, F. X., & Eckel, R. H.
(2006). Obesity and cardiovascular disease pathophysiology, evaluation, and effect of
weight loss. Arteriosclerosis, thrombosis, and vascular biology, 26(5), 968-976.
Polivy, J. (1996). Psychological consequences of food restriction. Journal of the American
dietetic association, 96(6), 589-592.
Potter, J. G., Coffman, K. P., Reid, R. L., Krall, J. M., & Albrink, M. J. (1981). Effect of test
meals of varying dietary fiber content on plasma insulin and glucose response. The
American journal of clinical nutrition, 34(3), 328-334.
Preacher, K. J., Rucker, D. D., & Hayes, A. F. (2007). Addressing moderated mediation
hypotheses: Theory, methods, and prescriptions. Multivariate behavioral research, 42(1),
185-227.
Qi, L., & Cho, Y. A. (2008). Gene-environment interaction and obesity. Nutrition reviews,
66(12), 684-694.
Reid, M., & Hammersley, R. (1998). The effects of blind substitution of aspartame-sweetened
for sugar-sweetened soft drinks on appetite and mood. British Food Journal, 100(5), 254-
259.
Reilly, J. J., Methven, E., McDowell, Z. C., Hacking, B., Alexander, D., Stewart, L., & Kelnar,
C. J. H. (2003). Health consequences of obesity. Archives of disease in childhood, 88(9),
748-752.
Renehan, A. G., Tyson, M., Egger, M., Heller, R. F., & Zwahlen, M. (2008). Body-mass index
and incidence of cancer: a systematic review and meta-analysis of prospective
observational studies. The Lancet, 371(9612), 569-578.
Rennie, K. L., Johnson, L., & Jebb, S. A. (2005). Behavioural determinants of obesity. Best
Practice & Research Clinical Endocrinology & Metabolism, 19(3), 343-358.
Reynolds, K. D., & Spruijt-Metz, D. (2006). Translational research in childhood obesity
prevention. Evaluation & the health professions, 29(2), 219-245.
Rhee, K. E., Phelan, S., & McCaffery, J. (2012). Early determinants of obesity: genetic,
epigenetic, and in utero influences. International journal of pediatrics, 2012.
Roberts, S. B. (2000). High-glycemic index foods, hunger, and obesity: is there a connection?
Nutrition reviews, 58(6), 163-169.
Rolls, B. J., Ello-Martin, J. A., & Tohill, B. C. (2004). What can intervention studies tell us
about the relationship between fruit and vegetable consumption and weight management?
Nutrition reviews, 62(1), 1-17.
Rose, D., & Richards, R. (2004). Food store access and household fruit and vegetable use among
participants in the US Food Stamp Program. Public health nutrition, 7(08), 1081-1088.
Rosenkranz, R. R., & Dzewaltowski, D. A. (2008). Model of the home food environment
pertaining to childhood obesity. Nutrition reviews, 66(3), 123-140.
Saelens, B. E., Sallis, J. F., & Frank, L. D. (2003). Environmental correlates of walking and
cycling: findings from the transportation, urban design, and planning literatures. Annals
of behavioral medicine, 25(2), 80-91.
Salmon, S. J., Fennis, B. M., de Ridder, D. T., Adriaanse, M. A., & de Vet, E. (2014). Health on
impulse: when low self-control promotes healthy food choices. Health Psychology, 33(2),
103.
152

