University of Southern California Dissertations and Theses

Using mixed methods to identify the characteristics of older fraud victims

(USC Thesis Other)

Using mixed methods to identify the characteristics of older fraud victims

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

USING MIXED METHODS TO IDENTIFY THE
CHARACTERISTICS OF OLDER FRAUD VICTIMS

by
Marguerite DeLiema

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
(GERONTOLOGY)

May, 2015

Copyright 2015 Marguerite Irene DeLiema

ii

ACKNOWLEDGMENTS

This dissertation was completed with financial support from the USC Provost Doctoral
Fellowship, the Haynes Lindley Dissertation Fellowship, and the NIA Training Grant (NIA
T32 AG000037).
First I would like to thank my family: my husband who patiently labored with me in the
trenches while we both completed our dissertations; my cousins—Latimer and Megan Smith—
for graciously inviting us stay in their Casita in Utah to write in solitude throughout the month of
October; and my parents and in-laws for showing interest in my research and for boasting about
my work despite not fully understanding my research. It’s true that parents are the best
cheerleaders.
I also would like to acknowledge the support of my lab group—Jeanine Yonashiro-Cho,
Adria Navarro, Yongjie Yon, Melanie Gironda and Zachary Gassoumis—for prodding my
research in the right direction, providing excellent feedback on drafts, organizing lab meetings,
and laughing at my jokes. I am especially grateful to Zachary Gassoumis for his guidance with
data management, statistical programming, and interpreting findings. I owe all my skills in SAS
and MPlus to his tutelage.
Thank you to my senior colleagues and committee members—Eileen Crimmins, Susan
Enguidanos, Mike Nichols, and many others on the gerontology faculty. They are incredibly
dedicated to mentoring the graduate students and have made my experience in the Gerontology
Ph.D. program better each year. I am now a competent student and researcher thanks to their
commitment to my scholarship.
iii

I also value the tremendous support from members of the LA County Elder Abuse
Forensic Center, particularly Allyson Young, Diana Homeier, and Sasha Lala, for working with
me on grant and IRB proposals and for guiding me through the Forensic Center case files. I am
eternally grateful for the “grounded” perspective I gained from participating in Forensic Center
case discussions. There is truly no better teacher than the real world for understanding elder
financial abuse and fraud.
And of course, none of this work would be possible without support and guidance from
by my faculty mentor, Kathleen Wilber. Thank you for spending countless hours on my
manuscript drafts, allowing me to take first authorship on our joint publications, encouraging me
to tackle challenging research questions, and for telling fantastic stories in our lab meetings.
Because of your efforts, I am a better writer, thinker, and teacher.

iv

TABLE OF CONTENTS

ACKNOWLEDGEMENTS ii
LIST OF TABLES vii
LIST OF FIGURES viii
ABSTRACT ix
CHAPTER I: INTRODUCTION 1
What is fraud and how is it different from elder financial abuse? 1
The scope of the problem 2
Older adults and fraud: Theoretical explanations 3
Social exchange and activity theories 4
Aging and neurological changes 5
Aging, emotion and goals 6
Routine Activity Theory 8
A new theoretical model 12
The present research 15
CHAPTER II: IDENTIFYING THE PREDICTORS OF FRAUD VICTIMIZATION
IN THE HEALTH AND RETIREMENT STUDY
16
Introduction 16
Background and Significance 17
Methods 20
Sample 20
Fraud measure 20
Non-victim control group 21
Socioeconomic & demographic variables 22
Stressful life events 22
Loneliness 23
Cognitive functioning 23
Logistic regression analysis 24
Results 24
Discussion 28
v

Limitations 30
Conclusion 31
CHAPTER III: USING LATENT CLASS ANALYSIS TO IDENTIFY FRAUD
VICTIM TYPOLOGIES IN THE HRS
32
Introduction 32
Identifying typologies using latent class analysis 34
Methods 35
Results 37
Class enumeration 37
Class frequencies (γ) & conditional probabilities (ρ) 37
Discussion 42
Conclusion 44
CHAPTER IV: QUALITATIVE CASE FILE ANALYSIS 46
Introduction 46
Background and significance 48
To what extent do financial abuse and fraud victims share similar
characteristics?
48
The Elder Abuse Forensic Center 52
Importance of capacity evaluations 53
Hypotheses 55
Methods 56
Sampling frame 56
Inclusion/exclusion criteria 56
Case selection procedure 59
Sample size 60
Coding strategy 61
Validity, dependability & credibility 62
Hypothesis testing with neurocognitive data 63
Hypothesis testing with qualitative data 63
Results 64
Victim characteristics 64
vi

Perpetrator characteristics 67
Co-occurring abuse and financial losses 70
Hypothesis 1: Financial abuse victims have poorer cognitive
functioning than fraud victims
71
Hypothesis 2: Fraud victims have better physical health and are less
dependent on others than financial abuse victims
77
Hypothesis 3: Fraud occurs during a “window of opportunity”—the
combination of impaired decision-making and no financial oversight
80
Differences in mental health and emotional status 81
Differences in social and environmental characteristics 82
Discussion 84
A narrative of fraud 86
Limitations 90
Conclusion 90
CHAPTER V: CONCLUSION 93
How to protect vulnerable adults 96
The future of fraud policy 99
The future of fraud research 100
REFERENCES 102
APPENDIX A: Consort diagram of HRS sample selection for Chapters II and III 119

vii

LIST OF TABLES

Table 2.1. Sample characteristics 25
Table 2.2. Logistic regression estimates (odds ratios) of select covariates on fraud 27
Table 3.1. Probability of class membership based on latent class assignment 38
Table 3.2. Item response probabilities (ρ) for each item conditional on fraud class
membership
39
Table 4.1. Criteria for selection into sample 57
Table 4.2. Forensic Center victim characteristics 66
Table 4.3. Forensic Center perpetrator characteristics 69
Table 4.4. Co-occurring abuse types and financial losses 71
Table 4.5. Victim cognitive functioning 72
Table 4.6. Differences in cognitive functioning between fraud and financial abuse
victims using Poisson regression and Fisher's Exact test
77
Table 4.7. Victim physical and mental health characteristics 78
Table 4.8. Differences in health and dependency between fraud and financial abuse
victims
79
Table 4.9. Care environment and grooming characteristics 83
Table 4.10. Differences in social isolation between fraud and financial abuse
victims

84
viii

LIST OF FIGURES

Figure 1.1. Theoretical model depicting fraud against vulnerable older adults 11
Figure 3.1. Latent class analysis model of fraud victim typologies 36
Figure 3.2. Conditional probabilities of gender, race, marital status, and SES of
older fraud victims in the HRS
41
ix

ABSTRACT

In 2011, 7.3% of U.S. adults ages 65-74 and 6.5% of adults ages 75 and older were
victims of financial fraud (Anderson, 2013). In addition to the billions of dollars lost annually to
scams, indirect societal costs include paying for the care and support of elders who lost their life
savings, and the expense of investigating cases and prosecuting offenders. One of the first steps
in stopping fraud is determining who is most vulnerable, yet the research literature on risk factors
presents a conflicting narrative. Some research studies and consumer protection agencies report
that older adults are the most vulnerable age group due to greater social isolation and
impairments in financial decision-making (Ganzini, McFarland, & Bloom, 1990; Lee &
Soberon-Ferrer, 1997; Federal Bureau of Investigation, 2014); however, national prevalence
studies have found that elders are the least likely to experience fraud (Anderson, 2013;
Schoepfer, & Piquero, 2009; Titus, Heinzelmann, & Boyle, 1995; Van Wyk & Mason, 2001).
According to the Financial Fraud Research Center (2012), research identifying the specific risk
factors that make elders susceptible to fraud is important to inform policies and where to target
resources.
After describing the prevalence, cost, and mechanisms of fraud, this dissertation
discusses multiple theoretical explanations for why older adults are vulnerable. Chapter I
presents a temporospatial framework for fraud based on the routine activity theory (Cohen &
Felson, 1979) and an ecological theory of elder financial exploitation (Rabiner, O’Keefe, &
Brown, 2004). The adapted model proposes that targets and scam artists are nested within the
broader social and political macrostructure comprised of public policy and legislation on
consumer protection, prevention/education efforts to reduce fraud, fraud reporting mechanisms,
x

and society’s values, beliefs and attitudes about older adults. Interacting with the macrostructure
is the mesosystem which represents the immediate social and financial environment surrounding
a target and offender—the availability of trusted others to help the target make financial
decisions and the presence/absence of legal mechanisms that protect the target’s assets.
Embedded within the macrostructure and mesosystem is the actual process of fraud. The model
proposes that individuals may be susceptible to fraud due to psychosocial, physical, and/or
cognitive/intellectual impairments associated with age. Susceptible targets interact with
motivated offenders (scam artists) at a time when there is no financial oversight from capable
guardians. This “window of opportunity” represents the timing of exploitation.
In the following three empirical chapters, a mixed methods approach is used to identify
the demographic, socioeconomic, psychological and cognitive characteristics of older fraud
victims using the nationally-representative Health and Retirement Study (HRS) and a sample of
victims in Los Angeles County. To clarify some of the discrepancies on reported risk factors in
the fraud literature, Chapter II uses logistic regression analysis to determine the relationship
between age, gender, and socioeconomic status on fraud in the HRS survey. Contrary to popular
assumptions about older women with low education and low wealth being the most susceptible
targets, results showed that younger age, higher education, and higher income are associated with
fraud. Compared to married respondents, those who are widowed were more likely to be
victimized. While stressful life events increase the odds of victimization, loneliness at baseline
was not statistically significant. Also, despite the widespread assumption that poor cognitive
functioning increases susceptibility, this study found that better cognitive functioning at baseline
increases the odds of victimization.
xi

Researchers have observed that different types of fraud target different socio-
demographic groups (Pak & Shadel, 2011). To identify heterogeneity among the respondents
who reported fraud in the HRS, Chapter III employs latent class analysis (LCA) to test the
hypothesis that older victims of fraud vary in terms of their socioeconomic and demographic
characteristics. LCA uses a maximum likelihood approach to categorize victims into groups
based on their shared characteristics. The second purpose of using LCA is to determine whether
the significant predictors of fraud described in Chapter II—stressful life events and cognitive
functioning—are associated with a specific victim profile.
Two distinct victim classes (i.e., typologies, profiles) emerged from the analysis. Based
on the distribution of socioeconomic and demographic characteristics within each group, Class 1
was descriptively labeled “high-SES middle-age married adults,” and Class 2 was labeled “low-
SES older widowed females.” The high-SES middle-aged married adult group was larger than the
low-SES older widowed female group, suggesting that the former group of victims is more
prominent in the U.S. population over age 50. High-SES middle-aged married adults had higher
average levels of cognitive functioning and also experienced a higher average number of
stressful life events in the past five years compared to low-SES older widowed females. Using a
national sample, this analysis provides comparable results to earlier studies that identified two
victim typologies using a smaller sample of older victims identified by law enforcement: “bogus
prize promotion” and “investment fraud” victim (Pak & Shadel, 2011; Financial Industry
Regulatory Authority; 2006).
To examine these findings in greater detail and to determine whether elders exploited by
friends or family differ from those exploited by predatory strangers, case files from victims
evaluated by clinicians at the Los Angeles County Elder Abuse Forensic Center were analyzed
xii

qualitatively in Chapter IV. No studies to date have used qualitative methods to analyze social
worker case notes, clinician reports, and victim interviews to create a comprehensive profile of
those who are exploited by fraud, and to determine how they differ from elders exploited by
“trusted others” (people in their close social networks).
The main finding of Chapter IV is that although fraud and financial abuse victims share
many of the same physiological, environmental, demographic, and psychosocial characteristics,
they differ in that fraud victims have significantly higher Mini Mental State Exam scores, better
mobility, and are more likely to be childless. A proposed explanation for the similarities is that
exploitation has less to do with the characteristics and risk factors associated with the victim, and
more to do with the people surrounding the victim (or lack thereof). In other words, the structure
of the victim’s social network determines whether he or she is more likely to be a victim of fraud
by strangers or financial abuse by family and friends. This analysis illuminates areas where
intervention and prevention strategies may differ between each type of exploitation.
This dissertation contributes to the literature in several important ways. First, it helps
clarify the relationship between fraud susceptibility, age, and socioeconomic status among adults
ages 50 and older in the U.S. Second, it uses prospective (pre-fraud) data on individuals to
determine whether the purported risk factors for fraud—loneliness, poor cognitive functioning,
stressful life events—actually do increase like likelihood of victimization later on. Third,
findings from the latent class analysis support previous research using victim complaint data to
categorize victims based on their SES and demographic characteristics. And fourth, the
qualitative findings in this study reveal the tremendous financial cost of fraud, and also the
importance of friends and family members in protecting older adults from predatory strangers.

1

CHAPTER I: INTRODUCTION
What is fraud and how is it different from elder financial abuse?
Fraud is defined as, “deliberately deceiving the victim with the promise of goods,
services, or other benefits that are nonexistent, unnecessary, never intended to be provided, or
grossly misrepresented” (Titus, Heinzelmann, & Boyle, 1995). Scams are a type of fraud that
may include bogus prizes and lotteries, sweetheart swindles, fake charities, sale of unnecessary
insurance policies, investments, and worthless consumer products and services. Fraud is
perpetrated via door-to-door sales representatives, direct mail, phishing emails, telemarketing,
and other modes of communication. The following chapters use the terms “scams”, “schemes”,
“swindles”, and “fraud” interchangeably. This dissertation focuses on fraud that targets
individuals, rather than fraud against corporations or the government (e.g., tax fraud, health care
fraud).
Fraud targeting older adults differs from elder financial abuse. Fraud requires that
victims actively comply with the perpetrator’s demands by sending cash, checks, or by
disclosing personal information—social security, bank account and credit card numbers (Van
Wyk & Benson, 1997). Financial abuse does not require victims’ compliance. It involves a
breach of trust between a vulnerable older person and a family member, close friend, or
caregiver (a.k.a., “trusted other”) who takes advantage of access to the elder’s funds, using these
funds to serve his or her own needs without any benefit to the elder. Financial abuse is also
referred to as material abuse, fiduciary abuse, exploitation of resources, and economic
victimization (Wilber, 1990). Fraud and financial abuse are similar in that both are categorized
as crimes of calculation rather than crimes of passion (Coleman, 1987), which would better
describe elder physical, psychological, and sexual abuse (Jackson & Hafemeister, 2011). In the

2

following chapters, “fraud” against elders and “financial abuse” are treated as distinct concepts
yet both are considered forms of “elder financial exploitation.”
While there is no such thing as a “typical scam” (Langenderfer & Shrimp, 2001),
confidence swindles generally follow a similar pattern: First, a scam artist selects potential
targets (“marks”) by contacting them directly or by casting a wider net via mass marketing such
as using phishing emails. Those who respond to the bait are solicited with the promise of
receiving a material or emotional reward—wealth, companionship, sex, a miracle cure, etc.—in
exchange money or property upfront. A key characteristic of fraud is that the reward does not
exist or it does not exist as it is represented by the scam artist. The reward is merely used to
persuade the target to comply.
The scope of the problem
In the United States, an estimated 9.1% of adults ages 55-64, 7.3% of adults ages 65-74,
and 6.5% of adults ages 75 and older are victims of consumer fraud each year (Anderson, 2013).
In 2011 there were an estimated 37.8 million incidents of fraud, yet only 1 million fraud
complaints were received by authorities (Anderson, 2013). Conservative approximations suggest
an annual loss of $50 billion among all adults 18 and older (Deevy, Lucich, & Beals, 2012), yet a
representative survey by the FINRA Investor Education Foundation (2013) found that more than
80% of respondents were solicited to participate in a potentially fraudulent offer and 11% of all
respondents lost money. Thus, the $50 billion dollar figure may underestimate total losses.
Moreover, the figure does not account for the emotional consequences of exploitation including
depression, shame, and loss of trust (Button, Gee, Lewis, & Tapley, 2010; Deem, 2000; FINRA,
2013; Ganzini, McFarland, & Bloom, 1990). Older victims may experience the worst
consequences because they have fewer opportunities to recoup losses and may become

3

dependent on public health, nutrition and housing programs after losing their life savings. For
elder abuse in general, research has shown that victims use 30% more mental health and
substance abuse services (Schonfeld, Larsen, Stiles, 2006); and are hospitalized more often than
non-victims (Dong & Simon, 2013). Thus, fraud exacts a significant toll on individuals and on
our society.
Older adults and fraud: Theoretical explanations
There has been very little theory development in elder financial exploitation (Jackson &
Hafemeister, 2012; Lowenstein, 2009) and in elder abuse in general (Bonnie & Wallace, 2003).
Part of the problem is that financial exploitation encompasses a wide range of crimes—theft,
scams, embezzlement, misuse of powers of attorney, unauthorized joint tenancies, etc.—each
involving a variety of different perpetrators with varying relationships to their victims.
Theoretical explanations for why older adults are susceptible to fraud are even more
sparse despite considerable research on risk factors like dependence, social isolation, and
cognitive impairment (e.g., Cohen, 2006; Jackson & Hafemeister, 2011; Lee & Geistfeld, 1999;
Lee & Soberon-Ferrer, 1997). In this dissertation, susceptibility to fraud is broadly defined as
the likelihood of involvement with any aspect of fraud—including being contacted by and
engaging with a scam artist in addition to losing money to fraud. Proposed theoretical
frameworks describing older adults’ susceptibility include psychosocial explanations such as
social exchange theory (Dowd, 1975), activity theory (Havighurst, 1961), and socioemotional
selectivity theory (Cartensen, 1992); emotion regulation theories involving self-control, credulity
(believing things that are unproven to be true), and visceral influence; and also criminological
theories such as the routine activity theory (Cohen & Felson, 1979). These theoretical
frameworks, applied to fraud against older adults, are reviewed below.

4

Social exchange and activity theories
According to Dowd’s (1975) application of the aging perspective on the social exchange
theory, power resources (money, status, vitality, prestige) are thought to decline with age. With
aging comes retirement, increased physical frailty, possible cognitive impairment, and death of
close friends and romantic partners. The loss of these social roles and partnerships reduces
activity and social interaction in later life, which can cause grief, depression, loneliness, and a
reduction in life satisfaction (Havighurst, 1961). Increased isolation may result in an older adult
being more open to engaging with strangers to fulfill social needs (Friedman, 1992; Lee &
Soberon-Ferrer, 1997).
Losses also create a power asymmetry between the older adult and others in society
(Dowd, 1975). Swindlers capitalize on this power differential. With fewer resources in the
exchange relationship, older adults may be willing to provide financial resources to the scam
artist in return for feelings of companionship and/or considerable promises of wealth (Lee &
Soberon-Ferrer, 1997). Some support for this theory is provided by Mead, Baumeister, Stillman,
Rawn, and Vohs (2011) who found that socially excluded individuals were more likely to
sacrifice their personal and financial well-being for the sake of their social well-being. In this
study, subjects were more willing to spend money on items they didn’t want or need if they
believed it would increase their chances of affiliation with a social partner.
While the combination of social exchange and activity theories explain how age-related
social losses can create a power imbalance between older adults and scam artists, it fails to
explain perpetrators’ motivation for exploiting elders and the changes in decision-making that
make elders more vulnerable to deception.