Samaha, F. F., Iqbal, N., Seshadri, P., Chicano, K. L., Daily, D. A., McGrory, J., . . . Stern, L.
(2003). A low-carbohydrate as compared with a low-fat diet in severe obesity. New
England Journal of Medicine, 348(21), 2074-2081.
Saris, W. H. M. (2003). Sugars, energy metabolism, and body weight control. The American
journal of clinical nutrition, 78(4), 850S-857S.
Schellekens, H., Finger, B. C., Dinan, T. G., & Cryan, J. F. (2012). Ghrelin signalling and
obesity: at the interface of stress, mood and food reward. Pharmacology & therapeutics,
135(3), 316-326.
Sengupta, J., & Zhou, R. (2007). Understanding impulsive eaters' choice behaviors: The
motivational influences of regulatory focus. Journal of Marketing Research, 44(2), 297-
308.
Shapiro, S. L., Carlson, L. E., Astin, J. A., & Freedman, B. (2006). Mechanisms of mindfulness.
Journal of clinical psychology, 62(3), 373-386.
Shearrer, G. E., O’Reilly, G.A., Belcher, B., Daniels, M., Goran, M., Spruijt-Metz, D., Davis,
J.N. . (2015). The impact of sugar sweetened beverage intake on hunger and satiety in
minority adolescents. . (under review, Appetite).
Silva, M. N., Vieira, P. N., Coutinho, S. l. R., Minderico, C. u. S., Matos, M. G., Sardinha, L. s.
B., & Teixeira, P. J. (2010). Using self-determination theory to promote physical activity
and weight control: a randomized controlled trial in women. Journal of behavioral
medicine, 33(2), 110-122.
Simon, G. E., Von Korff, M., Saunders, K., Miglioretti, D. L., Crane, P. K., van Belle, G., &
Kessler, R. C. (2006). Association between obesity and psychiatric disorders in the US
adult population. Archives of general psychiatry, 63(7), 824-830.
Singer, J. D., & Willett, J. B. (2003). Applied longitudinal data analysis: Modeling change and
event occurrence: Oxford university press.
Singh, A. S., Mulder, C., Twisk, J. W. R., Van Mechelen, W., & Chinapaw, M. J. M. (2008).
Tracking of childhood overweight into adulthood: a systematic review of the literature.
Obesity reviews, 9(5), 474-488.
Skinner, A. C., & Skelton, J. A. (2014). Prevalence and trends in obesity and severe obesity
among children in the United States, 1999-2012. JAMA pediatrics, 168(6), 561-566.
Slavin, J. L. (2005). Dietary fiber and body weight. Nutrition, 21(3), 411-418.
Slyper, A. H. (2013). The influence of carbohydrate quality on cardiovascular disease, the
metabolic syndrome, type 2 diabetes, and obesity–an overview. Journal of Pediatric
Endocrinology and Metabolism, 26(7-8), 617-629.
Smith, A. (2015). US Smartphone Use in 2015. Pew Research Center.
Song, Y. J., Sawamura, M., Ikeda, K., Igawa, S., & Yamori, Y. (2000). Soluble Dietary Fibre
Improves Insulin Sensitivity By Increasing Muscle Glut ‐4 Content In Stroke‐Prone
Spontaneously Hypertensive Rats. Clinical and Experimental Pharmacology and
Physiology, 27(1 ‐2), 41-45.
Sorof, J., & Daniels, S. (2002). Obesity hypertension in children a problem of epidemic
proportions. Hypertension, 40(4), 441-447.
Srinivasan, S. R., Bao, W., Wattigney, W. A., & Berenson, G. S. (1996). Adolescent overweight
is associated with adult overweight and related multiple cardiovascular risk factors: the
Bogalusa Heart Study. Metabolism, 45(2), 235-240.
153

Stanhope, K. L., Schwarz, J. M., & Havel, P. J. (2013). Adverse metabolic effects of dietary
fructose: results from the recent epidemiological, clinical, and mechanistic studies. Curr
Opin Lipidol. doi: 10.1097/MOL.0b013e3283613bca
Story, M., Neumark-Sztainer, D., & French, S. (2002). Individual and environmental influences
on adolescent eating behaviors. Journal of the American dietetic association, 102(3),
S40-S51.
Swallen, K. C., Reither, E. N., Haas, S. A., & Meier, A. M. (2005). Overweight, obesity, and
health-related quality of life among adolescents: the National Longitudinal Study of
Adolescent Health. Pediatrics, 115(2), 340-347.
Swinburn, B. A., Sacks, G., Lo, S. K., Westerterp, K. R., Rush, E. C., Rosenbaum, M., . . . Butte,
N. F. (2009). Estimating the changes in energy flux that characterize the rise in obesity
prevalence. The American journal of clinical nutrition, 89(6), 1723-1728.
Tate, D. F., Wing, R. R., & Winett, R. A. (2001). Using Internet technology to deliver a
behavioral weight loss program. JAMA, 285(9), 1172-1177.
Te Morenga, L., Mallard, S., & Mann, J. (2013). Dietary sugars and body weight: systematic
review and meta-analyses of randomised controlled trials and cohort studies. Bmj, 346,
e7492.
Telch, C. F., & Agras, W. S. (1996). Do emotional states influence binge eating in the obese?
International Journal of Eating Disorders, 20(3), 271-279.
Terry, P. C., Lane, A. M., Lane, H. J., & Keohane, L. (1999). Development and validation of a
mood measure for adolescents. Journal of Sports Sciences, 17(11), 861-872.
Thanos, P. K., Michaelides, M., Piyis, Y. K., Wang, G. J., & Volkow, N. D. (2008). Food
restriction markedly increases dopamine D2 receptor (D2R) in a rat model of obesity as
assessed with in ‐vivo μPET imaging ([11C] raclopride) and in ‐vitro ([3H] spiperone)
autoradiography. Synapse, 62(1), 50-61.
Thayer, R. E. (1987). Energy, tiredness, and tension effects of a sugar snack versus moderate
exercise. Journal of Personality and Social Psychology, 52(1), 119.
Thayer, R. E. (1989). The biopsychology of mood and arousal: Oxford University Press.
Thayer, R. E. (2003). Calm energy: How people regulate mood with food and exercise: Oxford
University Press.
Thayer, R. E., & McNally, R. J. (1992). The Biopsychology of Mood and Arousal (Vol. 5).
Thompson, T. G., & Veneman, A. M. (2005). Dietary guidelines for Americans 2005. United
States Department of Health and Human Services and United States Department of
Agriculture.
Thorpe, K. E., Florence, C. S., Howard, D. H., & Joski, P. (2004). The impact of obesity on
rising medical spending. HEALTH AFFAIRS-MILLWOOD VA THEN BETHESDA MA-,
23, 283-283.
Toledo-Corral, C. M., Davis, J. N., Alderete, T. L., Weigensberg, M. J., Ayala, C. T., Li, Y., . . .
Goran, M. I. (2011). Subclinical atherosclerosis in Latino youth: progression of carotid
intima-media thickness and its relationship to cardiometabolic risk factors. The Journal of
pediatrics, 158(6), 935-940.
Trevino, R. P., Marshall, R. M., Jr., Hale, D. E., Rodriguez, R., Baker, G., & Gomez, J. (1999).
Diabetes risk factors in low-income Mexican-American children. Diabetes Care, 22(2),
202-207.
Tsai, A. G., Williamson, D. F., & Glick, H. A. (2011). Direct medical cost of overweight and
obesity in the USA: a quantitative systematic review. Obesity reviews, 12(1), 50-61.
154