5

Aging and neurological changes
While the majority of older adults maintain functioning across most cognitive domains
well into late life (Hedden & Park, 2004), small neurological changes associated with age may
affect susceptibility. For example, studies have shown that aging is associated with increased
impairment in the prefrontal cortex, an area largely responsible for executive functioning tasks
such as organizing and planning behavior, inhibitory control, judgment and decision-making
(Craik & Jennings, 1992). Asp et al. (2012) found that subjects with damage to parts of their
ventromedial prefrontal cortex were more credulous toward misleading advertisements than a
control group of individuals without lesions to the prefrontal cortex. Denburg and colleagues
(2007) found that even seemingly healthy older adults with no obvious signs of neuropathology
performed poorly on laboratory tests of decision-making. Tasks involved analyzing rewards,
punishment, risk, and sorting through the ambiguity of different options.
Other regions of the brain involved in mathematical calculations, such as the parietal and
temporal lobes, are also affected by aging and disease (Delazer, 2003). Neuroimaging studies of
adults with Alzheimer’s disease have demonstrated that lesions in these areas affect money
management, particularly the ability to perform complex arithmetic (Knight, & Marson, 2012).
Memory is also important in processing and storing information. Both long-term and
short-term memory are affected by aging (Park et al., 1996). Spencer and Raz (1995) found that
older adults performed worse than young adults in remembering the source of new information.
This has implications for fraud because older adults may misremember if they heard something
from a legitimate source—like a physician or a trusted financial planner—versus an illegitimate
source—like a scam artist trying to sell something they do not need. Other research indicates
that compared to younger adults, older adults are less sensitive to “untrustworthy” facial

6

characteristics (Castle et al., 2012), and are less able to detect lying (Ruffman, Murray,
Halberstadt, & Vater, 2012). They are also less likely to recall information in advertisements but
more prone to be persuaded by the information (Phillips & Stanton, 2004).
Aging, emotion and goals
Aging is associated with changes in emotion regulation. Although working memory
declines with age (e.g., Salthouse, 1991), older adults demonstrate preserved memory
functioning for positive emotional stimuli compared to negative emotional stimuli (Charles &
Carstensen, 2008). As such, they more likely to remember information when it is framed
positively (Charles, Mather & Carstensen, 2003). This is called the “positivity effect” of aging
(Carstensen & Mikels, 2005). Some evidence shows that the positivity effect influences
financial decision-making. For example, Chou, Lee, and Ho (2007) found that risk-taking
tendency was greater for older adults who were in positive mood states than older adults in
neutral or negative mood states. It follows that scams with positive emotional appeal (e.g.,
winning a prize) will elicit compliance more effectively from older adults’ compared to middle-
aged and younger adults who process information differently.
Related to this, the socioemotional selectivity theory (Carstensen, 1992) states that goals
are embedded within the temporal framework of the life course. Individuals who perceive that
they have less time left in life become more focused on goals related to emotional gratification
and less focused on knowledge acquisition and novelty. Emotionally meaningful goals include
social interconnectedness and intimacy. This shift in motivation toward emotionally meaningful
experiences in older age means that more attention is devoted to goal-relevant information,
particularly positive emotional information (Carstensen & Mikels, 2005). Fung and Carstensen

7

(2003) found that older adults were more persuaded by advertisements promoting emotionally
meaningful rewards than advertisements promising to expand their horizons.
Research shows that decision-making is negatively affected by sad emotional states (e.g.,
Baumeister, DeWall, Ciarocco, & Twenge, 2005). Isolating life course events—retirement,
widowhood, onset of disability, death of a spouse or close friends—are more common in late
adulthood and can lead to greater feelings of depression and loneliness (e.g., Gentry, Kennedy,
Paul, & Hill. 1995), and perhaps poorer decision-making. Lerner and Weber (2013) found that
subjects who were manipulated to feel sad were more impatient than subjects in neutral
emotional states and chose to accept an immediate payout of less money instead of waiting to
receive more money in the future. Another study by Duclos, Wan, and Jiang (2013) found that
subjects feeling isolated or ostracized were more likely to pursue risky decisions but with a
higher potential payout, and Baumeister and colleagues showed that social exclusion produces a
decline in cognitive performance (Baumeister, Twenge, & Nuss, 2002) and self-regulation
(Baumeister, DeWall, Ciarocco, & Twenge, 2005). These findings have important implications
for isolated and depressed older adults whose attentional resources may be overtaxed, limiting
their ability to focus on resisting scams.
A central aspect of vulnerability is a person’s tendency toward gullibility and credulity
(Greenspan, Loughlin, & Black, 2001), which may be affected by age-related changes in
physical health, emotion, and cognitive functioning. Pinsker, McFarland, & Pachana (2010)
adapted Greenspan et al.’s (2001) social vulnerability model to explain elder financial
exploitation. According to their revised model, personal competence factors (social intelligence,
general intellectual functioning, cognitive processes, social skills, physical functioning, and
motivation/personality) interact to promote or protect against vulnerability to exploitation.

8

Cognitive, behavioral, and social factors affect a person’s credulity and gullibility (the tendency
toward being manipulated or deceived). The fewer personal competence factors an elder has, the
greater his risk of exploitation (Pinsker, McFarland, & Pachana, 2010). A strength of this model
is that it incorporates a range of variables that influence vulnerability, including acquired
knowledge (e.g., learned financial management skills) and personality, yet it ignores
environmental conditions such as the presence/absence of a perpetrator and the social and
political context that determine the degree to which elders are at risk.
Cognitive and emotion regulation theories help explain why some older adults are
vulnerable to scams, but do not focus attention on why perpetrators choose to target the elderly.
They also make no effort to include the social and institutional structures that make older persons
vulnerable. More narrow explanations such as the positivity effect and socioemotional
selectivity theory do not account for older adult’s vulnerability to scams that prey on negative
emotions (like fear, anger, and jealousy) or that use threats and intimidation to bully targets into
compliance. Therefore, these theoretical perspectives provide only a partial explanation for
fraud.
Routine Activity Theory
Criminological theories add an additional perspective by focusing attention on
perpetrators and the context of fraud. According to Cohen and Felson’s (1979) routine activity
theory, criminal acts require the convergence of three factors: (1) a motivated offender, (2) a
suitable target, and (3) the absence of capable guardians. This theory is not a motivational theory
of criminal behavior, but rather is a temporospatial model that focuses on how routine daily
activities create opportunities for crime to occur (Goergen & Beaulieu, 2010).

9

The first key component of the routine activity theory is the presence of a motivated
offender. In general, fraud perpetrators are motivated by a desire for economic gain. Coleman
(1987) argues that this desire stems from a cultural emphasis on competition for higher economic
status and the socialized belief that wealth is an appropriate proxy for success. Although scam
artists are motivated by financial gain, their willingness to actually commit fraud is modified by
the degree to which they internalize social values of “right” versus “wrong,” and by the
rationalizations they provide for engaging in fraud such as, “I need the money more than she
does,” “He wants me to have the money,” or “He’s such a sucker he deserves to be scammed.”
The second component of the routine activity theory is the presence of a suitable target.
Older adults are likely targeted for reasons including greater economic resources (remunerative),
the perception that they are more susceptible to being deceived (easy targets), or the belief that
they won’t recognize and report the crime to law enforcement (low risk of detection).
The action of financially exploiting an older person is contingent on an opportunity to do
so. Perpetrators must weigh the risks and benefits associated with committing fraud, including
how well the target is guarded and the risk of being caught. Factors that inhibit an older person
from being victimized include having a wide social network (more eyes and ears), having
trustworthy family members and legal/financial professionals involved in financial decision-
making, and the presence of legal documents that protect the elder’s estate. If these factors are
absent, the third criterion in the routine activity model is satisfied and fraud is more apt to occur.
A strength and a weakness of the routine activity theory is its breadth. Although it
focuses on victims, perpetrators, and the context that gives rise to crime, it does not seek to
explain what motivates people to commit fraud specifically (An addiction to drugs? A desire for
power?); what factors make for a suitable target (Wealth? Physical weakness? Gullibility?), and

10

who (or what) acts as a capable guardian (Law enforcement? Trusted family members? Legal
safeguards protecting a person’s assets?). It also does not describe the mechanisms of fraud, or
how money and/or personal information are extracted from victims.
A critique offered by Hotfreter, Reisig, & Pratt (2008) is that the routine activity theory
places too much emphasis on street crime involving deviant offenders, whereas fraud offenders
are not inherently deviant. For one thing, their sales pitches reflect many of the marketing tactics
and communication methods (e.g., internet, mail, telephone, in-person) used by legitimate
businesses to solicit customers. Another element that differentiates white-collar criminals from
street criminals is a false belief by many that they are not committing an offense. Deliema,
Yongjie, & Wilber (2014) found that frontline sales associates engaged in fraud were not clearly
aware of their role in predatory marketing to older clients. Company leaders actively trained
them to believe that their product was legitimate and that they were helping, not harming, elderly
consumers. Thus, although the fraud offender is still motivated to get the target to comply and
hopes to profit from the exchange, he may assume his motivations are guileless. Also, as
opposed to street crime, elderly fraud victims are not “pure victims” (Jackson & Hafemeister,
2011; Doerner & Lab, 2008; Wallace, 2007). There is some degree of complicity/complacency
involved in going along with the scam artist’s requests (see Jackman, 2002). This means that
opposed to violence and outright theft, victims of fraud play a key role in their own
victimization. Another critique of the routine activity theory is that it fails to explain why elders
are perhaps more vulnerable to financial exploitation than other age groups (Jackson &
Hafemeister, 2011). Are older people actually easier to manipulate or just perceived to be by
those who target them?

11

Figure 1.1. Theoretical model depicting a window of opportunity for fraud against
vulnerable older adults; adapted from Routine Activity Theory (Cohen & Felson, 1979)
and ecological model of financial exploitation framework (Rabiner, O’Keeffe & Brown,
2004)

Microprocess
Window of
opportunity

Suitable
Target

Motivated
Offender
Interaction
involving
deceptive offer
Financial
loss
Cognitive/
Intellectual
vulnerability
Financial
resources
Psychosocial
vulnerability
Desire for
financial
gain,
prestige,
power
Low ethical
threshold
Victim
compliance
Physical
vulnerability
Macrostructure
Availability of trusted others to
provide oversight
Fraud reporting
Prevention/education
initiatives
Aging stereotypes
Public policy &
consumer protection
legislation
Mesosystem
Use of legal safeguards to protect assets
Personal knowledge of scams and fraud

12

A new theoretical model
To overcome the limitations of the routine activity theory as an explanation for fraud of
older adults, the model presented in Figure 1.1 incorporates psychosocial theories of aging as
well as ideas from the applied ecological model of elder financial exploitation by Rabiner,
O’Keeffe, & Brown (2004). The current model builds on their 2004 framework by including the
broader policy and social context of fraud that influence an older person’s suitability as a target.
In this model, the social and political macrosystem interacts with and influences
intermediate legal, financial, and social systems that surround older adults and affect their
interactions with scam artists. The macrosystem includes public policies and legislation on
consumer privacy and protection, prevention/education efforts to reduce fraud, fraud reporting
mechanisms, and society’s values, beliefs and attitudes about older adults. For example, the
absence of definitive legislation on how our personal data can be used and shared and how
products/services can be marketed to seniors increase the prevalence of fraud schemes. The lack
of clear guidelines provides opportunities for scam artists to operate within the cover of the
consumer marketplace and avoid detection by law enforcement. Furthermore, low rates of
reporting and lack of consistency in how reports are addressed by complaint agencies make it
easier for scam artists to commit crimes without being detected; and negative stereotypes of
older adults as being unsophisticated, socially isolated, and more trusting increase the likelihood
that they are targeted.
The meso-level depicted in the model represents the social, legal, and financial
environment specific to a target that impacts the probability of being defrauded. These factors
are nested within and interact with the macrostructure systems. Based on the routine activity
theory, the social environment of victims and perpetrators can either inhibit or promote fraud.

13

Suitable targets are those who lack oversight in financial decision-making from friends, family,
or professionals; lack knowledge of common persuasion tactics and scams; and lack the
legal/financial tools (e.g., powers of attorney, living trusts) that help secure their assets.
At the micro level, the process of fraud is depicted an exchange between a vulnerable
older adult and scam artist. The adapted framework in Figure 1.1 presents this exchange by
incorporating ideas from cognitive and psychosocial theories of aging and the social
vulnerability model previously described. Factors that may influence an older person’s
suitability as a fraud target are physical vulnerabilities such as poor vision and limited hearing
which can increase dependency and may affect a person’s ability to pick up on visual/auditory
cues they are being manipulated (Quinn & Tomita, 1997). Another factor is
cognitive/intellectual vulnerability. This relates to age-associated changes in memory, emotion
regulation, and executive functioning that impair a person’s ability to encode and analyze new
information. Other age-related vulnerabilities include impaired problem solving, difficulties
managing complex transactions, and changes in personality (Triebel & Marson, 2012; Moye &
Marson, 2007). Cognitive/intellectual vulnerability also refers to a person’s acquired financial
literacy skills and sophistication in responding to persuasion attempts. Those with poorer money
management abilities and who are less sophisticated in analyzing persuasion messages will be
more susceptible to fraud. A third factor is psychosocial vulnerability. Lonely elders, perhaps
due to widowhood or isolation, may be more willing to socialize with strangers and other
solicitors to fulfill their social needs (Lee & Soberon-Ferrer, 1997). The fourth factor that
increases elders’ suitability as a target is wealth. Elders without financial resources will yield no
returns and are thus poor targets for scam artists.

14

Fraud also requires the presence of a motivated offender to orchestrate the exchange.
As stated previously, perpetrators are motivated by a desire for financial gain, power, and
prestige. They must also be willing to commit a crime to reach these goals, which means having
a low ethical threshold or robust strategies to rationalize immoral behavior and convince
themselves that their behavior does not constitute a criminal act.
The microprocess level also depicts the temporospatial context of fraud. Elders may be
most vulnerable to scams during the early stages of cognitive and physical decline when deficits
are hardest to recognize by others and by elders themselves. Even if they do notice subtle
changes, elders may try to minimize and overcompensate for their deficits because they fear
losing their independence. At these early stages of impairment, family members may not yet be
involved in helping to manage financial affairs (Karp, 2012; Martin et al., 2009). The lack of
financial oversight at the meso-level combined with psychosocial, health, and
cognitive/intellectual vulnerabilities provides a “window of opportunity” for scam artists.
In order to influence their targets, scam artists exploit weaknesses in judgment and
decision-making by deploying subtle persuasion tactics, usually involving emotional arousal and
other visceral cues (Langenderfer & Shrimp, 2001). They also mimic sales tactics used in the
legitimate consumer marketplace (FINRA, 2006, DeLiema et al., 2014). As shown in Figure 1.1,
targets must be willing to comply with the scam artists’ requests in order for the crime to occur,
which is one way fraud sometimes differs from financial abuse by a trusted other. Based on the
principals of the social exchange theory, older adults may be motivated to provide financial
resources to the scam artist in exchange for wealth, companionship or other promised rewards.
The end result of this interaction is financial loss.

15

The present research
The past few decades have seen escalating media attention on the threat of consumer
fraud to the aging population, yet researchers have made little headway in identifying at-risk
groups for targeted prevention and intervention (Deevy, Lucich, & Beals, 2012). Part of the
problem is that studies report a wide range of characteristics associated victimization, and some
of the literature on prevalence and risk factors, including age, gender, income, education, and
level of financial sophistication, have contradictory findings. For example, Lee and Soberon-
Ferrer (1997) reported that consumers were more vulnerable to fraud if they were older, less
educated, poor, and not living with a spouse. Alternatively, the Federal Trade Commission
found that consumers age 55 and older were less likely to have been victims than those between
ages 45 and 54 (Anderson, 2013); and the Financial Industry Regulatory Authority (FINRA,
2006) reported that investment fraud victims were more likely to be educated, married, and have
more financial knowledge than non-victims.
The purpose of the following chapters is twofold: First, to clarify the impact of age,
gender, socioeconomic status, cognitive functioning, loneliness, and negative life events on fraud
using nationally representative survey data and two analytic techniques—logistic regression and
latent class analysis; and second, to explore the differences between fraud and financial abuse by
“trusted others” using qualitative data from over fifty cases presented before an elder abuse
multidisciplinary forensic center in Los Angeles. The goal of these studies is to identify
promising areas for fraud prevention, and also to move the field closer to developing a
comprehensive theoretical model describing victimization of older adults by fraud.

16

CHAPTER II: IDENTIFYING THE PREDICTORS OF FRAUD VICTIMIZATION IN
THE HEALTH AND RETIREMENT STUDY

INTRODUCTION
Conventional wisdom suggests that older adults are more likely to be targeted by scam
artists because of age-related cognitive impairments and changes in emotion regulation (e.g.,
Boyle, Yu, Wilson, Gamble, & Buchman, 2012; Castle et al., 2012), perceived lack of financial
sophistication (Agarwal, Driscoll, Gabaix, & Laibson, 2009), greater wealth (Butrica, Smith, &
Iams, 2012), and greater social isolation compared to younger adults (Lee & Soberon-Ferrer,
1997). FINRA Investor Education Foundation found that upon being solicited for fraud, older
respondents were 34% more likely to lose money than respondents in their forties (FINRA,
2013). Older victims may experience the worst consequences because they have fewer
opportunities to recoup losses and may become dependent on public health, nutrition and
housing programs after losing their life savings. For elder abuse in general, research has shown
that victims use 30% more mental health and substance abuse services (Schonfeld, Larsen, Stiles,
2006); and are hospitalized more than non-victims (Dong & Simon, 2013). Thus, fraud exacts a
significant toll on individuals and on our society.
Although older adults are assumed to be the most vulnerable population, national
prevalence studies by the Federal Trade Commission (FTC) challenge these conventional
assumptions. According to FTC surveys of common types of consumer fraud, victimization is
actually highest among middle-aged adults between the ages of 45-54 (Anderson, 2013). Other
discrepancies among victims have been identified with respect to socioeconomic status (income
and educational attainment) (Anderson, 2013) and cognitive functioning (Ross, Grossman, &
Schryer, 2014).

17

These discrepancies may be attributed to differences in sampling and survey
methodology that limit cross-study comparison and hinder efforts by policy makers to identify
those who are most susceptible. Discrepancies may also result from the cross-sectional focus of
current studies. Longitudinal data are needed to differentiate the antecedents of fraud from the
consequences of fraud, such as loss of income and reduced psychosocial well-being. The present
study uses longitudinal data from the Health and Retirement Study (HRS), a nationally
representative sample of U.S. adults over age 50, to identify the predictors of fraud victimization.
Variables that are thought to affect susceptibility include age, sex, race, educational attainment,
income, marital status, cognitive functioning, loneliness, and stressful life events.
BACKGROUND & SIGNIFICANCE
According to Ross, Grossmann, and Schryer (2014), the perception that older adults are
more susceptible to consumer fraud than younger age groups is largely driven by laboratory
research on age-related cognitive and affective changes, in addition to negative stereotypes of
older adults. For example, a laboratory study examining trust (Castle et al., 2012) indicated that
older adults were less sensitive to “untrustworthy” facial characteristics than younger cohorts.
Similarly, a laboratory study by Ruffman, Murray, Halberstadt, & Vater (2012) suggested that
older adults perform worse than young adults when asked to determine whether people are
telling the truth or lying. And in neuroscience research, Asp et al. (2012) found that damage to
parts of the ventromedial prefrontal cortex, an area involved in decision-making that is shown to
atrophy with age, results in greater credulity toward misleading advertising.
While these studies provide essential information on how information processing changes
with age, Ross, Grossmann, and Schryer (2014) argue that laboratory studies don’t necessarily
translate to real-world behavior. A concern is that the media publicize these study implications

18

as though they were the actual findings (e.g., Lewis, 2012; Lloyd, 2012), despite that the authors
clearly state how their findings only suggest that older adults are more vulnerable to deception.
In contrast to studies conducted in laboratory settings, survey-based consumer fraud
studies do not necessarily corroborate the assumption that elderly people are more likely to be
victims. An early study using a national random probability sample of adults ages 18 and older
found that age and education were negatively associated with fraud victimization (Titus,
Heinzelmann, & Boyle, 1995). Similarly, three consecutive national telephone surveys by the
FTC all supported the finding that older persons are less likely to report fraud relative to adults
of other ages (Anderson, 2013; 2007; 2004). These findings were consistent across all specific
types of fraud with one exception: Respondents between ages 65 and 74 reported the highest
level of bogus prize-promotion fraud; yet in another survey by Schoepfer and Piquero (2009),
younger adults reported prize-promotion fraud more frequently than older adults.
Because these foundational fraud surveys use data that are cross-sectional and
retrospective (e.g., Lee & Soberon-Ferrer, 1997; Anderson, 2013) it has been not possible to
separate the antecedents of fraud from the outcomes of fraud. In fact, some purported risk
factors like social isolation and stressful life events—debt, divorce, illness, and death of a spouse
(Anderson, 2013)—may have followed rather than preceded victimization. Although the few
longitudinal studies that are available (Boyle et al., 2012) have suggested that cognitive
functioning is a predictor of victimization, none of these studies directly measured fraud as the
outcome variable. For example, in a longitudinal study of “cognitively healthy” older adults,
more rapid cognitive decline predicted poorer decision-making and increased susceptibility to
scams (Boyle et al., 2012). However, susceptibility was based on responses to a questionnaire
on poor and risky financial decision-making rather than actual loss of money to fraud.