Valente, T. W., Fujimoto, K., Chou, C.-P., & Spruijt-Metz, D. (2009). Adolescent affiliations
and adiposity: a social network analysis of friendships and obesity. Journal of Adolescent
Health, 45(2), 202-204.
Van Gaal, L. F., Mertens, I. L., & Christophe, E. (2006). Mechanisms linking obesity with
cardiovascular disease. Nature, 444(7121), 875-880.
Videon, T. M., & Manning, C. K. (2003). Influences on adolescent eating patterns: the
importance of family meals. Journal of Adolescent Health, 32(5), 365-373.
Volek, J. S., VanHeest, J. L., & Forsythe, C. E. (2005). Diet and exercise for weight loss. Sports
Medicine, 35(1), 1-9.
Volkow, N. D., Wang, G.-J., Fowler, J. S., & Telang, F. (2008). Overlapping neuronal circuits in
addiction and obesity: evidence of systems pathology. Philosophical Transactions of the
Royal Society B: Biological Sciences, 363(1507), 3191-3200.
Volkow, N. D., & Wise, R. A. (2005). How can drug addiction help us understand obesity?
Nature neuroscience, 8(5), 555-560.
Wadden, T. A., Butryn, M. L., & Wilson, C. (2007). Lifestyle modification for the management
of obesity. Gastroenterology, 132(6), 2226-2238. doi: S0016-5085(07)00578-1 [pii]
10.1053/j.gastro.2007.03.051
Wallston, B. S., Wallston, K. A., Kaplan, G. D., & Maides, S. A. (1976). Development and
validation of the health locus of control (HLC) scale. Journal of consulting and clinical
psychology, 44(4), 580.
Wallston, K. A., Wallston, B. S., & DeVellis, R. (1978). Development of the multidimensional
health locus of control (MHLC) scales. Health Education & Behavior, 6(1), 160-170.
Wang, Y. C., Gortmaker, S. L., Sobol, A. M., & Kuntz, K. M. (2006). Estimating the energy gap
among US children: a counterfactual approach. Pediatrics, 118(6), e1721-e1733.
Wen, X., Gillman, M. W., Rifas-Shiman, S. L., Sherry, B., Kleinman, K., & Taveras, E. M.
(2012). Decreasing prevalence of obesity among young children in Massachusetts from
2004 to 2008. Pediatrics, 129(5), 823-831.
White, M. A., Whisenhunt, B. L., Williamson, D. A., Greenway, F. L., & Netemeyer, R. G.
(2002). Development and validation of the food ‐craving inventory. Obesity Research,
10(2), 107-114.
Whiteside, S. P., & Lynam, D. R. (2001). The five factor model and impulsivity: Using a
structural model of personality to understand impulsivity. Personality and individual
differences, 30(4), 669-689.
Who, J., & Consultation, F. A. O. E. (2003). Diet, nutrition and the prevention of chronic
diseases. WHO technical report series(916), 1-60.
Wideman, C. H., Nadzam, G. R., & Murphy, H. M. (2005). Implications of an animal model of
sugar addiction, withdrawal and relapse for human health. Nutritional neuroscience, 8(5-
6), 269-276.
Willett, W., Manson, J., & Liu, S. (2002). Glycemic index, glycemic load, and risk of type 2
diabetes. Am J Clin Nutr, 76(1), 274S-280S.
Williams, E., Stewart-Knox, B., Helander, A., McConville, C., Bradbury, I., & Rowland, I.
(2006). Associations between whole-blood serotonin and subjective mood in healthy
male volunteers. Biological psychology, 71(2), 171-174.
Williams, G. C., Grow, V. M., Freedman, Z. R., Ryan, R. M., & Deci, E. L. (1996). Motivational
predictors of weight loss and weight-loss maintenance. Journal of personality and social
psychology, 70(1), 115.
155