19

The same can be said for determining psychosocial antecedents of fraud. FINRA found
that victims are more likely than non-victims to have experienced negative life events—
difficulties maintaining their homes, financial trouble, having a serious injury or illness, and
unemployment—in the months prior to falling for lottery and investment scams (FINRA, 2006).
And, the FTC found that negative life events are associated with victimization by debt-related
fraud and bogus prize promotion scams (Anderson, 2013), yet this and prior studies are cross-
sectional (e.g., Anderson, 2011; Alves & Wilson, 2008; Lee & Soberon-Ferrer, 1977). The one
study that did use longitudinal data to predict psychosocial vulnerability to fraud found that
depression measured in 2002 was a significant predictor of fraud occurring sometime between
2003 and 2008 (Lichtenberg, Stickney, & Paulson); however, the authors used cross-sectional
data to test the relationship between other variables such as financial satisfaction, social needs,
and fraud. These observations were drawn from the survey after fraud occurred instead of
before, so it’s not possible to differentiate predictors from outcomes.
To address these limitations in fraud research, the purpose of this study is to (1) clarify
the relationship between age, education, income and fraud, and (2) determine whether cognitive
functioning and psychosocial status at baseline—loneliness and stressful life events—predict
fraud victimization in a nationally representative sample of adults ages 50 and older. Based on
findings from national prevalence studies and cross-sectional studies on elder abuse, the three
study hypotheses are:
1. Younger age and higher socioeconomic status (SES) increase the risk of fraud
victimization.
2. Greater levels of loneliness and stressful life events at baseline increase the risk of fraud
victimization.

20

3. Poorer cognitive functioning at baseline increases the risk fraud victimization.
METHODS
Sample: Data come from the Health and Retirement Study (HRS), a longitudinal biennial
survey comprised of approximately 26,000 individuals over age 50. The core survey, first
administered in 1992, is divided into subject modules that ask participants detailed information
about their physical and cognitive status, family structure, employment, financial status, life
experiences, attitudes and behaviors (hrsonline.isr.umich.edu). Younger cohorts are added every
six years to account for attrition, aging, and death. Blacks and Hispanics are oversampled.
Data files are provided online by RAND Center for the Study of Aging. The present
study uses data from the 1998 through 2010 HRS core surveys and also from the four
Psychosocial Leave-Behind Participant Lifestyle Questionnaires (Leave-Behind) that were
administered to participants who completed an enhanced face-to-face interview that year.
Beginning in 2006, the Leave-Behind was administered to each half of the eligible sample on a
rotating schedule, so each person receives it every other wave. The overall response rates for the
Leave-Behind surveys are 68.3% (2004), 74% (2006), 71% (2008), and 77% (2010).
Fraud Measure: In the 2008 and 2010 Leave-Behind survey, respondents were asked
whether they had been a victim of fraud anytime within the previous five years (‘yes’ or ‘no’),
and if ‘yes’, to state the year fraud occurred. There were no significant differences between the
2008 and 2010 respondents in the independent variables of interest so samples were combined to
increase statistical power. Combining the 2008 and 2010 respondents (n=13,074), approximately
5% of the total sample reported fraud victimization (n=665). Three-quarters of those
respondents reported a valid year of fraud within the 5-year retrospective period (n=506). Those

21

who reported a year more than five years in the past (before 2003 for the 2008 subsample and
before 2005 for the 2010 subsample) were excluded because they could not be assigned an
accurate pre-fraud baseline year. Respondents or their spouses who were younger than age 50
when they were interviewed in either 2008 or 2010, and those who were living in a nursing home
or who responded via proxy, were also excluded. Thus, the total analysis sample included
11,984 respondents; 460 reported fraud. Appendix A presents a visual diagram of the
inclusion/exclusion procedures.
Non-victim Control Group: Of the total analysis sample, 11,524 respondents (~95%) did
not report fraud in the previous 5 years. To select a “pre-fraud” baseline year for this group,
non-victims were randomly assigned a baseline year based on the distribution of when fraud
occurred for the 460 individuals who did report fraud. For illustration of this procedure, assume
that 10% of the fraud sample stated that victimization occurred in 2005. To create a comparable
control group for these individuals, 10% of non-victims were randomly assigned that same year
as a reference. Their baseline data (and the baseline data for the corresponding 10% of fraud
victims) was drawn from 2004, the survey prior to the reference year. Using this approach,
victims’ baseline survey years were proportionally equal to non-victims’ baseline survey years.
Approximately 3% (n=343) of non-victims and 4.8% of fraud victims (n=22) were missing
baseline observations largely because the HRS sample was replenished in 2006 and these
subjects were part of the new and younger cohort.
The total sample contained 3,275 households in which multiple household members were
administered the Leave-Behind. A concern with this study design is that both the respondent and
his/her spouse could respond independently to the fraud question. Thus, single incidents of fraud
could be “double counted” by both members living in a household, thereby leading to biased

22

estimates, particularly for socioeconomic variables since partners tend to be similar in race,
educational attainment, income, and of course, marital status. Despite this concern, in 164
households only one household member reported fraud. Seventy-two of the reporters were
female and 92 were male. Of the households where both partners reported fraud (n=39), half of
them reported different years (n=19), suggesting that partners are reporting unique incidents of
exploitation. In preliminary analyses, clustering the sample by household did not change the
parameter estimates, so the issue of “double-counting” fraud did not pose a threat to the validity
of the regression model.
Socioeconomic & Demographic Variables: Demographic variables included in the
analysis are race (Non-Hispanic White/Non-White), education (less than college/some college or
more), sex (male/female), age (continuous), marital status (married/partnered, divorced,
widowed, and never married), and annual income. Pre-fraud baseline data were used for time-
varying characteristics. Income was adjusted to account for household size based on the
following equation: Adjusted individual income = 𝑡𝑜𝑡𝑎𝑙 ℎ𝑜𝑢𝑠𝑒 ℎ𝑜𝑙𝑑 income/√ (household size).
This formula takes into account “economies of scale” in a multi-person household (Rainwater,
1974). Respondents were divided into four income quartiles based on the income distribution in
the weighted sample.
Stressful life events: The Leave-Behind survey measures stressful life events based on a
widely cited study on the mental health effects of major life stressors by Turner, Wheaton, &
Lloyd (1995). Respondents were asked whether they experienced any of the following five
events sometime within the past five years: involuntary job loss, unemployment for longer than 3
months, someone else in the household unemployed for over 3 months, moved to a worse

23

residence, and victim of a robbery home burglary. Affirmative responses were added yielding an
index with range “0” (no stressful events) through “5” (all stressful events endorsed).
Loneliness: Social isolation is often cited as a risk factor for elder abuse and financial
exploitation (Dong & Simon, 2008; Lee & Soberon-Ferrer, 1997). For this study, loneliness,
rather than frequency of social contact with friends and family members, was selected for the
model because it is a direct measure of subjective feelings of social isolation. The Leave-Behind
survey measures loneliness using three questions based on Hughes, Waite, Hawkley & Cacioppo
(2004). Participants are asked how often they (1) lack companionship, (2) feel left out, and (3)
feel isolated from others. Response categories are: 1=often; 2=some of the time; and 3=hardly
ever or never. Items were reverse scored. A total loneliness score was calculated by averaging
responses across the three items (range 0-3), with higher values indicating greater levels of
loneliness.
Cognitive Functioning: Participants were administered the Telephone Interview for
Cognitive Status (TICS) at each survey wave (Folstein, Folstein, & McHugh, 1975). The TICS
is a standardized assessment containing 35 items that measure word list memory, semantic
knowledge, orientation, language, attention, mathematical skills, repetition, and nonverbal
praxis. One point is given for each correct answer and higher scores indicate better cognitive
functioning. Epidemiological studies and clinical trials using the TICS have shown it to have
high reliability and validity (e.g., Brandt, Spencer, Folstein, 1988; Welsh et al., 1993). This
analysis sample excludes individuals too impaired to complete the TICS. These respondents
generally rely on proxies to answer the core surveys for them. The Leave Behind was not
administered to proxies so whether subjects with severe cognitive impairment experienced fraud
(as reported by their proxies) was never ascertained.

24

Logistic Regression Analysis: Stepwise multivariable logistic regression was used to
determine whether younger age, having greater education and income (Model 1), loneliness and
stressful life events (Model 2), and poor cognitive functioning (Model 3) significantly predict
fraud. Sampling weights from the Leave-Behind were applied to ensure that estimates are
representative of adults over age 50 in the U.S. Weights were adjusted for sample size. A type-I
error rate of less than 5% was considered statistically significant, and 1% indicated a trend. SAS
9.3 was used for coding, descriptive, and bivariate statistics. MPlus 7.3 was used for the logistic
regression analysis because it uses a full information maximum likelihood (FIML) approach to
account for missing data, meaning that cases with missing observations are still included in
estimating model parameters.
RESULTS
As shown in Table 2.1, the mean age of respondents who reported fraud was 61.7 years
(SD=8.6), younger than the average age of non-victims (64.6 years, SD=9.3). In bivariate
analysis these differences were statistically significant (t=6.21, p<0.001). On average, fraud
victims reported a higher median annual income at baseline compared to non-victims—$39,466
versus $30,852—and one additional year of education—13.7 years versus 12.7 years.
Differences in income and years of education were significant: tincome=-1.97, p=0.049, and
teducation= -8.02, p<0.001. Although differences were not statistically significant (χ
2
=3.80,
p=0.051), 55.5% of non-victims were female compared to 48.6% of the fraud sample. Sixty-nine
percent of non-victims were married or partnered, 12.9% were divorced, 14.5% widowed, and
3.9% never married at baseline. In the fraud sample, 66.7% were married/partnered, 15.1%
divorced, 14.5% widowed, and 3.8% never married. The majority of both groups were non-
Hispanic White. Differences in race and marital status were not significant. The fraud

25

respondents performed significantly better on TICS at baseline compared to non-victims (t= -
7.57, p<0.001), and reported more stressful life events (t= -3.92, p<.001). Differences in
baseline levels of loneliness were not significant (t= -1.60, p= 0.11).
Table 2.1. Sample characteristics
Mean/Frequency (%)
Non-Victims Fraud Victims
n=11,524 n=460
Mean age (St. Dev.) 64.6 (9.3) 61.7 (8.6)
Female 55.5 48.6
White 83.9 85.2
Black 7.8 6.3
Hispanic 6.0 5.0
Other 2.3 3.5
Mean years of education (St. Dev.) 12.7 (3.04) 13.7 (2.62)
Married/partnered 68.7 66.7
Divorced 12.9 15.1
Widowed 14.5 14.5
Never Married 3.9 3.8
Median household income $30,852 $39,466
Mean # stressful life events 0.24 (0.57) 0.50 (0.89)
Mean loneliness 1.46 (0.54) 1.53 (0.65)
Mean cog. functioning (TICS score) 23.35 (4.61) 24.83 (3.99)

Table 2.2 presents results of the stepwise logistic regression. As predicted, younger age
and higher SES are positively associated with fraud in all three regression models. In Model 1,
for each year of age after 50, the odds of fraud victimization decrease by 3.6% (p<.001).
Attending some college or more increases the odds of victimization by 53% (p=.001). Compared
to those in the bottom income quartile ($0-19,460), respondents with the highest annual income
($55,356 and up) have the greatest risk of fraud at a trend level (OR=1.47, p=.051). Females are

26

less likely to be victims than males (OR=0.74, p=.015), and widowed respondents are more
likely to be victimized than married respondents (OR=1.98, p=.001). Race/ethnicity, never being
married, being divorced, and having median levels of income are not significant.
Psychosocial factors were added to the regression in Model 2. Although there is little
change in the relationships and significance of variables from Model 1, higher income is no
longer statistically significant (OR=1.45, p=0.10). As predicted, stressful life events increase the
odds of victimization, such that for every additional negative life experience in the previous five
years, the odds of fraud increase by 36% (p=.013). Contrary to the hypothesis that loneliness
increases risk of fraud, loneliness at baseline is not statistically significant (OR= 1.25, p=.240).
Cognitive functioning was added to the regression in Model 3. Despite the widespread
belief that poor cognitive functioning increases susceptibility, this study found that better
cognitive functioning at baseline increases the odds of victimization (OR=1.05, p=.007). In
other words, as performance on the TICS increases, so does the risk of fraud.
All models have good overall specification. Applying a Wald test to constrain the
parameters, the three models were deemed to be significant (Model 1 χ
2
=69.1, p<.001; Model 2
χ
2
=78.8, p<.001; Model 3 χ
2
=89.7, p<.001).

27

Table 2.2 Logistic regression estimates (odds ratios) of select covariates on fraud (n= 11,984)
Model 1 Model 2 Model 3
Independent Variables Weighted OR (95% CI) p-value Weighted OR (95% CI) p-value Weighted OR (95% CI) p-value
Demographic Characteristics
Age 0.96 0.95 0.98 <.0001 *** 0.97 0.96 0.98 <.0001 *** 0.98 0.96 0.99 0.001 **
Non-Hispanic White 1.07 0.82 1.40 0.684 0.99 0.75 1.30 0.950 1.01 0.77 1.34 0.932
Sex (female=1) 0.74 0.60 0.91 0.015 ** 0.70 0.57 0.86 0.005 ** 0.72 0.58 0.88 0.007 **
Some college or more 1.53 1.23 1.90 0.001 *** 1.42 1.13 1.78 0.011 * 1.39 1.11 1.74 0.017 *
Married/Partnered (reference) --- --- --- --- --- --- --- --- --- --- --- ---
Divorced 1.33 0.97 1.82 0.137 1.25 0.90 1.73 0.266 1.24 0.90 1.73 0.276
Widowed 1.98 1.43 2.75 0.001 ** 1.85 1.32 2.58 0.003 ** 1.81 1.30 2.52 0.003 **
Never Married 0.99 0.55 1.79 0.975 0.93 0.51 1.69 0.831 0.92 0.50 1.67 0.811
Income $0-19,460 (ref) --- --- --- --- --- --- --- --- --- --- --- ---
Income $19,461-47,664 1.16 0.86 1.57 0.412 1.11 0.82 1.50 0.582 1.14 0.85 1.55 0.466
Income $47,665-55,355 1.18 0.87 1.60 0.380 1.40 0.82 1.52 0.560 1.16 0.85 1.58 0.427
Income $55,356 and up 1.47 1.06 2.04 0.051
†
1.45 1.00 1.95 0.097 1.45 1.04 2.03 0.065
†
Psychosocial characteristics
Stressful life events (1999-2006) 1.36 1.11 1.67 0.013 * 1.35 1.09 1.66 0.019 *
Loneliness 1.02 0.97 1.84 0.081 1.28 0.94 1.73 0.190
Cognitive functioning
TICS performance 1.05 1.02 1.07 0.007 **

Note: Time-varying characteristics (e.g., marital status, household income, TICS score) are all measured at baseline (pre-fraud)

28

DISCUSSION
As hypothesized, this study found that age is negatively associated with fraud, such that
the risk of victimization declines after age 50. This finding is consistent with studies by
Anderson (2013), Lichtenberg, Stickney, & Paulson (2013), Titus, Heinzelmann, & Boyle (1995)
and Van Wyk & Mason (2001). For example, the FTC found that middle-aged adults experience
the highest rates of fraud while adults 75 and older report the lowest rates, followed by adults
ages 65-74 (Anderson, 2013). Some researchers have hypothesized that adults in their 40s and
50s face the greatest risk of exploitation because consumption peaks in middle-age and
consumers are exposed to more scams than non-consumers (Attanasio, Banks, Meghir, & Weber,
1999; Van Wyk & Mason, 2001).
The finding that age is negatively associated with fraud contradicts popular assumptions
that older adults are the most vulnerable age group due to cognitive impairment and social
isolation. On the contrary, fraud victims in this study had better cognitive functioning than non-
victims, and loneliness was not a significant predictor in the regression model. Although
Hypotheses 2 and 3 were not supported by the data, the result that fraud is significantly
associated with higher levels of cognitive performance aligns with the finding that victims are
also more likely to be middle-aged and more educated than non-victims.
An alternative explanation is that older adults actually do experience higher rates of fraud
yet fail to report their victimization to complaint agencies, law enforcement, and in surveys like
the HRS. There are a number of explanations for underreporting. First, reporting relies on
subjective memory of past events. Research indicates that aging is associated with declines in
episodic memory (Craik, 1986) and that older adults have a greater tendency to minimize
emotionally negative experiences (Charles & Carstensen, 2008). Elders may also choose to hide

29

victimization because of shame, embarrassment, or a belief that they are partially to blame for
being complaisant in the scam (Deem, 2000; Ganzini, McFarland, & Bloom, 1990). Thus,
studies like the HRS that rely on victim self-report and studies that use data from complaint
agencies may underestimate the frequency of fraud among older adults.
As predicted by Hypothesis 1, higher SES predicts fraud. Those in the highest income
quartile and those with more than a high school education face the greatest risk. Scam artists
may deliberately target adults with more wealth because they can potentially make a larger
profit. This finding is important because it runs counter to the common stereotype that fraud
victims are isolated elderly women with low education and low wealth. The alternative picture
painted by this study suggests that victims are actually quite young (closer to middle-aged),
male, wealthy and well educated, and cognitively intact. One critical finding that conflicts with
this picture is the result that widowed adults are more likely to be victims than married adults.
This is counterintuitive given that widowhood is generally associated with older age and being
female, yet once these variables are controlled for in the regression, widowhood was found to
independently increase risk. More research is needed to explore this unexpected and somewhat
atheoretical result.
FINRA (2006) found that compared to non-victims, fraud victims experience a greater
number of negative life stressors, including illness, disability, legal problems, and worry over
housing and finances. In this study, negative life events—e.g., unemployment, moving to a
worse residence, robbery—significantly increased the risk of fraud (FINRA, 2006). One
explanation is that adults with financial and housing trouble are more willing to comply with
“get rich quick” and debt consolidation schemes than adults without economic worries.

30

To test the hypothesis that financial problems are associated with fraud susceptibility, the
variable “ongoing financial strain” was added to the regression model (results not shown).
Although significant in bivariate correlations, financial strain was not significant in the
regression, suggesting that financial problems alone do not account for the effect of stressful life
events on fraud. An alternative explanation is that stressful life events increase depression
symptomology in later life, which subsequently increases the risk of experiencing additional
stressful events (Moos, Schutte, Brennan, & Moos, 2005), like falling for a scam. Indeed,
Lichtenberg, Stickney, & Paulson (2013) found that depression measured in 2002 was a
significant predictor of fraud between 2003 and 2008, but this finding was not replicated in the
present study (results not shown) which uses a larger sample and randomly assigned baseline
years for the non-victimized control group.
LIMITATIONS
A limitation of this study is that the HRS does not capture the amount of money lost from
scams or the type of fraud, so it is not possible to distinguish respondents who were attempted
targets of fraud from those who were “successful” victims. Some individuals may perceive
themselves to be victims when they were targeted but experienced little to no financial
consequences, whereas others who suffered large losses may not label the experience as fraud.
Moreover, the HRS did not capture respondents’ lifetime experience of fraud, so those who were
defrauded more than five years in the past were excluded from the fraud subsample even though
they could share some of the same characteristics and risk factors as the more recent victims.

31

CONCLUSION
This study adds empirical support to Ross, Grossmann, and Schryer’s (2014) argument
that older adults are not necessarily more susceptible to fraud than younger age groups, yet this
assumption permeates through the psychology literature and the popular media. Contrary to
laboratory studies that claim that age-related cognitive and affective changes increase
vulnerability to fraud, the present study found a negative association between age and fraud, and
that better cognitive performance predicted fraud. A major advantage of this study over previous
national prevalence studies is that it uses longitudinal data to disentangle the antecedents of
fraud, such as stressful life events, from the consequences of fraud. Because victims in the HRS
are more likely to be young and have higher SES than previously thought, prevention efforts
should include targeting middle-aged adults in upper income brackets. Further research should
investigate the mechanisms underlying the relationship behind stressful life events, widowhood,
and increased risk of fraud, with interventions designed to mitigate the negative psychosocial
effects of these events that put older people at risk.