Williams, G. C., Ryan, R. M., & Deci, E. L. Self-Determination Theory Perceived Competence
Scales. Retrieved July, 2014, from http://www.selfdeterminationtheory.org/perceived-
competence-scales/.
Williams, G. C., Ryan, R. M., & Deci, E. L. Self-Determination Theory Treatment Self-
Regulation Questionnaire. 2015, from http://selfdeterminationtheory.org/
Williams, J., Wake, M., Hesketh, K., Maher, E., & Waters, E. (2005). Health-related quality of
life of overweight and obese children. JAMA, 293(1), 70-76.
Wing, R. R. (1998). Behavioral approaches to the treatment of obesity. Handbook of obesity,
855-873.
Wolk, A., Gridley, G., Svensson, M., Nyrén, O., McLaughlin, J. K., Fraumeni, J. F., & Adami,
H.-O. (2001). A prospective study of obesity and cancer risk (Sweden). Cancer causes &
control, 12(1), 13-21.
Woods, S. C. (2004). Gastrointestinal satiety signals I. An overview of gastrointestinal signals
that influence food intake. American Journal of Physiology-Gastrointestinal and Liver
Physiology, 286(1), G7-G13.
Wurtman, R. J., & Wurtman, J. J. (1995). Brain serotonin, carbohydrate ‐craving, obesity and
depression. Obesity Research, 3(S4), 477S-480S.
Yanovski, S. Z., & Yanovski, J. A. (2011). Obesity prevalence in the United States—up, down,
or sideways? New England Journal of Medicine, 364(11), 987-989.
Yon, B. A., Johnson, R. K., Harvey-Berino, J., Gold, B. C., & Howard, A. B. (2007). Personal
digital assistants are comparable to traditional diaries for dietary self-monitoring during a
weight loss program. J Behav Med, 30(2), 165-175. doi: 10.1007/s10865-006-9092-1
Zhang, Z., Gillespie, C., Welsh, J. A., Hu, F. B., & Yang, Q. (2015). Usual intake of added
sugars and lipid profiles among the US adolescents: National Health and Nutrition
Examination Survey, 2005–2010. Journal of Adolescent Health, 56(3), 352-359.
Zillmann, D. (1988). Mood management: Using entertainment to full advantage.
Communication, social cognition, and affect, 31, 147-171.

156

Appendix A. FAME study ad libitum snack tray contents
Food or drink item Amount provided on snack tray (g or mL)
Cheetos Flamin Hot Crunchy 28.3 g
Milk Chocolate M&Ms fun size 18 g
100 Calorie Packs Oreo Thin Crisps 23 g
Hostess Cup Cakes (Frosted Chocolate Cake
with Creamy Filling)
50 g
Chewy Granola Bar Chocolate Chip 24 g
Fiber One Chewy Bars: Oats and Chocolate 40 g
Nutri-Grain Cereal Bars: Blueberry 37 g
Pringles Sour Cream and Onion 23 g
Snyder’s of Hanover 100 Calorie Pack: Mini
Pretzels
26 g
Wheat Thins Original: Baked Snack Crackers 35 g
Mott’s Natural Apple Sauce (no added sugar) 111 g
Del Monte: Diced Peaches 106 g
Handi-Snacks: Dunk ‘ems: Outrageously
Cheesey
27 g
Sun-Maid Natural California Raisins 14.1 g
Ready Pac Cool Cuts: Carrots with Ranch Dip 63.7 g
Disney garden: Sweet Red Apple Slices 79.2 g
Orowheat Whole Grain 100% Whole Wheat
Sandwich Style Bread (2 slices)
62 g
Weber’s Enriched Bread Sliced Small (2
slices)
50 g
Buddig: The Original Deli Thin Beef (Smoked,
Chopped, Pressed)
56 g
Buddig: The Original Deli Thin Turkey
(Smoked, Chopped, Pressed)
56 g
Heinz Mayonnaise (2 packets)
Heinz Mustard (2 packets)
Mango (150 g +/- 2 g )
Diet 7-Up Lemon Lime 240 mL
Diet Coke 237 mL
Dr. Pepper 240 mL
Sprite 237 mL
Kirkland Organic Chocolate Reduced Fat Milk
with Vitamins A and D
244 mL

157

Appendix B. Study 3 Interview Protocol
Introduction:
Thank you for coming, I appreciate your taking the time out to speak with me today. My
name is Gillian O’Reilly, I’m a PhD student at the University of Southern California, and
I’m going to interview you today as part of a project that I’m completing for my
dissertation.