32

CHAPTER III: USING LATENT CLASS ANALYSIS TO IDENTIFY FRAUD VICTIM
TYPOLOGIES IN THE HRS

INTRODUCTION
As discussed in Chapter II, the fraud literature presents a conflicting narrative about who
is most vulnerable to exploitation: Middle or older aged adults? People with low-SES or high-
SES? Cognitively impaired or cognitively intact adults? Contradictory findings suggest that
there is not a “typical” profile that characterizes all fraud victims over age 50 in the U.S. More
likely, scams differentially target individuals with specific socioeconomic characteristics that
make them either more profitable targets or more susceptible victims. Some individuals are
vulnerable because they engage in risky consumer behaviors, such as opening spam mail or
responding to telemarketers (AARP, 1996; Lee & Geistfeld, 1999; Van Wyk & Benson, 1997);
whereas others are defrauded without their knowledge, such as in cases of identity theft. Some
scams market to older adults, like lottery, charity, and prize-promotion scams, and others market
to young people, like bogus weight-loss products and work-at-home scams. In sum, the wide
variety of schemes and targeting of various groups makes it difficult to identify common
vulnerability traits across victims in different surveys and samples (Deevy, Lucich, & Beals,
2012).
Studies that have profiled victims according to the type of fraud support this pattern. The
FINRA Investor Education Group found that not only do victims of investment fraud and bogus
lotteries differ demographically and socioeconomically from non-victims, they also differ from
one another (FINRA, 2006). AARP also found that victims differ based on fraud type (Pak and
Shadel, 2011). Compared to the general adult population, AARP found that prize promotion
victims are more likely to be single, female, less educated, and have an annual income of less

33

than $50,000; whereas investment fraud victims are more likely to be male, better educated, and
have an annual income of more than $50,000 (Pak & Shadel, 2011). In a sample of over 2,300
adults ages forty and older FINRA (2013) found that willingness to take investment risk was
positively correlated with household income, education, and being male, and negatively
correlated with age.
The victim typologies described by AARP (Pak & Shadel, 2011) and FINRA (2006) were
identified using administrative data on verified fraud victims in cases investigated by law
enforcement, not in a random sample survey of adults. Generally, consumer fraud studies either
rely on a non-representative sample of confirmed victims provided by law enforcement and
complaint agencies (e.g., Pak & Shadel, 2011; FINRA, 2006), or a random sample of adults who
may have experienced fraud, but did not necessarily reported it to authorities (e.g., Anderson,
2013; Schoepfer & Piquero, 2009; Titus Heinzelmann, & Boyle, 1995). These methodological
differences make it hard to compare findings across studies and draw reliable conclusions about
the populations that are most vulnerable.
To address these issues, the purpose of this study is to use nationally representative
survey data to classify victims of fraud ages 50 and older into distinct socioeconomic and
demographic groups and to determine whether the significant predictors of fraud described in
Chapter II—stressful life events and cognitive functioning—are associated with a specific victim
profile. This study hypothesizes that fraud victims aged 50+ in the United States reflect the
characteristics of “bogus prize promotion” and “investment fraud” victims identified in samples
provided by complaint agencies (FINRA, 2006; Pak & Shadel, 2011). These typologies have not
yet been replicated using a random, nationally representative sample of middle-aged and older
Americans interviewed before and after fraud.

34

Identifying Typologies using Latent Class Analysis
Multivariable logistic regression is a statistical approach used to test the relationship
between multiple independent variables and a dichotomous dependent variable while adjusting
for potential confounders (Hidalgo & Goodman, 2013). In Chapter II, logistic regression was
used to determine whether fraud in the age 50+ population is predicted by being middle-aged
(closer to age 50), having higher levels of income and education, poorer cognitive functioning,
and greater levels of loneliness and stressful live events.
While multivariable logistic regression can test the strength and direction of relationships
between predictor variables and the outcome measure, the method treats all independent
variables in the same model as orthogonal. It is therefore difficult to determine which variables
are highly correlated and whether there are unique differences among respondents in the same
sample. In other words, there may be substantial variation among victims that cannot be
determined using regression techniques.
To overcome these methodological limitations, this study uses a statistical technique
called latent class analysis (LCA) to test the hypothesis that victims of fraud not only differ from
non-victims, but also differ from one another. LCA uses a maximum likelihood approach to
categorize individuals based on their shared characteristics and behaviors (McCutcheon, 1987).
Individuals within a group are similar to each other but different from individuals in other
groups. Since past research has shown that scams are crafted to target individuals with a certain
set of characteristics (such as being poor and isolated or wealthy and married), this study
hypothesizes that the unobserved “latent” variable that differentiates victims from one another is
the type of fraud they succumbed to.

35

Although similar to latent factor analysis, LCA is a “person-centered” method whereas
latent factor analysis is a “variable-centered” method (Muthén & Muthén, 2002). One advantage
of LCA is that it does not rely on assumptions of linearity, normal distribution, or homogeneity
that are often violated in regression analysis leading to biased interpretations of the parameter
estimates (Magidson & Vermunt, 2004). Furthermore, the latent class structure can be estimated
using binary, nominal, ordinal, and Likert-scale variables. When all the indicator variables are
continuous the procedure is called latent profile analysis (LPA), but since this study uses both
continuous and binary observations to identify fraud typologies, the term latent class analysis,
“LCA”, will be used throughout.
METHODS
Mixture modeling was conducted using MPlus (Muthén & Muthén, 2006), a software
program that can fit both categorical and continuous data and uses a maximum likelihood
approach to account for missing observations. Using only the fraud subsample (n=460), a single-
class solution (k =1) was estimated first by regressing the observed indicator variables—age,
race, sex, education, marital status, income, cognitive functioning and stressful life events—on
the categorical latent variable. Age, cognitive functioning and stressful life events were
measured continuously; race and education were dichotomous (as opposed to categorical) to
simplify model interpretation; and income and marital status were categorical. Models were
estimated using participant weights so that the results are representative of the U.S. population
age 50 and older. Figure 3.1 depicts the latent class model with both continuous and binary
indicator variables and one latent variable with multiple levels (victim classes).

36

Unobserved
latent
variable

Figure 3.1. Latent class analysis model of fraud victim typologies

Note: Age, cognitive functioning and stressful life events are measured continuously (circles); all
other variables are categorical (squares)

Using stepwise addition, k +1 classes were modeled until the best class solution for the
data was reached. Class selection was based on conceptual interpretability of the distribution of
characteristics within each class as well as five indicators of model fit: the Akaike’s Information
Criterion (AIC; Akaike, 1987), the sample-size adjusted Bayesian Information Criterion (BIC;
Schwartz, 1978), the Vuong-Lo-Mendell-Rubin test (VLMR; Lo, Mendell, & Rubin, 2001), the
parametric bootstrapped likelihood ratio test (BLRT; McLachlan & Peel, 2000), and entropy.
When comparing nested models, lower AIC and BIC values are preferable, and each class must
be conceptually coherent and distinct from other classes (Kass & Wasserman, 1995). Entropy is
a measure of classification certainty, or how well individuals are assigned to a group on a 0 to 1
Race/
Ethnicity
Age Sex
Marital
Status: 4
categories
Income: 4
categories
College
Cognitive
Functioning
Stressful
Life
Events
Fraud
Victim
Typology
Observed
variables
in dataset

37

scale, with values closer to 1 indicating better classification. Entropy values greater than .80
indicate an acceptable delineation of people into classes (Ramaswamy et al., 1993).
Both VLMR and BLRT compare the improvement in fit between nested models: k
classes versus k -1 classes. These tests provide a p-value indicating whether the k − 1 class
solution is rejected in favor of the k class solution. When choosing the optimal number of
classes to describe the data, the VLMR and BLRT estimates should be significant for k classes
(indicating improvement over the k -1 model), but not significant for the k + 1 model, which
suggests that an additional class offers no further improvement in model fit.
RESULTS
Class enumeration: A two-class solution offered the most robust delineation of fraud
victims into classes. In the two-class solution, the BIC = 9400 and AIC = 9373, a decrease from
the one-class (full fraud sample) solution (BIC=9613; AIC= 9598). The entropy of the two-class
solution was 0.85 compared to 0.76 for the 3-class solution, suggesting that the 3-class model
was less precise in assigning individuals into groups. The VLMR test was not statistically
significant in the 3-class model indicating a lack of statistical support for the inclusion of a 3
rd

class. Conversely, the BLRT value was significant the three-class model suggesting that a third
class may be appropriate; however, based on the other four model fit indices and a qualitative
assessment of the 2-class versus the 3-class model, it was determined that two classes were a
better solution for the data.
Class frequencies (γ) & conditional probabilities (ρ): Based on the distribution of
socioeconomic and demographic characteristics within each group, Class 1 was descriptively

38

labeled “high-SES middle-aged married adults,” and Class 2 was labeled “low-SES older
widowed females.” Classification probabilities are shown in Table 3.1.
Table 3.1. Probability of class membership based on latent class assignment

Class Assignment

High-SES Middle-aged
Married Adults

Low-SES Older White
Widowed Females
Classification probability:
Class 1 0.98 0.02
Class 2 0.13 0.87

Although individuals have some probability of being in either class, those who are
characterized as “high-SES middle-aged married adults” have an average probability of 0.98 for
being in Class 1, and only a 0.13 estimated probability of being in Class 2. Those who were
characterized as “low-SES older widowed females” have a .87 mean estimated probability of
being in Class 2, and only a .02 average probability of being in Class 1. These results provide
further evidence that fraud victims are appropriately differentiated into two groups.
One parameter estimated by LCA is the relative size of each latent class (γ), or the
proportion of individuals assigned to each class. The proportion of fraud victims assigned to
Class 1 (γ = 0.834) is greater than those assigned to Class 2 (γ = 0.166), suggesting that low-SES
older widowed female victims are less prevalent in the population than high-SES middle-aged
married adult victims. Table 3.2 presents class frequencies and item-response probabilities for
each variable, conditional on class membership.

39

Table 3.2. Item response probabilities (ρ) for each item conditional on fraud class membership
Class # (frequency): 1 (γ=0.834) 2 (γ=0.166)

Class Label:
Total Fraud
Sample
(n=460)
High-SES
middle-aged
married
adults
(n=384)

Low-SES
older
widowed
females
(n=76)
Odds ratio:
Class 1
compared to
Class 2

Mean age (Std. Dev) 61.73 (8.0) 58.9 (4.2) 75.0 (4.2)
Female 0.49 0.45 0.64 0.47**
Male 0.51 0.55 0.36 ref
White 0.85 0.84 0.89 0.70**
Non-White 0.15 0.16 0.11 ref
College or more 0.64 0.71 0.32 5.17**
Less than college 0.36 0.29 0.68 ref
Married/Partnered 0.67 0.70 0.51 0.46**
Divorced/Separated 0.15 0.17 0.06 0.20**
Widowed 0.15 0.09 0.42 10.18***
Never married 0.04 0.05 0.01 ref
Household Income Q1: $0-24,361 0.25 0.18 0.57 5.91**
Household Income Q2: $24,362-49,144 0.25 0.25 0.27 7.06*
Household Income Q3: $49,145-79,911 0.25 0.28 0.10 6.47*
Household Income Q4: $79,912 and up 0.25 0.29 0.06 ref
Mean cognitive functioning (Std. Dev) 25.23 (1.3) 25.7 (3.68) 22.8 (3.68)
Mean stressful life events (Std. Dev) 0.54 (0.91) 0.60 (0.85) 0.17 (0.85)

Note: Variance of continuous variables held constant within class

Conditional response probabilities (ρ) represent the probability that a person assigned to a
particular latent class has a certain set of characteristics. These item response probabilities are
used to assign descriptive labels to each group. For example, individuals in Class 1 have a 0.55
probability of being male, a 0.70 probability of being married/partnered, and a 0.71 probability
of having completed some college or more. These individuals also have a higher probability of

40

being in mid- to upper-income quartiles: ρ=.25 (probability of income between $24,362 and
$49,144), ρ=.28 (probability of income between $49,145 and $79,911), and ρ=.29 (probability of
income $79,912 and greater). Thus, this group is labeled “high-SES middle-aged married
adults.”
Aside from the similarity of race (both groups have above an 80% probability of being
non-Hispanic White), the characteristics of those in Class 2—low-SES older widowed females—
are opposite those in Class 1. Individuals in Class 2 have a high probability of being female
(ρ=.64), having less than a college education (ρ=.68), and being in the lowest annual household
income quartile: ρ=.57 (Q1) versus ρ=.06 (Q4). The gender and marital characteristics of each
victim class reflect the distribution of gender and marital status in the United States at those two
relative ages. That is, the older victims in Class 2 (mean age = 75.0 years) had a higher
probability of being female (ρ=.64) and widowed (ρ=0.42) compared to the relatively younger
victims in Class 1 (mean age=58.9 years), who were significantly less likely to be female (ρ=.45)
or widowed (ρ=.09).
The two classes also differed in average cognitive functioning and mean number of
stressful life events. While the logistic regression results in Chapter II indicated that better
cognitive functioning and more stressful life events predict fraud, these risk factors more closely
categorize the victims in Class 1. High-SES middle-aged married adults have higher average
levels of cognitive functioning than low-SES older widowed females. The conditional mean
TICS score for Class 1 was 25.7 versus 22.8 for Class 2. The finding that Class 1 has higher
mean cognitive functioning is expected given that this group is younger on average than low-SES
older widowed females, who presumably experience more age-related cognitive declines.

41

High-SES middle-aged married adults also experienced a higher average number of
stressful life events in the past five years than low-SES older widowed females (mean = 0.60 vs.
0.17). Low-SES older widowed females actually had a lower mean number of stressful life
events than the general population (mean=0.54). Low overall values indicate that stressful life
events, like losing a job or being robbed, are relatively infrequent among adults ages 50+. Class
1 likely experiences more stressful life events than Class 2 because the index contains three
items related to job loss and unemployment. Given that low-SES older widowed females are
older, this group is less likely to be participating in the paid labor force than high-SES middle-
aged married adults, and are thus less likely to be fired from a job and/or seeking work.

Figure 3.2. Conditional probabilities of gender, race, marital status, and SES of older
fraud victims in the HRS

0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Probability of trait within class
High SES younger married adults
Low SES older widowed females

42

DISCUSSION
Chapter III uses latent class analysis to answer a similar question that was posed in
Chapter II: What are the characteristics associated with being a fraud victim age 50 and older in
the U.S.? Based on the logistic regression results in Chapter II, one may conclude that victims
are more likely to be middle-aged (closer to age 50), male, wealthy, educated, cognitively intact,
widowed (relative to being married or partnered), and have experienced more stressful life events
in the previous 5 years compared to non-victims. Using a different statistical procedure
however, LCA results present a different story. Rather than just one set of uniform
characteristics associated with fraud, we find that there are actually two distinct victim
typologies that fall on opposite ends of the socioeconomic/demographic spectrum. These two
groups were descriptively labeled high-SES middle-aged married adults and low-SES older
widowed females based on the conditional probabilities of observed characteristics within each
group.
In addition to socioeconomic differences, the values for average cognitive functioning
and stressful life events differed between classes. The logistic regression showed that higher
cognitive functioning and more stressful life events significantly predicted fraud, yet the
conditional means for these risk factors were lower for low-SES older widowed females than they
were for both high-SES middle-aged married adults and non-victims (see Table 2.1).
The differences between logistic regression and LCA results provide a possible
explanation for discrepancies in fraud risk factors reported in the literature (e.g. Lee & Soberon-
Ferrer, 1997 versus Titus, Heinzelmann, & Boyle, 1995). Including all victims in the same
logistic regression equation obscures the unique differences between groups on specific
countervailing risk factors, such as middle-age and widowhood. Only the most prevalent

43

characteristics associated with fraud reach a level of significance, thereby painting a somewhat
contradictory picture of the type of older person who is at risk. LCA offers an alternative
explanation by modeling unobserved heterogeneity in the victim subsample. Most of the factors
that were significant in the logistic regression model are reflected by individuals in Class 1. The
most prevalent victim group—84.4% of victims—is categorized by a high probability of being in
their late 50s, married, with middle and upper levels of income and education, yet there is also a
smaller group of victims (~16.6%) who are more likely to be older (70+), female, widowed (or
still married), with low-education and low-income.
Using a national frame, this analysis provides comparable results to earlier studies that
identified two victim typologies using a smaller sample of victims identified by law
enforcement: “bogus prize promotion” and “investment fraud” victim (Pak & Shadel, 2011;
FINRA, 2006), which are named after the type of fraud each group was victimized by. The
distribution of characteristics in the two latent classes are consistent with these typologies: Low-
SES older widowed females share a socio-demographic profile with bogus prize promotion
victims and high-SES middle-aged married adults share a socio-demographic profile with
investment fraud victims. This implies that the unmeasured “latent” variable differentiating the
two victim groups in the HRS is type of fraud, yet this hypothesis cannot be tested because the
survey did not ask respondents to describe how they were victimized.
Previous research by the FTC found that for incomes below $100,000 there is roughly a
U-shaped association between income and fraud, such that the rate of victimization is greatest at
both high and low levels of income compared to middle income levels (Anderson, 2013). While
the logistic regression results in this study show that those in the highest income quartile
experience the greatest likelihood of victimization, LCA results support the U-shaped curve

44

hypothesis: while the larger group of victims in Class 1 has high SES, there is also a smaller
group of victims in Class 2 who are more likely to be poor.
When cognitive functioning and stressful life events were regressed on latent class rather
than included as indicator variables for determining class differentiation, both were significantly
associated with Class 1 (Class 2 was the reference group). This suggests that cognitive
functioning and stressful life events are not universal risk factors for fraud, and that low-SES
older widowed females and high-SES middle-aged married adults may be vulnerable to fraud for
different reasons. If this is the case, fraud prevention programs will be more effective if they
specifically address the risk factors that are unique to each sector.
CONCLUSION
This study advances the literature by identifying victim subgroups in the general U.S.
population over age 50 who reported fraud. Chapters II and III underscore the advantage of
using multiple statistical approaches to better understand victimization of middle-aged and older
Americans by fraud, a phenomenon that is still lacking in conceptual clarity and theory
development. While logistic regression can identify general trends, LCA is advantageous
because it groups individuals who share a particular experience, such as fraud, but who differ in
other characteristics that may predispose them to that experience, such as level of income and
education. Logistic regression was not be able to identify the heterogeneity among these
individuals.
Findings support the current understanding among researchers that there is no single type
of fraud victim (Jackson & Hafemeister, 2011). More likely, fraud schemes are specifically
targeted to reach particular population groups with distinct socioeconomic/demographic profiles.

45

This has noteworthy implications for public policy because it indicates that many types of people
need protection. Further investigation into the risk factors specific to each victim class will
enhance prevention programs to improve the overall safety and financial security of older
Americans. Given that the largest group of fraud victims in Class 1 (high-SES middle-aged
married adults) were younger than previous research and popular assumptions suggest,
prevention efforts must begin earlier in the life course to be effective.