I want to remind you about the project and what my goal is by talking to you today. I am
interested in talking to people who track their diet, exercise, stress, sleep, and/or weight
to learn about what motivates them to track these things and how tracking these things
helps them with their personal health goals. By interviewing people like you, I am hoping
to gather information that could potentially help other people learn how to start and stick
with tracking similar health metrics. This interview will take about one hour.

Confidentiality:
My goal is to gather information that could help people who want to learn about
themselves and improve their health through self-tracking . While I want to listen to and
tell others about your experiences with self-tracking, I want to assure you that your
identity and all information you give me is strictly confidential. I will not be reporting
your name to anyone. I will not attach your name to any comments you make and will
change identifying details if I use your information in any report.

Tape recording:
I will be tape recording this interview so that I can focus on our discussion and so I don’t
miss anything important. Is that okay with you? We can stop recording the interview at
any time, just let me know if that’s what you want to do.

Do you have any questions about the project or anything that I’ve told you so far?
Interview questions:
I’m going to ask you questions now about your experiences with self-tracking your health
metrics and health behaviors.

1. Can you please describe your self-tracking habits?

Probe: What kinds of behaviors and/or health metrics do you self-track?
Probe: How often do you self-track these behaviors and/or health metrics?

Follow-up question: How long have you been self-tracking for?

2. What kinds of tools do you use to self-track?

Probe: Do you use smartphone apps? Which ones? How long have you used them for?
Probe: Do you use websites? Which ones? How long have you used them for?
158

Probe: Do you use any commercial products (ex: Fitbit, Nike fuel band)? Which ones?
How long have you used it/them for?

Follow-up question: Do you tend to stick with a self-tracking tool for a while?

3. What do you like about the tools you use to self-track?

4. What do you dislike about the tools you use to self-track?

5. If you could change something about the tools you use to self-track, what would you
change?

6. Why did you start self-tracking [behavior/health metric]?

Probe: Was there an event or health issue that motivated you to start self-tracking?
Probe: Was there a person or people who motivated you to start self-tracking?
Probe: Was there a product that you wanted to try (ex: Nike Fuel band, Livestrong diet
tracking app, etc.)

7. What motivates you to stick with self-tracking?

Probe: Tell me what you like about self-tracking.
Probe (if appropriate): Tell me about the social aspect of self-tracking with the tools you
use.

8. What have you learned about yourself through self-tracking?

9. Tell me about how self-tracking helps you with your health goals

Probe: Has self-tracking helped you change something about your health? Tell me more
about that.
Follow-up: Has self-tracking helped you maintain that change? Tell me more about that.

10. What are your long-term [diet, exercise, stress, sleep, weight] goals?

11. Tell me how you think self-tracking will help you reach those goals.

12. Is there anything else about self-tracking you’d like to tell me that I haven’t asked you
about yet?

Conclusion:
That’s the end of the interview. Thank you again for taking the time to talk with me today. This
has been very informative. If you think of anything else later on that you’d like to tell me please
feel free to contact me by email.

159

Appendix C. Study 3 Questionnaires

A. Exercise Self-Regulation Questionnaire

I try to exercise on a regular basis:

Not At All True Somewhat True Very True
A1. Because I would feel bad
about myself if I did not
1 2 3 4 5 6 7
A2. Because others would be
angry at me if I did not
1 2 3 4 5 6 7
A3. Because I enjoy exercising 1 2 3 4 5 6 7
A4. Because I would feel like a
failure if I did not
1 2 3 4 5 6 7
A5. Because I feel like it’s the best
way to help myself
1 2 3 4 5 6 7
A6. Because people would think
I’m a weak person if I did not
1 2 3 4 5 6 7
A7. Because I feel like I have no
choice about exercising; others
make me do it.
1 2 3 4 5 6 7
A8. Because it is a challenge to
accomplish my goal
1 2 3 4 5 6 7
A9. Because I believe exercise
helps me feel better
1 2 3 4 5 6 7
A10. Because it’s fun 1 2 3 4 5 6 7
A11. Because I worry that I would
get in trouble with others if I did
not
1 2 3 4 5 6 7
A12. Because it feels important to
me personally to accomplish this
goal
1 2 3 4 5 6 7
A13. Because I feel guilty if I do
not exercise regularly
1 2 3 4 5 6 7
A14. Because I want others to
acknowledge that I am doing what
I have been told I should do
1 2 3 4 5 6 7
A15. Because it is interesting to
see my own improvement
1 2 3 4 5 6 7
A16. Because feeling healthier is
an important value for me.
1 2 3 4 5 6 7

160

B. Treatment Self-Regulation Questionnaire - Diet

The following question relates to the reasons why you would either start eating a healthier diet
or continue to do so. Different people have different reasons for doing that, and we want to know
how true each of the following reasons is for you. All 15 response are to the same question.

Please indicate the extent to which each reason is true for you, using the following 7-point scale:

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

The reason I would eat a healthy diet is:

B1. Because I feel that I was to take responsibility for my own health.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B2. Because I would feel guilty or ashamed of myself if I did not eat a healthy diet.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B3. Because I personally believe it is the best thing for my health.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B4. Because others would be upset with me if I did not.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B5. I really don’t think about it.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

161

B6. Because I have carefully thought about it and believe it is very important for many
aspects of my life.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B7. Because I would feel bad about myself if I did not eat a healthy diet.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B8. Because it is an important choice I really want to make.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B9. Because I feel pressure from others to do so.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B10. Because it is easier to do what I am told than think about it.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B11. Because it is consistent with my life goals.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B12. Because I want others to approve of me.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B13. Because it is very important for being as healthy as possible.

1 2 3 4 5 6 7
162

Not at
all true
Somewha
t true
Very
true

B14. Because I want others to see I can do it.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

B15. I don’t really know why.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

C. Multidimensional Health Locus of Control

Each item below is a belief statement about your medical condition with which you may agree or
disagree. Beside each statement is a scale which ranges from strongly disagree (1) to strongly
agree (6). For each item we would like you to circle the number that represents the extent to
which you agree or disagree with that statement. The more you agree with a statement, the
higher will be the number you circle. The more you disagree with a statement, the lower will be
the number you circle. Please make sure that you answer EVERY ITEM and that you
circle ONLY ONE number per item. This is a measure of your personal beliefs; there are no
right or wrong answers.

1=STRONGLY DISAGREE (SD)
2=MODERATELY DISAGREE (MD)
3=SLIGHTLY DISAGREE (D)
4=SLIGHTLY AGREE (A)
5=MODERATELY AGREE (MA)
6=STRONGLY AGREE (SA)
SD MD D A MA SA
C1. If I get sick, it is my own behavior which determines
how soon I get well again.
1 2 3 4 5 6
C2. No matter what I do, if I am going to get sick, I will get
sick.
1 2 3 4 5 6
C3. Having regular contact with my physician is the best
way for me to avoid illness.
1 2 3 4 5 6
C4. Most things that affect my health happen to me by
accident.
1 2 3 4 5 6
C5. Whenever I don't feel well, I should consult a medically
trained professional.
1 2 3 4 5 6
C6. I am in control of my health. 1 2 3 4 5 6
C7. My family has a lot to do with my becoming sick or
staying healthy.
1 2 3 4 5 6
163

D. Five Facet Mindfulness Questionnaire

This instrument is based on a factor analytic study of five independently developed mindfulness
questionnaires. The analysis yielded five factors that appear to represent elements of
mindfulness as it is currently conceptualized. The five facets are observing, describing, acting
with awareness, non-judging of inner experience, and non-reactivity to inner experience. More
information is available in:
Please rate each of the following statements using the scale provided. Write the number in the
blank that best describes your own opinion of what is generally true for you.

Never or
very rarely
true
Rarely
true
Sometimes
true
Often
true
Very
Often or
always
true
D1. When I’m walking, I deliberately notice the
sensations of my body moving.
1 2 3 4 5
D2. I’m good at finding words to describe my
feelings.
1 2 3 4 5
D3. I criticize myself for having irrational or
inappropriate emotions.
1 2 3 4 5
D4. I perceive my feelings and emotions without
having to react to them.
1 2 3 4 5
D5. When I do things, my mind wanders off and
I’m easily distracted.
1 2 3 4 5
D6. When I take a shower or bath, I stay alert to the
sensations of water on my body
1 2 3 4 5
D7. I can easily put my beliefs, opinions, and
expectations into words.
1 2 3 4 5
D8. I don’t pay attention to what I’m doing because 1 2 3 4 5
C8. When I get sick, I am to blame. 1 2 3 4 5 6
C9. Luck plays a big part in determining how soon I will
recover from an illness.
1 2 3 4 5 6
C10. Health professionals control my health. 1 2 3 4 5 6
C11. My good health is largely a matter of good fortune. 1 2 3 4 5 6
C12. The main thing which affects my health is what I
myself do.
1 2 3 4 5 6
C13. If I take care of myself, I can avoid illness. 1 2 3 4 5 6
C14. Whenever I recover from an illness, it's usually
because other people (for example, doctors, nurses, family,
friends) have been taking good care of me.
1 2 3 4 5 6
C15. No matter what I do, I 'm likely to get sick. 1 2 3 4 5 6
C16. If it's meant to be, I will stay healthy. 1 2 3 4 5 6
C17. If I take the right actions, I can stay healthy. 1 2 3 4 5 6
C18. Regarding my health, I can only do what my doctor
tells me to do.
1 2 3 4 5 6
164