46

CHAPTER IV: QUALITATIVE CASE FILE ANALYSIS
INTRODUCTION
Chapters II and III used national survey data from the HRS to identify the socioeconomic
and psychosocial characteristics of fraud victims aged 50 and older. Results indicated that fraud
victims in the HRS fall into two groups. The larger group is comprised primarily of cognitively
intact adults who are more likely to be middle-aged, married, with mid to high levels of income
and education, and the smaller group is comprised of individuals who are more likely to be
female and widowed, age 70 and older, with low education and low income.
Although the HRS sample is representative of people over age 50 in the United States, it
is not representative of financial exploitation victims whose cases receive APS and/or law
enforcement services. Forty-six states and the District of Colombia have laws that mandate
professionals, such as healthcare providers and caregivers, to report suspected abuse and neglect
of vulnerable adults in the community to APS, even if the victims themselves do not
acknowledge being mistreated. Thus, victims who receive abuse intervention services tend to be
much older and more cognitively impaired than HRS participants and generally do not report the
fraud themselves. In fact, when interviewed, many APS clients refuse to acknowledge being
victimized and are unaware of losing money.
As discussed in Chapter II, a limitation of the HRS is that respondents were not asked
about the nature of the fraud and how much money they lost, if any. There is no information on
how the victims were contacted by the scam artist(s), if a trusting relationship developed, and
how, exactly, they were vulnerable to fraud in the first place. There is also no indication of

47

whether HRS respondents, or others, reported the incident to fraud complaint agencies and other
authorities such as APS and the police.
Given these limitations, the purpose of Chapter IV is twofold. The first objective is to
gain a deeper understanding of victims of fraud and financial exploitation by analyzing cases
(n=63) that were investigated by the Los Angeles County Elder Abuse Forensic Center (Forensic
Center). Because of the Forensic Center’s in-depth investigation and documentation process,
Forensic Center cases offer rich data on exploitation, including information on victims and
perpetrators, and narratives describing the crime, and the inter-agency response to allegations.
The Forensic Center case files contain far more detail about the event than is available in
the HRS survey. Case files include evaluations by medical practitioners, neuropsychologists,
APS workers, and law enforcement officers who document information on the victims’ physical
and mental health, cognitive functioning, social behaviors, and living environment, as well as
facts about the perpetrator and the amount of assets that are stolen. Analysis of this qualitative
case data will provide a greater account of how elder financial exploitation transpires, including
data on the etiology of fraud that was not collected in survey research.
The second objective of this study is to compare victims of fraud to victims of financial
abuse by a “trusted other.” Half of the case files in the sample involve financial abuse by a
family member, long-time friend, neighbor, or a paid caregiver. The other cases involve fraud
by a stranger or someone who befriended the victim for the sole purpose of exploitation.
Existing conceptual frameworks categorize fraud against older people as a type of elder financial
abuse, and in doing so they minimize potential differences between victims, perpetrators, and the
psychosocial and environmental context of exploitation. This has important implications for

48

determining influential risk factors, and also for developing effective prevention and intervention
programs.
A review of the elder abuse and financial fraud literature failed to identify studies that
used qualitative methods to analyze case investigation notes, clinician reports, and victim
interviews to create a comprehensive profile of those who are exploited by fraud. There are also
no studies examining how fraud victims differ from elders exploited by people in their close
social networks (“trusted others”). Using case data from real crimes, this study informs our
conceptualization of fraud and may help identify ways to prevent it.
BACKGROUND AND SIGNIFICANCE
To what extent do financial abuse and fraud victims share similar characteristics?
A number of frameworks describing elder financial exploitation are presented in the
literature. Some of these explicitly present fraud by strangers as distinct from financial abuse by
trusted others (e.g., Conrad, Iris, Ridings, Langley, & Wilber, 2011; Wilber & Reynolds, 1996),
whereas other frameworks incorporate fraud within the general financial exploitation paradigm
(e.g., Rabiner, O’Keeffe, and Brown, 2005). Although the four models offer varying levels of
specificity, common among all of them is the vulnerable status of the older adult, the
perpetrator’s sense of entitlement, and the unequal balance of power and status between the two
parties.
The first conceptual framework for financial exploitation was proposed by Wilber and
Reynolds in 1996. They describe four interacting criteria to distinguish elder financial
exploitation from acceptable monetary exchange: (1) vulnerability characteristics of the elder, (2)
the nature of the relationship between the elder and the perpetrator, (3) the interests served by the

49

relationship, and (4) the type of influence used by the perpetrator. The second model by
Rabiner, O’Keeffe, and Brown (2005) presents an applied ecological perspective of financial
exploitation in which the victim and perpetrator are embedded within a larger sociocultural and
policy macrosystem. The macrosystem includes cultural perceptions of aging, public policies,
state statutes, and interventions to protect elders from exploitation. The microsystem includes
the victim-perpetrator dyad, the social networks of each person, and power and exchange
dynamics underlying the relationship.
A third conceptual framework by Kemp and Mosqueda (2005), which built on Wilber
and Reynolds (1996), was developed and validated with input from police detectives, district
attorneys, and Adult Protective Service (APS) workers. The model lays out eight process-related
features of exploitation describing how a vulnerable elder is persuaded to provide money,
property, or other assets to another person when it is not in the elder’s best financial interests. In
the most recent conceptual framework by Conrad and colleagues (2011), experts were recruited
to generate a concept map of financial exploitation using a statement sorting procedure. This
yielded six ranked clusters: (1) theft and scams, (2) coercion, (3) financial victimization, (4)
signs of possible financial abuse, (5) financial entitlement, and (6) money management
difficulties. Of all clusters, statements pertaining to theft and scams were rated as the most
severe.
These frameworks describe the general types of elder financial exploitation and the
associated risk factors, but they do not explain the specific processes of fraud that sets it apart
from elder financial abuse by a trusted other. It can be argued that there are major theoretical
and conceptual differences in victim-perpetrator dynamics, personal risk factors, the context in
which financial transfers occur, and the tactics that are used. Importantly, fraud against

50

individuals involves misrepresentation, deception, and some degree of compliance from the
victim, whereas financial abuse by a trusted other does not require the elder to have knowledge
of or to agree to the financial transaction.
The perpetrators of financial abuse, particularly their relationship with the older adult,
also differ from perpetrators of fraud. Those who orchestrate scams deliberately target their
victims. They may enter professions where they have access to vulnerable older people, such as
personal care assistance, estate law and financial planning. As opposed to financial abusers,
fraud perpetrators can live close (e.g., paid caregiver) or far (e.g., telemarketers) from their
targets and may pursue one to multiple targets at the same time. Alternatively, family members,
friends and neighbors who take advantage of elderly loved ones often commit the crime because
there was an opportunity to do so and they had easy access. Some may feel entitled to the money
in exchange for providing care, whereas others justify their behavior as simply cashing in on
their inheritance early. In other words, financial abusers do not actively seek out victims for
monetary gain; rather, they find themselves in opportune positions where they can financially
benefit from their relationship with an elder. Because of these important distinctions, financial
exploitation frameworks may be inadequate in describing the concepts, mechanisms, and risk
factors of fraud.
Undue influence frameworks may conceptualize fraud better than financial exploitation
frameworks. Undue influence of a vulnerable older adult occurs when one person uses his or her
position of power to exploit the elder’s trust and dependency in order to benefit at the elder’s
expense (Quinn, 2000). Although they use slightly different terminology, undue influence
models by Blum (2003) and Bernatz (2008) describe how susceptible targets are isolated by
perpetrators who act to control elders’ social environments, thereby creating a sense of

51

powerlessness and increased dependency. As elders begin to rely more and more on perpetrators
for functional and emotional support and for access to information, they become easier to
manipulate to consent to financial transactions that ordinarily they would not agree to. These
theoretical models need to be evaluated in terms of their relevance to fraud of older adults.
One hypothesized difference between victims of fraud and financial abuse is cognitive
functioning and dependency. Researchers and practitioners have suggested that elders may be
most vulnerable to scams during the early stages of decline when cognitive deficits are hardest to
recognize and when family members are not yet involved in helping the elder manage his/her
estate (Karp, 2012; Martin et al., 2008). Elders in the early stages of decline may exhibit minor
symptoms of impaired problem solving, trouble learning new information, difficulties making
complex transactions, and changes in personality, yet they may function normally in other areas
of cognition, like language and semantic knowledge, and need no assistance with personal care
(Triebel & Marson, 2012; Moye & Marson, 2007). This combination of subtle impairments in
financial capacity and lack of oversight from family members provides a “window of
opportunity” for scam artists. These concepts are depicted in the model presented in Chapter I
(Figure 1.1).
Alternatively, it’s hypothesized that financial abuse by friends and family occurs after the
elder has lost the capacity to manage his or her finances independently. In many cases the abuser
is an adult child or close friend/relative appointed to manage the person’s financial affairs and
who has decided to take advantage of this position of trust. The elder may also depend on that
individual to also assist with personal care needs, such as help with activities of daily living
(ADLs) and instrumental activities of daily living (IADLs). To determine whether fraud and
financial abuse victims differ in cognitive functioning and dependency, this study tests whether

52

victims of financial abuse are more physically and cognitively impaired and more dependent on
others for care than victims of fraud. The analysis also examines the extent to which financial
abuse victims cognitively differ from fraud victims and in what areas of cognition: memory,
financial decision-making ability, attention and processing.
The Elder Abuse Forensic Center
The data in this study derive from The Los Angeles County Elder Abuse Forensic Center,
an abuse intervention that brings together professionals from various fields to provide expert
case examination, consultation, and documentation of abuse, neglect, and financial exploitation.
Members include APS, LA Police Department, LA Sheriff’s Department, LA District Attorney’s
Office, Victim/Witness Assistance Program, LA City Attorney’s Office, LA County Office of
the Public Guardian, LA County Department of Mental Health (GENESIS), forensic
neuropsychologists, and USC Keck School of Medicine. Two to four cases of abuse are
presented weekly, typically by law enforcement officers or APS workers. Team members work
synergistically across disciplines to identify common goals for resolving each case and to decide
on a course of action. Remedies might include conservatorship to protect elders and their assets
or a criminal justice response to prosecute the abuser (Navarro, Yonashiro, Homeier & Wilber,
2010).
So far, the Forensic Center MDT is the only empirically tested elder abuse intervention
with significant positive outcomes. Compared to financial exploitation cases in Los Angeles
County that did not receive the Forensic Center intervention, cases seen by the Forensic Center
have ten times greater odds of submission to the district attorney for criminal prosecution of
abusers (Navarro, Gassoumis, & Wilber, 2013), and eight times greater odds of referral to the
public guardian for conservatorship protection of elders (Navarro, Gassoumis, & Wilber, 2014).

53

Importance of capacity evaluations
In financial exploitation cases it is often necessary to evaluate an alleged victim’s
financial decision-making capacity (Naimark, 2001) because even capable and seemingly sharp
older adults have been reported to give away valuable assets and personal information to
convincing strangers (Quinn, 2000). Documentation of financial capacity is important because it
provides evidence that the elder did not understand what was being agreed to, or did not fully
understand the consequences of the transaction. Despite the importance of capacity evaluations,
there are a lack of experts to administer these tests and to testify on the elder’s behalf (Connolly,
2010). The Forensic Center helps by sending clinicians to evaluate the living environment,
cognitive functioning, and financial decision-making capacity of alleged abuse victims through
administering a flexible battery of tests in the victim’s home setting. Tests are selected based on
the elder’s perceived cognitive ability, mental status, financial capacity concerns, and the clinical
judgment of the evaluator. The goal is to determine the elder’s probable financial decision-
making capacity at the time of the crime and whether s/he reasonably understood the risks and
benefits of the transaction to provide informed consent. If the elder is determined to have lacked
financial capacity at that time, the findings are often presented to the District Attorney’s Office
and/or the Public Guardian’s Office to pursue criminal prosecution of the abuser or
conservatorship (guardianship) of the elder (Gassoumis, Navarro, & Wilber, 2014).
Using data from twenty-seven of the Forensic Center cases that had neurocognitive
evaluations completed by Forensic Center clinicians between May 2008 and 2011, a pilot study
indicated that victims of financial exploitation performed significantly worse on tests of
processing speed, attention, financial decision-making and memory compared to a control
sample of 32 healthy older adults (Wood et al. 2014). When the study was conducted, only 375

54

total cases had been seen by the Forensic Center; over a quarter of them did not involve financial
exploitation, and of those cases, less than 20% contained neurocognitive testing. Also, the pilot
study did not specifically examine victims of fraud, who may differ in functioning from those
who are exploited by a trusted other. The present study expands the time frame of the pilot study
to include cases presented between March 2006 and August 2013, and compares victims’ scores
on the Mini Mental State Exam (MMSE; Folstein & Folstein, & McHugh, 1975) and the Money
Management subtest of the Independent Living Scales (Loeb, 1996).
In addition to administering the neurocognitive battery during the home visit, Forensic
Center clinicians often document the condition of the living environment and the elder’s physical
appearance, mental status, and how they interact with the evaluator. Their summary reports
provide substantial information on the elder’s dress and physical appearance, signs of possible
mental illness (e.g., depression, paranoia), and style of social engagement (e.g., friendly versus
withdrawn). Some victims are administered the Geriatric Depression Scale (GDS; Montgomery
& Asberg, 1979) to assess depression. Forensic Center case files also contain meeting notes
(“minutes”) from the team’s review of the allegations and any follow-up activities that were
carried out. These evaluations may provide detail and support for the risk factors of fraud
identified in the quantitative analyses using the HRS in Chapters II and III.
Many case files also contain victims’ medical records, including physician and clinic
notes on past hospitalizations, chronic conditions, medications, cognitive status, mobility,
functioning, and mental health. These additional documents are important to the analysis
because they provide a holistic presentation of the victim—physiological and cognitive
characteristics in addition to mental health and substance use. They also offer a unique data
source that was not utilized in the pilot study.

55

HYPOTHESES

Grounded research is needed to explore key differences between financial exploitation
and fraud so that policy-makers and practitioners can combat each problem effectively. Based
on the proposed model for fraud presented in Chapter 1 and the results of the analyses in
Chapters II and III, the following three hypotheses will be tested using the Forensic Center case
files:
(1) Victims of fraud present with better cognitive functioning than victims of
financial abuse by “trusted others”—a.k.a., relatives, friends, neighbors, and well-
known paid caregivers.
(2) Fraud victims are less dependent on others for care and support than victims
of financial abuse by trusted others.
(3) Elders experience fraud during a “window of opportunity” when subtle declines in
cognition impair financial decision-making and there is no financial oversight from
trusted others.

This study tests these hypotheses using a mixed method approach using quantitative
neurocognitive data and qualitative case data on the characteristics of victims, perpetrators, and
the crime itself. Characteristics include cognitive, physical and mental health status, social
isolation, relationship with the perpetrator, method of coercion, and many others. The objective
of this mixed-methods multiple case analysis is to generate a broad psychosocial, cognitive, and
contextual profile of fraud victims and to understand how they differ from financial abuse
victims. Results will add to the evidence on what factors are linked with fraud susceptibility, and

56

may offer a clearer target for allocating resources that protect the health, financial security, and
safety of elders.
METHODS
Sampling frame: All cases presented between March 30
th
, 2006 and August 8, 2013—a
total of 924 cases—were included in the sampling frame. The majority of these cases (72%)
involved financial exploitation of an elder or dependent adult. Approximately 120 of these cases
(13% of total) involved some fraud by a stranger. Elder abuse investigations in LA County
generally begin with a report to APS or law enforcement. These agencies begin investigating the
allegations before bringing the case to the Forensic Center if a more coordinated response is
needed. Generally, two to four cases are presented each week.
Cases were initially identified using an existing database developed under a National
Institute of Justice (NIJ)-funded study on the overall effectiveness of the Forensic Center (see
Navarro, Gassoumis, & Wilber, 2013). This database includes case intake information: socio-
demographic information about the alleged victim, the type of abuse/exploitation, characteristics
of the abuser, as well as the APS case worker’s assessment of whether the victim presents with
any suspected cognitive impairment. The database also identifies whether a neurocognitive
battery was administered to the victim by the Forensic Center geriatrician or neuropsychologist.
Each Forensic Center case in the electronic database has a unique ID number that is linked to a
hard copy file stored in a locked file cabinet at the Forensic Center.
Inclusion/exclusion criteria: The sample was divided into fraud and financial abuse
cases, with an equal number of cases in each category. Only cases involving an elder (age 65
and older) were included. Cases involving dependent adult victims ages 18-64 were excluded.

57

Only cases where the allegations of fraud or financial abuse were substantiated were included in
the analysis. A substantiated report means that the case worker who conducted the investigation
decides that the allegations constitute elder financial exploitation based upon evidence that
makes it more likely than not that exploitation occurred (California Penal Code Section
11165.12).

Table 4.1. Criteria for selection into sample
Inclusion Criteria Exclusion Criteria
1. Crime involves financial abuse by trusted
other (n=31) OR fraud by non-family
member (n=31)
1. Crime involves BOTH financial abuse
and fraud
2. Victim is 65 years old or older 2. Victim is a dependent adult (less than
age 65)
3. Victim is vulnerable (criteria for receiving
APS services)
3. Case was never substantiated by
Forensic Center investigation team
4. Case contains intake information from APS
and/or law enforcement
4. Case only includes a declaration of
capacity by victim’s physician
5. Case includes report on victim’s cognitive
status from forensic neuropsychologist

6. Case contains documentation on victim’s
health status

7. Case contains sufficient documentation on
the crime

Based on the GAO (2012) report on elder financial exploitation, the types of cases
included in the sub-sample of fraud cases could involve, for example, (1) sale of fraudulent
investments or products/services unsuitable for an older adult’s circumstances (such as long-term
annuities for elders of very advanced age); (2) lottery, mail, telephone, or internet prize
promotion/sweepstake scams; (3) door-to-door home repair fraud; (4) real estate fraud; and (5)

58

predatory sweetheart scams where a perpetrator initiates a romantic relationship with an elder for
the purpose of financial gain.
The comparison sample of financial abuse cases includes crimes where the perpetrator is
a “trusted other,” such as a family member, hired caregiver or close friend/neighbor that has a
long-standing relationship with the victim. Many of these crimes involve abuse of fiduciary
duty, whereby the individual who was trusted to manage the elder’s finances (formally or
informally) makes decisions that are not in the elder’s best interest (Wilber & Reynolds, 1996).
Forms of financial abuse could involve, (1) misappropriation of an elder’s bank account funds or
credit cards; (2) inappropriate transfer of real property; (3) misuse of an elder’s income or assets;
and (4) altering an elder’s will/trust to benefit the perpetrator.
There is a high co-occurrence of abuse types in cases presented to the Forensic Center, as
there is in elder abuse in general (Bonnie & Wallace, 2003). Selecting isolated cases of fraud or
financial abuse would severely limit the size of the sample. Thus, cases where the older adult
was a victim of fraud or financial abuse in addition to another type of abuse (e.g., physical abuse,
caregiver neglect, self-neglect) were included. However, cases in which the elder was a victim
of both fraud and financial exploitation were excluded.
Only cases in which the victim was evaluated by a forensic neuropsychologist were
included in the analysis. These evaluations involve multiple validated cognitive instruments that
can identify specific areas of cognition (e.g., processing speed, attention, financial competency)
that are impaired. An evaluation by the forensic neuropsychologist typically includes the Money
Managing Subtest of the Independent Living Scales (Loeb, 1996), and the Mini-Mental State
Exam (MMSE; Folstein & Folstein, & McHugh, 1975). The MMSE measures orientation to
place, time and date, immediate and delayed recall, attention, and calculation via serial seven

59

subtraction. The Managing Money subtest is a 17-item scale that assesses checkbook balancing,
making change, and bill payment. In addition to these two tests, some victims are administered
the Trail Making A and B, and sub-sections of the Repeatable Battery for the Assessment of
Neuropsychological Status (RBANS; Randolph, Tierney, Mohr, & Chase, 1998), which includes
the Digit Span, Animal Naming, list recall, FAS (controlled oral word association), and figure
drawing. Each of these tests is widely used in clinical practice to determine a person’s abilities
in multiple areas of cognition, however, only the MMSE and Managing Money Subtest are
consistently administered.
Cases that only include a capacity declaration by the elder’s physician for the purpose of
conservatorship were excluded. Capacity declarations do not involve the use of standardized
instruments to evaluate functioning and therefore lack the specificity needed for this analysis.
Case selection procedure: This study used purposive multiple case sampling to ensure
representativeness of the data (Miles, Huberman, & Saldana, 2013). According to Yin (2009),
multiple case sampling increases the precision, validity, stability, and trustworthiness of findings
in qualitative research. In order to remain “blind” to the type of financial exploitation (fraud
versus financial abuse) during data extraction and coding, case selection was aided by the
Forensic Center’s Project Manager. The Project Manager used the existing NIJ database to
select cases according to the inclusion criteria provided, and placed information on the crime at
the end of each file so that it would be coded last. The goal was to prevent bias in how the data
were coded. The project manager ensured that case selection was balanced, that is, that an equal
number of financial abuse and fraud cases were selected into the sample.
Case selection began with Forensic Center case ID #066 and continued forward
chronologically until 30 fraud cases and 30 financial abuse cases were coded. Case #066 was the

60

first fraud-specific case presented to the Forensic Center within the analytic time frame that
contained a neurocognitive assessment and adequate information on the victim and the crime.
Additional case selection was driven by the quality and quantity of documents contained in
selected files that met criteria for inclusion. The aim was to select a sufficient number of cases
to reach the point of theoretical “saturation” (Strauss and Corbin, 1998). Saturation refers to the
process of adding new cases to the sample until no new insights or concepts emerge (Strauss and
Corbin, 1990), or where there is sufficient replication or redundancy of information to answer the
research questions (Bowen, 2008). Saturation also ensures that the researcher has captured the
full range of variation in the concepts being explored and that the sample is representative
enough to draw valid conclusions about the research hypotheses.
Sample size: While there are no definitive rules for determining how many cases are
needed to reach saturation (Morse, 1995), analyzing all 31 Forensic Center fraud cases (all fraud
cases that met the inclusion criteria), and 31 financial abuse cases, was determined to be a
sufficient sample size to cover the scope of exploitation scenarios and explore each phenomenon
in-depth. Given that files contain between 5 and 35 documents that range between 1 to 30 pages
each, 62 total files out of the 924 available cases presented at the Forensic Center was feasible
for the time frame of the study.
After these cases were selected and coded, four graduate student researchers and one
faculty member who was an expert in elder abuse met to discuss cases in which the crime was
ambiguous and could not be definitively assigned to either the “fraud” or “financial abuse”
category. Of the 62 coded cases, nine were ultimately excluded from the final analysis sample.
Five were dropped because the researchers could not determine whether the case involved fraud
or financial abuse, one was excluded due to missing data, one was excluded because the crime

61

involved theft (not fraud or financial abuse of a vulnerable elder), and two were excluded
because the crime was not substantiated by the Forensic Center investigators. Thus, the final
analysis included 53 cases—28 were fraud and 25 were financial abuse.
Coding strategy: Codes are defined as, “labels that assign meaning to the descriptive or
inferential information compiled during a study” (Miles, Huberman, & Saldana, 2013; pg. 71).
The purpose of coding is to identify, retrieve, and consolidate relevant information in and across
documents. Coding divides large amounts of qualitative data into manageable pieces that can be
reviewed and compared across cases (Saldana, 2013).
This study used a hypothesis coding strategy (Saldana, 2013) in which the researcher
begins with a predetermined list of codes based on the theoretical framework and what concepts
need to be explored. Hypothesis coding is better suited for testing predictions because it follows
a deductive analytic approach rather than an inductive approach and is best for seeking support
or counter evidence for study predictions about the differences between financial abuse and fraud
victims.
Many qualitative analysis projects use analytic computer software like Atlas.ti or NVivo
to code electronic versions of text documents. No content categorization software was used in
this study for two reasons. First, many of the documents—social worker notes, medical records,
etc.—are hand-written and the software is unable to recognize script. Second, converting
documents into electronic files could potentially compromise victims’ personal identifying
information by making it virtually accessible. Therefore, rather than electronically coding the
case documents, the information was extracted from the files by typing key words and phrases
from the files into an Excel spreadsheet. The spreadsheet was used to organize, sort, summarize
and display the data. Each row in the spreadsheet corresponded to single a case (victim), and

62

columns represented the predetermined codes for hypothesis testing, such as “victims’ medical
history”. Example information that would be entered into this coding category include, “high
blood pressure,” “stroke in 2003,” and “decubitus ulcer on right heel.” Other contextual and
psychosocial factors for testing the hypotheses—such as victims’ marital and family status,
knowledge of finances, physical appearance and the condition of the home—were also extracted
and entered into the Excel columns. New columns were added as additional concepts emerged
during the file review. The final database contained a total of 142 columns (codes).
Data extraction from Forensic Center case files was supervised by the Forensic Center
Project Manager. No information that could identify elders, reporters of abuse, or alleged
abusers was extracted. Cases in the database were identified using their original Forensic Center
intake number so that they could be linked to the hardcopy file. They were also assigned a new
number for study purposes (ID #01-62). Personal identifying information was secured by
keeping all documents in locked file cabinets at the Forensic Center between coding sessions.
Data extraction took place at a secure workstation within the Forensic Center meeting room.
Cases were coded sequentially until all relevant information was extracted and entered into the
Excel database. The study protocol was reviewed and approved by the University of Southern
California Institutional Review Board prior to data extraction.
Validity, dependability & credibility: To improve transferability of findings, sampling
was designed to include a range of victims and mechanisms of financial exploitation that were
available in the Forensic Center case files. An important step in qualitative research is to assess
the validity, or “trustworthiness,” of conclusions, with the intent to produce findings that are both
plausible and confirmable (Miles, Huberman, & Saldana, 2013). This required ensuring that
protocols were carried out consistently throughout the project, and that identified patterns

63

accurately represented the original case narratives (a.k.a., remained “true” to the actual data).
Quality checks were carried out intermittently throughout the project by soliciting feedback on
the findings from fellow graduate students and Forensic Center team members who are experts in
elder abuse.
Hypothesis testing with neurocognitive Data: To test the first hypothesis that fraud
victims are less cognitively impaired than financial abuse victims, the unstandardized
neurocognitive testing scores from each victim were extracted from the case files and entered
into Excel. There was significant variation in which neurocognitive tests were administered to
victims resulting in substantial missing data. Only the MMSE and the Managing Money subtest
of the Independent Living Scales were consistently administered and could be compared using
Poisson regression. Poisson was used to account for the non-normal distribution of victim’s
MMSE and Managing Money scores. The sample clustered toward low scores—poor
performance on these tests—so OLS regression was deemed inappropriate. The independent
variable of interest the regression analyses was whether the elder had been a victim of fraud or
financial abuse, adjusting for age and education.
Hypothesis testing with qualitative data: Once all relevant information was extracted
from the cases, some of the information within the coding categories was further condensed by
assigning it a numerical value. For example, “overall health” was recoded on a scale of 0 to 3,
where 0=good health and 3=very poor health. Thus, a victim who had a long list of chronic
illnesses, multiple hospitalizations, and poor mobility would have his/her health recoded
numerically as a “3.” This process of consolidating information is called magnitude coding. In
magnitude coding the researcher adds an alphanumeric code to indicate the intensity, frequency,
direction, presence, or evaluative content of qualitative information (Saldaña, 2013, p. 72-73).

64

Recoding this information numerically helped summarize patterns across cases and made it
possible to do quantitative analyses with primarily qualitative data.
After some cells were recoded numerically, the sample was divided into two groups:
fraud and financial abuse. Values were compared across groups to identify differences between
fraud and financial abuse victims and to test Hypotheses 1 and 2. The objective was to
determine what characteristics in health and functioning are common among victims of financial
abuse and fraud, and what characteristics are different.
Fisher’s exact test was used to determine if there were significant differences between
financial abuse and fraud victims in quantitatively recoded measures of memory functioning,
financial capacity, overall health, dependency, mobility, quality of the care environment, social
isolation, and marital status. For example, qualitative information on financial capacity was
recoded as 0=mildly impaired, 1=moderately impaired, and 2=severely impaired, for each
victim. Dependency was recoded as 0=independent/little assistance needed, 1=some IADL/no
ADL assistance needed, 2= IADL/some ADL assistance needed, and 3=fully dependent.
Fisher’s exact was selected as the appropriate analysis to identify differences in distributions
between fraud and financial abuse victims because many items had expected cell sizes of less
than five.
Hypothesis 2 posited that victims of fraud are less dependent on others for care and
support than victims of financial abuse. To test this prediction, the subsamples were compared
on the following factors: (1) To what extent does each victim-type have physical and cognitive
disabilities that may interfere with routine activities and self-care? (2) To what extent does each
victim-type have others available to assist with these activities? And (3) How socially active or

65

isolated from friends, family members and community members are each victim-type? (Do they
live alone, do they have children, do they attend religious services, etc.)?
Another objective of the study was to determine whether there is a window of
opportunity when elders are most vulnerable to fraud due to mild cognitive impairment and lack
of financial oversight by trusted others (Hypothesis 3). To test Hypothesis 3, fraud cases were
compared to identify common processes, narratives, and patterns. The objective was to
determine whether the majority of fraud victims had impairments in decision-making at the time
they were victimized and whether they had family members or others to assist them during that
period. The goal of this subsample analysis was to summarize the chronological flow of events
to determine how factors that purportedly increase vulnerability, such as cognitive decline and
social support, relate in time and affect the process of victimization.
RESULTS:
Victim characteristics:
Table 4.2 presents the demographic characteristics of fraud and financial abuse victims in
the sample. While the average age of both groups is similar—83.3 and 84.1 years—there are
differences in gender although they were not statistically significant (p=0.32): Over fifty-four
percent of fraud victims were male, compared to 44% of financial abuse victims. Sixty percent
(n=17) of fraud victims were non-Hispanic White compared to 48% (n=12) financial abuse
victims (p=0.26).

66

Table 4.2. Forensic Center victim characteristics

Financial Abuse
Victims (n=25)

Fraud Victims
(n=28)
Count Frequency

Count Frequency
Average age (St.Dev.) 25 84.1 (6.8)

28 83.3 (5.0)
Sex (female) 14 56%

13 46.4%
Race

Non-hispanic White 12 48.0%

17 60.7%
Black 6 24.0%

4 14.3%
Hispanic 4 16.0%

3 10.7%
Asian/Pacific Islander 1 4.0%

4 14.3%
Other 2 8.0%

0 0.0%
Education

Less than high school 6 24.0%

2 7.1%
High school graduate 5 20.0%

13 46.4%
Some college (1-3 yrs) 6 24.0%

6 21.4%
Trade School 3 12.0%

1 3.6%
College graduate (4 yrs) 1 4.0%

1 3.6%
Advanced degree (Master's/PhD) 3 12.0%

5 17.9%
Unknown 1 4.0%

0 0.0%
Marital Status

Single/never married 4 16.0%

8 28.6%
Married/Partnered 2 8.0%

4 14.3%
Widowed 17 68.0%

11 39.3%
Divorced/Separated 1 4.0%

3 10.7%
Married to/Dating the SA 0 0.0%

2 7.1%
Unknown 1 4.0%

0 0.0%
Living Arrangement at interview

Alone 8 32.0%

11 39.3%
Own home with spouse/partner 1 4.0%

2 7.1%
Own home with other family/friends 4 16.0%

3 10.7%
Own home with paid caregiver 1 4.0%

2 7.1%
SNF/AL/board and care 5 20.0%

8 28.6%
Own home with SA (relative) 2 8.0%

0 0.0%
Own home with SA (non-relative) 4 16.0%

2 7.1%
Lived with SA at time of exploitation 9 36.0%

4 14.3%
Multiple perpetrators 3 12.0%

4 14.3%

Note: SA is "suspected abuser"; SNF is “skilled nursing facility”; AL is “assisted living”

67

Nearly half of the fraud sample reported that high-school was their highest level of
educational attainment; six victims (21.4%) had some college education, and five victims
(~18%) had an advanced degree such as a Masters or Doctorate. Comparatively, five (20%) of
financial abuse victims reported high school as their highest degree, six (24%) did not graduate
from high school, and another quarter listed some college. Only three (12%) had an advanced
degree. These differences were not significant (p=0.21).
Although not statistically significant (p=0.19), a higher proportion of fraud victims
compared to financial abuse victims were married (28.6% versus 16%), and more financial abuse
victims were widowed (68% versus 39%). Of those who were widowed, the majority had been
without their partner for six or more years. Three fraud victims (7%) were dating the suspected
abuser, commonly referred to as a “sweetheart scam.” A greater number of financial abuse
victims lived with the suspected abuser at the time of exploitation compared to fraud victims:
nine financial abuse victims (36%) versus four fraud victims (14.3%.); and a higher proportion of
fraud victims lived in a long-term care residential setting, such as a skilled nursing facility,
assisted living, or board and care when they were interviewed by APS or the neuropsychologist.
Nearly 40% of fraud victims lived alone compared to 32% of financial abuse victims.
Differences in living arrangement were not statistically significant (p=0.70).
Perpetrator characteristics:
Perpetrator characteristics are presented in Table 4.3. There was less information in the
case files on perpetrators of fraud than perpetrators of financial abuse because none were related
to the victim and thus were harder to track and interview. Significantly more fraud perpetrators
were female compared to financial abusers (fifteen people versus seven people; p= 0.01).
Financial abusers were proportionally older: ten (40%) financial abusers were between ages 50-

68

59 and eight (32%) were between ages 40-49, compared to only six (21.4%) fraud perpetrators
between ages 50-59 and eight (28.6%) between ages 40-49. These differences were not
significant. Over a quarter of the fraud cases contained no documentation on the age of the
perpetrator. Both perpetrator groups were racially diverse: seven (25%) fraud perpetrators were
non-Hispanic White compared to nine (36%) financial abusers; three (10.7%) fraud perpetrators
and six (24%) financial abuse perpetrators were Black; eight (28.6%) fraud perpetrators and nine
(36%) financial abuse were Hispanic (p=0.23).
The majority of financial abuse perpetrators were related to the victim: 10 were children
(40%), two were grandchildren (8%), and two were other relation (8%). Four were neighbors
(16%), one was a long-time friend (4%) and three were roommates/tenants (12%). Fraud
perpetrators were more likely to be strangers (6 perpetrators; 21.4%), non-relative caregivers (4
perpetrators; 14.3%), business professionals such as estate planners, attorneys, and general
contractors (5 perpetrators; 17.9%) or new romantic partners (6 perpetrators; 21.4%). In
general, fraud perpetrators had known their victims for less than 5 years (average length of
relationship=2.75 years), whereas financial abusers generally knew their victims 11+ years or for
a lifetime if they were a relative. These differences in length of relationship were statistically
significant (p<0.00).

69

Table 4.3. Forensic Center perpetrator characteristics

Financial Abuse
Perpetrators (n=28)
Fraud
Perpetrators (n=25)
Count Percent Count Percent
Age group
29 and younger 1 4.0% 2 7.1%
30-39 0 0.0% 4 14.3%
40-49 8 32.0% 8 28.6%
50-59 10 40.0% 6 21.4%
60 and older 2 8.0% 0 0.0%
Unknown 4 16.0% 8 28.6%
Sex (female) 7 28.0% 15 53.6%
Both male and female perpetrators 1 4.0% 5 17.9%
Unknown 0 0.0% 2 7.1%
Race
Non-Hispanic White 9 36.0% 7 25.0%
Black 6 24.0% 3 10.7%
Hispanic 9 36.0% 8 28.6%
Asian/Pacific Islander 0 0.0% 3 10.7%
Other 1 4.0% 4 14.3%
Unknown 0 0.0% 3 10.7%
Relationship to Victim
Stranger 0 0.0% 6 21.4%
Son/Daughter 10 40.0% 0 0.0%
Grandchild/grandchildren 2 8.0% 0 0.0%
Other relative 2 8.0% 0 0.0%
Non-relative Caregiver 2 8.0% 4 14.3%
Business Professional 0 0.0% 5 17.9%
Tenant/roommate 3 12.0% 2 7.1%
New romantic partner 1 4.0% 6 21.4%
Neighbor 4 16.0% 2 7.1%
Friend (long-time) 1 4.0% 2 7.1%
Friend (new or acquaintance) 0 0.0% 1 3.6%
Length of relationship
Less than 1 year 0 0.0% 0 0.0%
1-2 years 2 8.0% 5 17.9%
3-5 years 2 8.0% 8 28.6%
6-10 years 1 4.0% 4 14.3%
11+ yrs 6 24.0% 6 21.4%
Lifetime 13 52.0% 2 7.1%
Unknown 1 4.0% 3 10.7%

70

Co-occurring abuse and financial losses
As shown in Table 4.4, 12 victims of financial abuse and 11 victims of fraud had been the
subjects of prior abuse reports to APS. One fraud victim had seven prior reports. Recurrent
abuse is indicative of cases brought to the Forensic Center for additional investigation because
victims often need more services than can be provided by APS or law enforcement alone.
Victims also experienced other forms of mistreatment that were reported to APS at the same time
as financial exploitation. Financial abuse victims in particular experienced high-levels of co-
occurring emotional abuse (6 victims; 24%), caregiver neglect (4 victims; 16%), and isolation (2
victims; 8%). The most frequent co-occurring abuse type for fraud victims was self-neglect (4
victims; 14.3%); otherwise, fewer fraud victims had additional reports of abuse.
Fraud victims experienced greater total losses than financial abuse victims from financial
exploitation. Of the 24 fraud victims who had data available, the estimated total cash losses was
nearly $15 million, approximately $619,000 per victim. Five of these victims lost a total of nine
properties (not included in the cash loss estimate). Financial abuse victims (n=20) lost a
combined total of almost $4 million dollars—$199,000 per victim—and six victims lost a total of
eight properties. Losses for financial abuse victims ranged from $5,000 to $275,000 with a
standard deviation of $251,173. Losses for fraud victims ranged from $1,700 to $5,000,000 with
a standard deviation of $1,142,295.

71

Table 4.4. Differences in co-occurring abuse types and financial losses
between fraud and financial abuse victims

Financial Abuse
Victims (n=25)
Fraud Victims
(n=28)
Count Percent Count Percent
Prior abuse reports 12 48.0% 11 39.3%
Physical abuse
Yes 1 4.0% 2 7.1%
None reported 24 96.0% 26 92.9%
Emotional/psychological abuse
Yes 6 24.0% 2 7.1%
None reported 19 76.0% 26 92.9%
Sexual abuse
Yes 1 4.0% 0 0.0%
None reported 24 96.0% 28 100.0%
Neglect by caregiver
Yes 4 16.0% 2 7.1%
None reported 21 84.0% 26 92.9%
Self-neglect
Yes 1 4.0% 4 14.3%
None reported 24 96.0% 24 85.7%
Isolation
Yes 2 8.0% 2 7.1%
None reported 23 92.0% 26 92.9%
Abduction
Yes 0 0.0% 1 3.6%
None reported 25 100.0% 27 96.4%
Estimated losses per victim 20 $198,887 24 $618,949
Total losses
$3,977,756
$14,854,795
Property Lost 6 8 properties 5 9 properties

Hypothesis 1: Financial abuse victims have poorer cognitive functioning than fraud victims
Cognitive functioning was divided into different cognitive domains based on the
neuropsychologists’ assessment of the victims’ performance in those areas. As shown in table

72

4.5, both financial abuse and fraud victims had high levels of impairment in areas of memory,
financial capacity, executive functioning, attention, learning, language, verbal fluency, and
reasoning. The following findings are based on reports of victims’ cognitive status derived from
the victims’ medical records and their neurocognitive testing scores on tests such as Trail
Making A, Trail Making B, RBANS, Figure Copy, and other neurocognitive batteries. Because
there was significant variation in which standardized tests were administered during the
neuropsychologists’ evaluations, the following findings are preliminary and present a general
overview of cognitive performance.
Table 4.5. Victim cognitive functioning

Financial Abuse
Victims (n=25)
Fraud Victims
(n=28)
Count Percent Count Percent
Documented history of cognitive impairment 16 64% 18 64%
Executive functioning
Not impaired based on evaluation 0 0.0% 2 7.1%
Evidence of Impairment 16 64.0% 19 67.9%
Mixed testing results 3 12.0% 4 14.3%
Unknown 6 24.0% 5 17.9%
Attention
Not impaired based on evaluation 1 4.0% 4 14.3%
Evidence of Impairment 9 36.0% 8 28.6%
Mixed testing results 5 20.0% 4 14.3%
Unknown 10 40.0% 12 42.9%
Visuospatial
Not impaired based on evaluation 2 8.0% 8 28.6%
Evidence of Impairment 14 56.0% 10 35.7%
Mixed testing results 2 8.0% 2 7.1%
Unknown 7 28.0% 8 28.6%
Learning
Not impaired based on evaluation 1 4.0% 3 10.7%
Evidence of Impairment 9 36.0% 10 35.7%
Mixed testing results 0 0.0% 0 0.0%
Unknown 15 60.0% 15 53.6%