I’m daydreaming, worrying, or otherwise
distracted.
D9. I watch my feelings without getting lost in
them.
1 2 3 4 5
D10. I tell myself I shouldn’t be feeling the way
I’m feeling.
1 2 3 4 5
D11. I notice how foods and drinks affect my
thoughts, bodily sensations, and emotions.
1 2 3 4 5
D12. It’s hard for me to find the words to describe
what I’m thinking.
1 2 3 4 5
D13. I am easily distracted. 1 2 3 4 5
D14. I believe some of my thoughts are abnormal
or bad and I shouldn’t think that way
1 2 3 4 5
D15. I pay attention to sensations, such as the wind
in my hair or sun on my face.
1 2 3 4 5
D16. I have trouble thinking of the right words to
express how I feel about things
1 2 3 4 5
D17. I make judgments about whether my thoughts
are good or bad.
1 2 3 4 5
D18. I find it difficult to stay focused on what’s
happening in the present.
1 2 3 4 5
D19. When I have distressing thoughts or images, I
“step back” and am aware of the thought or image
without getting taken over by it.
1 2 3 4 5
D20. I pay attention to sounds, such as clocks
ticking, birds chirping, or cars passing.
1 2 3 4 5
D21. In difficult situations, I can pause without
immediately reacting.
1 2 3 4 5
D22. When I have a sensation in my body, it’s
difficult for me to describe it because I can’t find
the right words.
1 2 3 4 5
D23. It seems I am “running on automatic” without
much awareness of what I’m doing
1 2 3 4 5
D24. When I have distressing thoughts or images, I
feel calm soon after.
1 2 3 4 5
D25. I tell myself that I shouldn’t be thinking the
way I’m thinking.
1 2 3 4 5
D26. I notice the smells and aromas of things. 1 2 3 4 5
D27. Even when I’m feeling terribly upset, I can
find a way to put it into words.
1 2 3 4 5
D28. I rush through activities without being really
attentive to them.
1 2 3 4 5
D29. When I have distressing thoughts or images I
am able just to notice them without reacting.
1 2 3 4 5
D30. I think some of my emotions are bad or
inappropriate and I shouldn’t feel them.
1 2 3 4 5
165

D31. I notice visual elements in art or nature, such
as colors, shapes, textures, or patterns of light and
shadow.
1 2 3 4 5
D32. My natural tendency is to put my experiences
into words.
1 2 3 4 5
D33. When I have distressing thoughts or images, I
just notice them and let them go.
1 2 3 4 5
D34. I do jobs or tasks automatically without being
aware of what I’m doing.
1 2 3 4 5
D35. When I have distressing thoughts or images, I
judge myself as good or bad, depending what the
thought/image is about.
1 2 3 4 5
D36. I pay attention to how my emotions affect my
thoughts and behavior.
1 2 3 4 5
D37. I can usually describe how I feel at the
moment in considerable detail.
1 2 3 4 5
D38. I find myself doing things without paying
attention.
1 2 3 4 5
D39. I disapprove of myself when I have irrational
ideas.
1 2 3 4 5

E. Perceived Competence for Maintaining a Healthy Diet

Please indicate the extent to which each statement is true for you, assuming that you were
intending either to permanently improve your diet now or to maintain a healthy diet. Use the
following scale:

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

E1. I feel confident in my ability to maintain a healthy diet.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

E2. I now feel capable of maintaining a healthy diet.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

E3. I am able to maintain a healthy diet permanently.

166

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

E4. I am able to meet the challenge of maintaining a healthy diet

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

F. Perceived Competence for Exercising

Please indicate the extent to which each statement is true for you, assuming that you were
intending either to begin now a permanent regimen of exercising regularly or to permanently
maintain your regular exercise regimen. Use the following scale:

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

F1. I feel confident in my ability to exercise regularly.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

F2. I now feel capable of exercising regularly.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

F3. I am able to exercise regularly over the long term.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true

F4. I am able to meet the challenge of exercising regularly.