73

Language
Not impaired based on evaluation 9 36.0% 14 50.0%
Evidence of Impairment 8 32.0% 7 25.0%
Mixed testing results 3 12.0% 3 10.7%
Unknown 5 20.0% 4 14.3%
Verbal Fluency
Not impaired based on evaluation 4 16.0% 1 3.6%
Evidence of Impairment 9 36.0% 16 57.1%
Mixed testing results 0 0.0% 0 0.0%
Unknown 12 48.0% 11 39.3%
Reasoning
Not impaired based on evaluation 0 0.0% 4 14.3%
Evidence of Impairment 11 44.0% 11 39.3%
Mixed testing results 1 4.0% 1 3.6%
Unknown 13 52.0% 12 42.9%
Current memory impairment
No memory impairment 0 0.0% 2 7.1%
Mildly impaired 2 8.0% 5 17.9%
Moderate impairment 8 32.0% 3 10.7%
Severely impaired 14 56.0% 16 57.1%
Mixed findings 1 4.0% 2 7.1%
Memory impairment onset
Less than 1 year prior 0 0.0% 1 3.6%
1-2 years prior 1 4.0% 6 21.4%
3-4 years prior 7 28.0% 4 14.3%
5-7 years prior 4 16.0% 5 17.9%
8+ years prior 2 8.0% 1 3.6%
Unknown 10 40.0% 7 25.0%
Current financial capacity
Full capacity to manage all finances 0 0.0% 1 3.6%
Capacity for only simple transactions 1 4.0% 3 10.7%
Mildly compromised capacity 2 8.0% 4 14.3%
Moderately compromised capacity 5 20.0% 5 17.9%
Lacks capacity to manage all finances 14 56.0% 15 53.6%
Inconclusive 3 12.0% 0 0.0%
Past financial sophistication
Excellent 0 0.0% 3 10.7%
Good 4 16.0% 9 32.1%
Okay/normal 1 4.0% 1 3.6%
Poor/depended on others 7 28.0% 6 21.4%
Unknown 13 52.0% 9 32.1%

74

The examiners noted that most victims had poor insight into their cognitive deficits and
care needs. Sixty-four percent of each group had a history of cognitive impairment that was
either documented in their medical records or mentioned by secondary sources such as family
members, friends, neighbors, and caregivers who were present at the neurocognitive interview or
spoke with APS at some point during the investigation. Both financial abuse and fraud victims
had evidence of impairment in executive functioning (64% and 68%, respectively), attention,
visuospatial abilities, list learning, verbal fluency, and reasoning. Most victims were oriented to
person, place, and situation, but half of fraud victims were incorrect on the date, and the vast
majority of financial abuse victims were not oriented to date. Although 20-50% of the sample
had missing data in various cognitive domains, the evidence is strong that both groups were
impaired in most areas of cognition.
Limited support was found for Hypothesis 1, that victims of financial abuse are more
cognitively impaired than victims of fraud. More data was available on memory functioning and
financial capacity, perhaps because these areas of cognition are important in determining
whether a victim understands and is able to consent to financial transactions or can reasonably
appoint another person to manage his or her financial affairs. Based on the distribution of
proportions within each group, financial abuse victims presented with slightly greater memory
impairment and poorer financial capacity than fraud victims. A small number of fraud victims (8
people) were determined to have limited financial capacity and only mild memory impairment
during the neuropsychologists’ evaluation, whereas none of the financial abuse victims
demonstrated full financial capacity. In terms of the number of individuals with severe levels of
impairment, both groups were similar. Over half the victims in both samples lacked the capacity

75

to manage their finances independently, and over half presented with severe memory
impairments.
There was some evidence that fraud victims were more financially sophisticated in the
past than financial abuse victims, meaning that a lower proportion depended on others for help
with paying bills, managing the estate, and investing money when they were middle aged. Three
fraud victims (10.7%) were described as having excellent financial sophistication/financial
literacy in the past, nine were described as having good financial sophistication (32.1%), and six
were poor/dependent on others (21.4%). Of the fraud victims with poor financial literacy in the
past, one neuropsychologist reported that the victim “cannot write out checks or manage
checkbook; didn't know name of bank; financially naïve; doesn't understand the concept of a
reverse mortgage; lived with parents who managed finances before they died in 1997.” No
information on prior financial management was available for 32% of fraud victims.
In comparison, no financial abuse victims were determined to have “excellent” financial
sophistication in the past, four (16%) were described as having “good” financial sophistication,
seven (28%) were poor/dependent on others, and 13 (52%) had no documentation available on
their past financial abilities. For fraud victims, the average estimated length of time in which the
person lacked financial capacity (for those deemed to have impairments) was three years,
whereas for financial abuse victims the average length was 4.4 years. In sum, while there is
some evidence that fraud victims were more financially sophisticated in the past and less
cognitively impaired in the present, these findings are specific to this sample of cases brought to
the Forensic Center where cases are generally more complex. These findings should be tested in
a larger and more representative sample of cases that only receive APS intervention.

76

Of all the raw neurocognitive data included in the case files, only the MMSE and Money
Management subtest were consistently administered and could be analyzed quantitatively. The
relationship between cognitive functioning and type of exploitation was also tested by regressing
victims’ MMSE scores (and separately Money Management scores) on type of financial
exploitation—fraud versus financial abuse. Age and education were included as control
variables. The dependent variable in the Poisson regression was whether the individual had been
a victim of or financial abuse (=0) or fraud (=1); the independent variable was the
unstandardized test score. It was hypothesized that the regression coefficient would be positive
and significant in both equations; that is, victims of fraud would perform significantly better on
the MMSE and Managing Money subtest than victims of financial abuse.
As shown in Table 4.6, fraud was not a significant predictor of Money Management
(β=0.157, p=0.127), but it was significant for MMSE (β=0.142, p=0.026), such that fraud victims
performed better on the MMSE than financial abuse victims. Age and education were not
significant in the MMSE regression model but were statistically significant in the Money
Management model (results not shown). To help assess the fit of the Poisson regression models,
the estat gof command was applied in Stata 13.1. The goodness-of-fit χ
2
test was statistically
significant indicating that the data do not fit the model well and that additional predictor
variables are needed to improve model fit; however, the sample size is too small to include more
independent variables beyond age and education.
In addition to Poisson regression, Fisher’s exact test was used to determine if there were
significant differences between groups in categorically-coded cognitive data. As shown in Table
4.6, there were no significant differences in memory impairment (p=0.143) or financial capacity
(p=0.589) when these characteristics are analyzed quantitatively.

77

Table 4.6. Differences in cognitive functioning between fraud and financial abuse victims using
Poisson regression and Fisher's Exact test

Financial
Abuse
Victims
Fraud
Victims
Fisher's
Exact
test Poisson regression

Mean
(SD)
/frequency
Mean
(SD)
/frequency p-value Coef.
95%
Confidence
Interval p-value

Poisson Model
Fraud on MMSE 18.8 (5.0) 21.4 (6.9) 0.142 0.017 0.266 0.026*
Fraud on money
management 14.9 (9.8) 20 (10.3) 0.157 -0.045 0.359 0.127
Difference Tests
Memory Impairment 0.143
Not impaired 0.0% 7.7%
Mild 8.3% 19.2%
Moderate 33.3% 11.5%
Severe 58.3% 61.5%
Financial capacity
0.589
Mildly impaired 4.6% 14.3%
Moderately impaired 31.8% 32.1%
Severely Impaired 63.6% 53.6%
Note: The Poisson regression models predicting MMSE and Money Management scores adjust for
victims' age and education.

Hypothesis 2: Fraud victims have better physical health and are less dependent on others than
financial abuse victims

Table 4.7 summarizes the physical and mental health characteristics of victims. Based on
information extracted from the victim’s historical medical records and the neuropsychologists’
report, fraud victims presented with only slightly better overall physical health than financial
abuse victims. Nine fraud victims (32.1%) and ten financial abuse victims (40%) were in very
poor health.

78

Table 4.7. Victim physical and mental health characteristics

Financial
Abuse Victims
(n=25)
Fraud Victims
(n=28)
Count Percent Count Percent
Overall Health
good 0 0% 2 7.1%
fair 9 36% 8 28.6%
poor 6 24% 9 32.1%
very poor 10 40% 9 32.1%
Mobility
No difficulty ambulating 4 16.0% 6 22.2%
Unsteady, slow ambulation, poor balance, no assistive device 6 24.0% 12 44.4%
Uses an assistive device to ambulate (e.g., cane, walker) 12 48.0% 6 22.2%
Uses a wheelchair/bedbound 3 12.0% 3 11.1%
Level of dependency
Fully independent 0 0.0% 1 3.7%
Requires assistance with only 1 or 2 IADLs 3 12.0% 3 11.1%
Requires assistance with some (3-4) IADLs and few or no ADLS 6 24.0% 9 33.3%
Requires assistance with many (5-6) IADLs or some (1-2) ADLS 2 8.0% 5 18.5%
Requires assistance with all IADLs and some (1-2) ADLS 5 20.0% 2 7.4%
Dependent on others for help with all IADLs and most (3-5)
ADLs 5 20.0% 5 18.5%
Fully dependent on others 5 16.0% 2 7.4%
Mood at evaluation
Negative mood--sad, tearful, angry, aggressive, impatient,
irritated 4 16.0% 7 25.0%
Flat/blunted affect--quiet, reserved 2 8.0% 3 10.7%
Positive mood--friendly, cheerful, cooperative, upbeat, engaging 13 52.0% 11 39.3%
Mixed affect 6 24.0% 7 25.0%
Post-exploitation depression 6 11
History of depression 11 13
History of psychosis 5 6
History of substance abuse problems 3 3
History of stroke or transient cerebral ischemia 7 8

As shown in Table 4.8, these differences in overall physical health were statistically
significant using Fisher’s exact test (p=0.369). Both groups had high levels of cardiovascular

79

disease—arteriosclerosis, hypertension, congestive heart failure—and other aging-related
diseases such as cataracts, dementia, diabetes, stroke, and Parkinson’s disease.
Table 4.8. Differences in health and dependency between fraud and financial abuse
victims

Financial
Abuse
Victims
Fraud
Victims
Fisher's
Exact Test

Mean (SD)
/frequency
Mean (SD)
/frequency p-value
Level of dependency
Overall health
Good 0.0% 10.7% 0.369
Fair 36.0% 28.6%
Poor 24.0% 32.1%
Very poor 40.0% 28.6%
Mobility

0.049*
No difficulty/no assistive device 40.0% 66.7%
Uses an assistive device 60.0% 33.3%
ADL/IADL Dependency

Independent/little assistance needed 12.0% 14.8% 0.841
Some IADL & No ADL assistance 24.0% 33.3%

Full IADL & Some ADL assistance 28.0% 25.9%
Fully dependent on others 36.0% 25.9%

Fifteen (60%) financial abuse victims relied on an assistive device to ambulate compared
to just nine (33.3%) fraud victims. These differences in mobility were significant using a one-
way Fisher’s exact test (p=0.049), such that the proportion of fraud victims who could ambulate
independently was significantly greater than the proportion of financial abuse victims.
Qualitatively, fraud victims were less dependent on others for care than financial abuse
victims, although differences in ADL and IADL functioning were not statistically different
(p=0.841). Seventeen fraud victims reported they needed no assistance with activities of daily

80

living yet the evaluator noted that two of these individuals had poor insight about functioning
and personal care needs. The majority of both victim groups relied on others for assistance with
driving, shopping, and finances, but most could independently manage basic activities such as
bathing, toileting, transferring and getting dressed. Five financial abuse victims were completely
dependent on their caregivers for all assistance with ADLs.
Hypothesis 3: Fraud occurs during a “window of opportunity”—the combination of impaired
decision-making and no financial oversight
Qualitatively, findings from this study strongly support hypothesis 3 as demonstrated by
the common narrative structure running through most of the fraud cases. Although the type of
fraud varied between cases, the general story outline is this: An elder with some degree of
physical and/or cognitive impairment is fleeced by a stranger because s/he has sufficient
economic resources to be a suitable target and there is no protective social network to monitor
the elder’s finances and intervene. Overall, findings support the adapted routine activity theory
as a model for fraud against older adults (see Figure 1.1).
The elders victimized by fraud typically lived alone either because they were widowed or
had never been married. Approximately two-thirds of the fraud victims were childless and many
had no relatives living nearby. It may be that elderly people without children lack oversight by a
trusted individual, thus making them more vulnerable to exploitation by strangers (Wilber &
Reynolds, 1996). In addition to social isolation, all victims had impaired decision-making
capacity when they were first approached. Some were recovering after being hospitalized and
others had dementia.
The majority of victims in both groups were identified by the neuropsychologist as being
highly susceptible to undue influence. One report stated that a financial abuse victim was

81

vulnerable due to having “no third party financial advisors, fully dependent on adult children for
assistance with ADL/IADLs, emotional dependency, dementia, and having a trusting and
confidential relationship with the abuser [adult son]”. Another report stated that the victim was
vulnerable because of “isolation…lack of insight into current needs, depression, emotional
dependency on the perpetrator, and poor health leading to functional dependency.”
Differences in mental health and emotional status
Victims presented wide variation in affect during the neurocognitive interview. Moods
ranged from sad, tearful, aggressive and impatient, to flat, quiet, and reserved, to positive,
friendly, cooperative and engaged. One fraud victim was “exceptionally friendly” and invited
the examiner and social worker to join her at functions. This inappropriate level of openness
toward strangers could potentially increase her risk of fraud. Not all victims were friendly to the
neuropsychologists. A financial abuse victim was described as, “Easily distracted and
digressive; did not want to discuss finances with examiner.” Although fraud victims were
generally less friendly and cooperative than financial abuse victims (four fraud victims versus
seven financial abuse victims were described as being in a generally “positive mood”), a quarter
of both groups fluctuated in their affective states throughout the interview and were described as
“emotionally labile.”
Both groups had high rates of depression based on historical medical records and the
neuropsychologists’ evaluations. A history of depression was noted in thirteen fraud victims’
case files and eleven financial abuse victims’ case files. Eleven fraud victims and six financial
abuse victims experienced episodes of depression after they were exploited. It cannot be inferred
that the act of exploitation caused symptoms of depression since many victims experienced other
negative events—e.g., widowhood, stroke—during the same time as exploitation. Only three

82

victims in each group had a recorded history of substance abuse problems; five financial abuse
and six fraud victims had a history of psychosis, and seven financial abuse and eight fraud
victims had a history of ischemia. The percentages of victims who experienced these
events/mental illnesses out of the total sample are not presented in Table 4.7 because it was not
possible to differentiate those who never had the condition/illness from those who were simply
missing the information from their case files.
Differences in social and environmental characteristics
Table 4.9 displays the distribution in living environment characteristics, physical
appearance/hygiene, and isolation for both groups. The neuropsychologists documented a wide
range of living conditions, yet qualitatively there were no substantial differences between
financial abuse and fraud victims. Some victims lived in very clean, well-organized homes and
apartments and were neatly groomed during the interview, and others lived in hoarding
conditions where there were noticeable health and environmental concerns, such as lack of
running water and fall hazards. For example, one fraud victim’s apartment was described as
“dilapidated—urine soaked chairs, used disposable diapers, vermin, odors, trash everywhere,
non-functional plumbing; no furniture, limited food (this is why she was taken to the emergency
room); pictures of the suspected abuser and some letters on the floor in the living room; piles of
financial papers on the floor.”
Victims in unsafe living environments often presented with poor hygiene: inappropriate
dress, body odor, and soiled clothing. One financial abuse victim answered the door “wearing a
robe over her clothes; teeth needed care…Doesn’t take care of medical needs; severe edema.” A
different fraud victim had signs of financial victimization throughout his apartment: “bait letters
and money packs throughout the home…The examiner noticed at least six inches of opened mail

83

on kitchen table all indicating that elder was the recipient of a lottery wining, and requesting that
he provide a sum of money to collect the winnings.”
Table 4.9. Care environment and grooming characteristics

Financial
Abuse Victims
(n=25)
Fraud
Victims (n=28)
Count Percent Count Percent
Care Environment
Non-remarkable (e.g., clean/neat AND/OR suitable residential
care) 9 36.0% 10 35.7%
Poor living environment (e.g., dirty, hoarded, poor repair,
hazardous) 9 36.0% 12 42.9%
Missing 6 24.0% 6 21.4%
Physical Appearance
Acceptable hygiene; appropriate dress and grooming 13 52.0% 15 53.6%
Poor grooming; inappropriate dress; body odor; soiled clothing 8 32.0% 5 17.9%
Not documented 4 16.0% 8 28.6%
Isolation prior to/during exploitation
Not isolated; socially active; connected to others 2 8.0% 1 3.6%
Somewhat isolated; low social engagement; few social partners 1 4.0% 3 10.7%
Very lonely or isolated; minimal or no social contact 5 20.0% 7 25.0%
Isolated from others by the perpetrator 14 56.0% 13 46.4%
Unknown 3 12.0% 4 14.3%

Approximately 70% of fraud victims lived alone at the time of exploitation compared to
44% of financial abuse victims, eight co-resided with the abuser. Two financial abuse victims
still lived with their abusers when they were evaluated. A few elders, approximately 20%, had
lived in unsafe conditions at the time of exploitation or lived with the abuser but then were
moved to residential care settings once family members recognized the need for higher levels of
care. These victims were interviewed in their new care setting, not the setting where the
financial exploitation took place, so no conclusions can be made about the original context of
mistreatment.

84

Table 4.10. Differences in social isolation between fraud and financial abuse victims

Financial
Abuse
Victims
Fraud
Victims
Fisher's Exact
Test/Poisson
Regression

Mean (SD)
/frequency
Mean (SD)
/frequency p-value
Social Isolation
Level of isolation 0.655
Not isolated 13.64% 16.67%

Isolated 86.36% 83.33%
Marital status

0.186
Married/Partnered 16.7% 28.6%
Divorced 8.3% 21.2%
Widowed 70.8% 39.3%
Never Married 4.2% 10.7%
Mean number of children 1.96 (2.28) 1.10 (1.71) 0.013*

Fisher’s Exact test showed that there were no significant differences in level of social
isolation between victim groups (p=0.665), nor were there differences in marital status (p=0.186)
(See Table 4.8). Fourteen (56%) financial abuse victims and 13 (46.4%) fraud victims were
isolated from others by the alleged perpetrator. Fraud victims were significantly less likely to
have children than financial abuse victims (β=-0.571, p=0.013). Eighteen out of the 28 fraud
victims were childless, whereas only 9 of the 25 financial abuse victims were childless. Among
those who did have children, the mean number of children was three for both groups.
DISCUSSION
Research shows that money management is one of the first areas of cognitive functioning
to become impaired with age (Marson et al., 2000). Some victims in the Forensic Center sample

85

maintained limited knowledge of financial concepts and skills in money management: They
demonstrated their ability to count out change and could describe the purpose of health
insurance, yet most could not describe how to prevent being cheated out of money and could not
write out checks properly. A common pattern that emerged from this analysis was that fraud and
financial abuse victims had poor insight about their abilities and limitations. The vast majority
were unaware that they had been victimized, and most of those who did acknowledge
victimization believed they would get their money back eventually. This is an important finding
because it indicates that many, if not most, older victims do not recognize their need for help
managing money and do not report being exploited. This lack of insight poses an additional risk
because it means that perpetrators can continue to exploit older adults without fear that they will
be caught.
The results of the quantitative and qualitative tests of Hypothesis 1 indicate that there
were minimal differences in cognitive functioning between fraud and financial abuse victims at
the time of the neurocognitive evaluation. Although fraud victims performed significantly better
on the MMSE compared to financial abuse victims, other quantitative differences in memory,
money management, and overall financial capacity were unremarkable. Lack of significance
may be due to the small sample size, but another explanation is floor effects: overall cognitive
performance on most tests was very low in both groups.
There was also only limited support for Hypothesis 2: Victims of fraud presented with
somewhat lower levels of dependency than victims of financial abuse, and these differences were
largest for mobility. Fraud victims had significantly better mobility than financial abuse victims,
such that a smaller proportion relied on assistive devices to ambulate. These difference could be
clinically relevant because better mobility means that fraud victims can provide for more of their

86

own needs. Despite small qualitative differences in health and cognitive functioning between
groups, overall, both samples showed considerable physical and cognitive impairment. More
than 50% of both groups needed assistance with ADLs, and almost all victims required at least
some help with IADLs. Common health problems included cardiovascular disease, dementia,
metabolic syndrome, vision impairment, anemia, osteoporosis, and many other debilitating
illnesses that increase dependency on others.
Despite many similarities between fraud and financial abuse cases, one major difference
was that fraud victims lost significantly more than financial abuse victims—an average of nearly
$619,000 compared to $200,000 per person. This difference may be explained by the fact that
many financial abuse perpetrators are opportunistic; they take what they can from vulnerable
older adults within their social networks who are most often relatives, neighbors or friends.
Predatory scam artists, on the other hand, actively pursue targets with means rather than relying
on susceptible elders already within their social networks. The sum total for all 54 victims who
had financial data available is alarming: Victims lost $19 million in savings and investments and
17 properties. This high figure suggests that the current $2.6 billion dollars per year calculation
provided by MetLife in 2009 underestimates the total losses suffered by elder financial abuse
victims in the U.S.
A narrative of fraud
A common condition among fraud victims was social isolation. Some became less social
after losing a spouse or becoming disabled, while others did not socialize by choice. Over half
of the fraud sample was isolated intentionally by the suspected abuser. Prior research studies
have found that social isolation increases vulnerability to elder financial exploitation (Lee &
Soberon-Ferrer, 1997; Choi & Mayer, 2000; Tueth, 2000). First, there are fewer eyes to guard

87

the elder’s estate and less people to provide caregiving assistance. Second, without other
meaningful social interactions, the victim may become increasingly dependent on the financial
predator to fulfill their social and emotional needs.
By secluding the elder and restricting access to friends and family, scam artists can both
hide the exploitation and use the power differential to force the victim to comply (Bernatz, 2008;
Quinn, 2000). For example, a perpetrator in one fraud case “isolated the victim from family and
previous employees,” and “withheld food to coerce victim to sell strip mall and fire long-term
attorney and accountant.” Another suspected abuser changed the victim’s locks and unplugged
her telephones to cut her off from her son. Two other perpetrators moved the elder into a long-
term care setting. By the time relatives, neighbors, healthcare providers, law enforcement
officials or financial institutions noticed and reported the situation, victims had already lost
hundreds of thousands of dollars.
Another common thread running through the fraud cases was how perpetrators met their
victims and how they gained their trust. Perpetrators positioned themselves in locations often
frequented by older adults such as the supermarket, church, hospital, and bank. One fraud victim
met the perpetrator outside a Von’s parking lot, another met her at a 99₵ store, and another in an
assisted living facility. Perpetrators also worked in service jobs as personal care aides,
housekeepers, financial advisors, and home maintenance workers. In these occupations they
could use the guise as legitimate professional to establish trust. For example, one perpetrator
pretended to be a patient advocate at a hospital and specifically targeted an older person
recovering from emergency prostate surgery. She introduced the heavily medicated man to a
string of other perpetrators who convinced him to grant them Powers of Attorney and transfer
over the title of his home so they could “fix it up” to be ready for him when he was discharged.

88

In four cases the perpetrators were neighbors and tenants who noticed an opportunity to exploit,
but most perpetrators in the Forensic Center cases were not opportunists who stumbled into the
situation and took advantage of it; they were actively seeking vulnerable targets. Included in the
sample were five repeat criminals who had a rap sheet of going after older adults.
Some perpetrators in the sample offered more than just commercial and financial
services. Rather, they charmed their way into romantic partnerships with the victims in order to
gain control. One repeat female offender befriended an older gentleman living in an assisted
living facility. She pretended to be a friend of another resident. She and her victim began a
secret romantic and sexual relationship that lasted between 6 to 10 visits, according to the
investigator’s report. During that time she convinced him to sign his name on loan documents to
purchase her a used Mercedes with the promise that she would pay him back. Repayment never
occurred and the car was gone by the time the police investigated this case. In another scenario,
an older male was the target in three different sweetheart scams, and in another, a cognitively
impaired man was seduced by the woman hired to provide care to his developmentally delayed
brother. These cases illustrate how promises of romance and companionship are also used to
persuade victims.
To get a foot in the door, a common tactic fraud perpetrators used was providing small
favors for their victims, such as taking them grocery shopping or doing light housework. Based
on the norm of reciprocity, receiving small acts of goodwill can pressure people to comply with
larger requests later on (Cialdini, 2012). For example, in return for some help with her laundry,
one victim allowed the perpetrator to live in her home rent-free. In another case, a wealthy elder
agreed to invest money in a marijuana growing operation after the perpetrator helped him with

89

some home repairs. In all cases, it appears the victims trusted the perpetrators and assumed they
were acting in accordance with their interests.
Not all cases of fraud aligned with this narrative pattern. For example, two victims in the
sample had never met the perpetrators in person. In these cases, scam artists claimed the victims
had won foreign lotteries and needed to pay the international taxes before receiving their
winnings. All correspondence took place over the phone or by mail. Although these
perpetrators didn’t target elders by going to the places they often frequent, it is possible that they
actively solicited elders over the phone because of the perception that older people are easier
targets.
Another deviation from the common pattern is that not all victims were socially isolated
from friends and family. A few received frequent support from their neighbors and adult
children, and one lived around the corner from her son. Another victim was still very active in
church functions and another went on regular outings with her sorority sisters from college.
These observations conflict with the adapted model presented in Figure 1.1 which shows that in
order for fraud to occur, a suitable target and motivated offender must meet in a “window of
opportunity” which is defined by the absence of capable (trustworthy) guardians. Although these
elders had regular social contact, it is possible that their open and friendly personalities increased
their exposure to unscrupulous people. In support of this hypothesis, Wyk and Mason (2001)
found that socialization significantly increased the risk of exploitation in random sample of 400
adults of all ages.

90

LIMITATIONS
Forensic Center cases represent some of the most complex and severe cases of abuse,
neglect, and exploitation that occur in Los Angeles County, and thus are not representative of the
majority of reports to APS. However, because exploitation is so well documented in the
Forensic Center case files, this sample provides insights that would not be available in other data
sources such as the HRS. Broader applicability is important for guiding fraud detection and
intervention policy and practice, so future qualitative studies should include cases reported to
APS and law enforcement that never receive a Forensic Center intervention.
Another limitation of this study was the lack of complete neurocognitive data in the files
and the variability in which tests were administered. While all files contained comprehensive
reports on the victims’ functioning from the neuropsychologist, most were missing the raw
scores on which these reports are based, thereby limiting quantitative analysis. There are many
unanswered questions about why certain neurocognitive tests were never administered: Was it
because the evaluator believed the victim was too impaired to attempt them, not impaired enough
for the test to provide meaningful information, or were there time constraints? There was a
similar problem with living environment observations in the case files: Was the home not
remarkable enough to document or did the neuropsychologist neglect to take notes? These
problems can be eliminated in future studies by standardizing interviews and improving
documentation protocols.
CONCLUSION
Based on this mixed-methods analysis of Forensic Center case files, there is only limited
evidence that victims of fraud present with better cognitive functioning and functional status

91

compared to victims of financial abuse by friends and family. Other than significantly higher
MMSE scores, better mobility, and more childlessness among fraud victims, the two groups
share many of the same physiological, environmental, demographic, and psychosocial
characteristics. The vast majority of victims presented with mental and physical health
conditions including diabetes, hypertension, stroke, vision problems, depression, and other
mental illnesses and age-related diseases. They also had severe deficits in memory and
executive functioning, although very few victims had insight about their limitations and needs.
Given the many similarities across cases, current elder financial exploitation and undue influence
frameworks (Blum, 2003; Bernatz, 2008) are sufficient in modeling the major concepts,
mechanisms, and risk factors associated with fraud against older adults.
One proposed explanation for the similarity between fraud and financial abuse victims is
that exploitation has less to do with characteristics and risk factors associated with the victim,
and more to do with the people surrounding the victim (or lack thereof). Take for example an
impaired widow who lives with her unemployed drug-abusing son. This widow faces a high risk
of financial abuse. Now imagine that same widow with no children and no close friends or
relatives. She is more vulnerable to fraud by a stranger. In other words, it is the structure of the
victim’s social network that determines whether s/he is more likely to be victimized by fraud or
financial abuse. As the hypothetical scenario illustrates, the presence of adult children decreases
risk of fraud, yet increases the risk of financial abuse because adult children are the most
common perpetrators of financial exploitation (Acierno et al., 2009).
This in-depth analysis illustrates that victims of fraud and financial abuse share many of
the same vulnerability characteristics even though the crimes and perpetrators differ. These
results should be validated using quantitative measures and a more geographically diverse

92

sample. Future qualitative studies should follow victims over time with follow-up evaluations
conducted every year until death. Longitudinal research could inform researchers and the
practice community on case outcomes: Do victims receive justice and/or compensation? What
are the health and disability outcomes following exploitation? Are victims better protected from
future scams and exploitation after their cases are investigated by APS or law enforcement?
What is the value-added of the Forensic Center intervention? Using the Forensic Center’s rich
case narratives, researchers can begin to answer these questions and move closer to developing
and testing effective prevention programs for vulnerable adults.

93

CHAPTER V: CONCLUSION
“Fraud is, at its core, successful marketing for illegitimate purposes. To understand fraud is to
understand deception—its persuasive strategies and practical methods.”
–Deevy, Lusich, & Beals (2013)

According to findings in Chapters II and III, the common stereotype that fraud victims
are lonely older widows with poor cognitive functioning paints an incomplete picture of who is
actually being defrauded. The latent class analysis in Chapter III showed that most fraud victims
over age 50 do not conform to these assumptions, but rather fall into different socioeconomic and
demographic groups. Of the two classes identified, older widowed females, whose
characteristics conform to our assumptions of the “typical” fraud victim, represent a smaller
proportion of victims over age 50. However, an alternative explanation is that older age is
associated with higher rates of fraud, yet adults in older age groups fail to recognize and report
their victimization in the HRS. Although it was not possible to determine what type of fraud was
responsible for each groups’ victimization in the HRS, it is hypothesized that the middle-aged
married adults with high-SES (Class 1) were victims of investment fraud, and older widowed
females (Class 2) were victims of bogus lottery scams. These typologies align with findings
from prior research that used law enforcement and fraud complaint data to identify victim
profiles.
Chapter II identified risk factors for fraud. Non-modifiable characteristics related to
vulnerability were age, gender and SES. Although these characteristics are not amenable to
intervention, the results show which groups should receive fraud prevention and intervention
services, and provide information on which interventions may be the most appropriate. Given
the heterogeneity of fraud victims in the older adult population in the United States, practitioners

94

and policy makers should tailor educational materials to each victim group’s specific consumer
behaviors, psychosocial characteristics, and financial planning needs. And, since age is
negatively associated with fraud, actions to safeguard people’s assets should begin in middle-age
when consumer activity is high and planning for retirement is a priority.
For older populations, prevention strategies should target modifiable risk factors like
unmet care and emotional support needs. In addition to cognitive impairment, nearly all the
Forensic Center victims in Chapter IV had health problems and lacked the necessary social
support to help them meet these needs. If scam artists look for cognitive changes and
social/emotional losses when selecting a suitable target, adult protection workers, elder justice
advocates, and family members must be one step ahead in detecting these cues that the person is
at risk. It is possible that fraud can be intercepted by mobilizing support for elders immediately
after negative events, such as the loss of a spouse, the onset of a debilitating physical condition,
job loss, or at the first signs of declines in financial decision-making. By reinforcing elders’
informal and formal support networks with family members, friends, and even health care
providers, perpetrators are prevented from filling those gaps with their own promises of wealth
and/or companionship. For those who are already victims, support groups may help overcome
the shame and stigma attached to falling for a scam by normalizing fraud as something that
happens to people across age and socio-demographic group. Victim support groups would also
provide safe opportunities for social interaction and emotional support.
There is not sufficient evidence to show that victims in the Forensic Center case files
reflect the two socioeconomic/demographic profiles of fraud victims in the HRS. First, in
general, Forensic Center clients are far more physically and cognitively impaired than HRS
respondents, and none of the Forensic Center clients self-reported exploitation. Other major

95

differences between the samples is that more Forensic Center clients were widowed (40%
compared to 14.5%), and they were over 20 years older on average—83.4 years for Forensic
Center clients compared to 61.7 years for HRS fraud respondents. These differences suggest one
of two things. The first explanation is that HRS respondents are simply younger versions of the
Forensic Center clients and will eventually reflect their characteristics when they reach eighty
and ninety years old. At these older ages the HRS victims will presumably have more disease
burden, more physical and cognitive impairment, and more of them will be widows.
The alternative explanation is that Forensic Center clients are an entirely different group
of people than victims in the HRS. Perhaps HRS respondents do not enter the adult protective
service system because they have other resources, such as friends and family members, to help
them resolve financial exploitation on their own without relying on social services. The Forensic
Center victims, on the other hand, enter the system due to their very high levels impairment and
lack of capable and trustworthy guardians. Perhaps this extremely vulnerable population is
underrepresented in survey research like the HRS because they have no proxy decision-
makers/respondents for surveys.
Regardless of which explanation is correct, the results of this analysis reveal that
intervention for fraud begins far too late in life. Based on what was found in this analysis,
Americans are being victimized as young as ages 50 and 60, or perhaps at even younger ages
than were included in the HRS sample. They may go on to experience more fraud victimization
in later life (although this cannot be verified with the current data). Victims in the Forensic
Center sample had already lost an estimated $15 million dollars and multiple properties by the
time APS, the police, or other friends and family members intervened. Many lost their entire life
savings and presumably must now turn to public services for health and housing support.

96

Chapter IV findings reinforce how costly fraud is to victims and to the public agencies charged
with investigating and responding to these crimes.
How to protect vulnerable adults
Intervention for fraud must address each level of the model presented in Figure 1.1. For
example, one factor that raises fraud risk is the absence of capable guardians to protect a
susceptible target. Thus, a promising solution for preventing fraud is ensuring adequate
safeguards are in place to protect an elder’s personal identifying information, real estate,
personal property, and financial account information. Financial, health care, and social services
professionals should initiate conversations with their clients/patients about appointing a
trustworthy friend, relative, or hired professional to help them with their finances in the event
they can no longer manage independently. These discussions should begin in mid-life, as
findings from Chapter II showed that adults closer to middle age were more vulnerable to fraud
than the oldest ages. Furthermore, financial advisors and estate planners need training in how to
recognize signs that a client may require higher levels of oversight in financial decision-making.
According to the Federal Bureau of Investigation (FBI, 2014), smart practices that
prevent fraud include setting up caller ID on landlines, registering phone numbers on the “do not
call list”, discussing possible investment opportunities with trusted friends, relatives, and
professionals, and instituting purchasing rules to avoid bad decisions. Purchasing rules include
never responding to unsolicited advertisements, never paying for things in advance, researching
the legitimacy of unfamiliar companies/products, having an attorney review business contracts
before any agreements are signed, and never disclosing personal information over the phone.
Another strategy is registering for online banking so that scam artists are blocked from creating a
new online account that is linked to the target’s funds.

97

Titus (2001) found that single incidents of fraud increase the risk of future victimization;
and, as shown in Chapter IV, perpetrators took money from their victims gradually rather than
taking everything in a single transaction. One cause of re-victimization is that susceptible
victims’ contact information is added to “sucker lists” (also called “mooch lists”) and then sold
to other scam artists (FBI, 2014). Implementing safeguards after the first incident of fraud, such
as appointing a power of attorney, can help ensure that elders do not fall into repeated cycles of
victimization.
Basic training in financial literacy is not sufficient to protect individuals from fraud. In
fact, one study found that investment fraud victims performed better than non-victims on
traditional financial literacy tests (FINRA, 2006). In a more recent study by FINRA (2013), 40%
of respondents rated an investment promising a “110% annual return” as highly appealing, and
another 43% were attracted to offers that were “100% guaranteed,” despite that any investment
promising returns over 100% is highly improbable. Because financial literacy is not a good
indicator of vulnerability and because many Americans are easily swayed by bogus information,
prevention efforts must incorporate specific training on common types of scams and persuasion
tactics (AARP, 2003; FINRA, 2006).
Americans also need information on how to report fraud. Studies have shown that those
who are knowledgeable about reporting are less likely to be victimized and more likely to report
if they are (Copes, Kerley, Mason, & Van Wyk, 2001; Titus, 2001). However, the FTC
(Anderson, 2004) found that only 55.4% of victims between ages 55-64 reported fraud to
complaint agencies compared to 75% of those under age 35; and AARP (Pak & Shadel, 2011)
found that only 25% of victims 55 and older reported compared to 44% of victims under age 55.
This paradox suggests that prevalence studies relying on self-report, such as the HRS and FTC

98

surveys (Anderson, 2013, 2007, 2004), may be skewed. In other words, older people may in fact
experience higher rates of fraud than middle and younger aged adults, yet fail to disclose fraud in
self-reported surveys. This is a major concern that has implications affecting the present
research. The divide between experiencing fraud and reporting fraud needs to be addressed in
future research.
The United States Senate Special Committee on Aging has a reporting hotline staffed by
experienced fraud investigators that is intended to assist victims with their complaints and refer
them to the appropriate authorities (U.S. Senate Special Committee on Aging, 2013). This
single, universal hotline may help link incidents of fraud nationwide to show where investigative
and supportive resources should be allocated. Increased reporting would also help researchers
identify what groups experience the highest rates of fraud and whether age plays a role in the
frequency of reporting. Furthermore, higher rates of reporting would provide better estimates on
the cost of fraud. From what could be discerned from the case files analyzed in Chapter IV, none
of the victims reported fraud, meaning that that nearly $15 million in stolen assets was never
included in estimates of the total cost of fraud collected by federal fraud complaint agencies.
Banks, credit unions, and other financial institutions are necessary participants in the
fight against fraud. In most states, financial institutions are mandated to report elder abuse and
exploitation. Too often banks facilitate criminal activity by authorizing unusual money transfers
from accounts held by clients with diminished capacity or by inadvertently giving criminals
access to these accounts. Bank staff should undergo mandated training on the warning signs of
fraud and undue influence, and learn where to report their suspicions to prevent further
fraudulent transactions (Nerenberg, 2006). The Government Accountability Office (2012)
recommends that financial institutions receive information on when they are legally permitted to

99

release a client’s financial records to law enforcement and engage in information sharing.
Partnerships between these stakeholders can improve efforts to track, apprehend, and prosecute
scam artists.
The future of fraud policy
Fraud may be an intractable problem that is embedded within the cultural and economic
fabric of our society. First, fraud is an exceptionally adaptable enterprise. Methods of
exploitation evolve with advances in technology and communication, and persuasion tactics tend
to keep pace with popular trends in advertising and mass marketing. Modes of exploitation
began with door-to-door sales pitches, moved to telemarketing, and now rely primarily on virtual
exchange via the internet. For those aiming to protect Americans against fraud, the next frontier
is predicting what mechanisms will be used to commit fraudulent acts in the future.
One of the most troubling characteristics of fraud is how scam artists adopt the same
sales and marketing tactics used by legitimate businesses to stay competitive and concealed in
the marketplace, even taking advantage of current events to make a profit. For example, when
President Obama signed the Affordable Care act in 2010, scam artists were at the ready, calling
households nationwide to inform older adults that changes would be made to their Medicare
insurance if they did not follow the directions provided. These bogus health insurance scams
resulted in nearly 83,000 complaints to fraud agencies (Gold, 2013).
Even more alarming is how the line between legitimate and fraudulent products, services,
and information is intentionally blurred by scam artists. These vague boundaries make it hard
for people of all ages to distinguish between information that is true from information that is
false. The proliferation of numerous media sources, each with different political agendas,
provides a platform for many illegitimate ideas to take hold. How are Americans, particularly

100

those with age-related cognitive impairments, expected to rationally assess all of this incoming
information and separate fact from fiction?
In addition to the hazy distinction between legitimate and fraudulent, the growth of social
media and information technology has outpaced policy makers’ ability to regulate these
industries. Americans have very few rights and protections regulating how private corporations
can use and share personal information collected online. The lack of oversight and slow
implementation of protections provide opportunities for scam artists to learn all about our
purchasing behaviors, the places we visit, who we communicate with, and our daily habits and
interests. With so much of our lives imprinted in the virtual world, it’s possible that marketers
and scam artists alike know more about us than even our closest friends and family members.
Consumers should not maintain the sole responsibility for detecting and protecting
themselves against scams. Defending the next generation of older adults requires a top-down
prevention approach from government agencies, including the passage of new laws that protect
our virtual privacy. Increased oversight of the financial services and insurance industries and
enhanced collaboration between law enforcement, financial institutions, adult protective services,
and federal investigatory agencies is needed (AARP, 2003; GAO, 2012). Partnerships can
improve efforts to track, apprehend, and prosecute scam artists.
The future of fraud research
In addition to studying those who are victims, perhaps a more promising avenue for
future research is to study elders who successfully evade fraud attempts. What qualities and