1 2 3 4 5 6 7
Not at
all true
Somewha
t true
Very
true
167

Appendix D. Study 3 Coding Scheme

A priori codes
Code Family Code Definition
Sex sex : : male Sex of participant

sex : : female

Behavior tracked track : : nutrition Type of behavior(s) that participant tracks

track : : exercise

track : : sleep

track : : multi

Participant source source : : QS Where participants were recruited from

source : : reddit fitness
forums

Adherence

Discussion of adherence to self-tracking

adher_length
how long participants have been self-
tracking

adher_strategies
strategies participants use to help them
adhere to self-tracking

adher_facilitate factors that promote adhere to self-tracking

adher_barrier
factors that inhibit/act as barriers to self-
tracking adherence

Tracking motivation motivation
Discussion of factors that motivate
participant to adhere to self-tracking
SDT self-regulation
style

Discussion of why participant self-tracks in
SDT behavior self-regulation framework

self-
reg_autonomous/intrinsic
Self-tracking behavior is driven by
autonomous/intrinsic motivation/reasons

self-reg_controlled/extrinsic
Self-tracking behavior is driven by
controlled/extrinsic motivation/reasons

self-reg_amotivation
Self-tracking behavior is not driven by any
intention or motivation
SDT competence

Discussion of the degree to which
participant feels competent in relation to
self-tracking

comp_adherence
Competence in relation to long-term
adherence to self-tracking

comp_behavior change
Competence in relation to changing a
health-related behavior with the aid of self-
tracking
Behavior change
success change
Discussion of behavior change success
aided by self-tracking
168

Goals

Discussion of health goals

goal_short term Short-term health goals

goal_long term Long-term health goals

goal_tracking
How self-tracking will help participant
achieve health goals

goal_extrinsic
SDT goal construct - ex: financial success,
appearance, and popularity/fame

goal_intrinsic
SDT goal construct - ex: community, close
relationships, and personal growth

How tracking helps

Ways in which self-tracking helps with
behavior change/maintenance

helps_awareness/attention

helps_goals/intentions

helps_self-efficacy

Purpose purpose
Discussion of why participant self-tracks
(ex: to help with weight loss)
Tracking method

Type of method participant uses to self-
track

method_app

method_on-body device

method_website

method_manual
(spreadsheet, notes, etc)

method_multiple

method_like
What participant likes about self-tracking
method

method_dislike
What participant dislikes about self-
tracking method

method_improve
What could be improved about the self-
tracking method
Data use data_use
Discussion of how participants use self-
tracking data
Initiation initiation
Discussion of why participant started self-
tracking
Insights insight
Discussion of what participants have
gained through self-tracking
Inductive codes
Self-tracking
adherence adherence_on&off
Discussion of on and off adherence to self-
tracking
How self-tracking
helps helps_build habit
Discussion of how self-tracking helps with
goal attainment
Change self-tracking method_change frequently Discussion of frequently changing self-
169

method tracking methods

method_stick to one you
like or not
Discussion of adhering or not adhering to a
particular self-tracking method
Negative
consequences of
tracking negative
Discussion of negative consequences of
self-tracking
Self-tracking won't
always be necessary not always necessary
Discussion of whether or not self-tracking
will be necessary to continue behavior
change/goal maintenance
Social social_negative
Discussion of negative impacts of social
factors on self-tracking

social_positive
Discussion of positive impacts of social
factors on self-tracking
What one wants out
of tracking self-tracking_wants
Discussion of what one hopes/wishes self-
tracking will/would be able to do

Abstract (if available)

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Motivation and the meanings of health behavior as factors associated with eating behavior in Latino youth

PDF

The acute relationship between affective states and physiological stress response, and the moderating role of moderate-to-vigorous physical activity

Asset Metadata

Creator O'Reilly, Gillian A. (author)

Core Title Effects of sugar and fiber consumption in minority adolescents and self-tracking as a potential dietary intervention tool

School Keck School of Medicine

Degree Doctor of Philosophy

Degree Program Preventive Medicine (Health Behavior Research)

Publication Date 08/01/2016

Defense Date 05/25/2016

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag Adolescent,fiber consumption,OAI-PMH Harvest,obesity,overweight,self-tracking,sugar consumption,weight-related behaviors

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Black, David S. (committee chair), Davis, Jaimie (committee member), Huh, Jimi (committee member), Spruijt-Metz, Donna (committee member), Unger, Jennifer (committee member)

Creator Email gillian.gentner@gmail.com,goreilly@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-290111

Unique identifier UC11279448

Identifier etd-OReillyGil-4700.pdf (filename),usctheses-c40-290111 (legacy record id)

Legacy Identifier etd-OReillyGil-4700.pdf

Dmrecord 290111

Document Type Dissertation

Format application/pdf (imt)

Rights O'Reilly, Gillian A.

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA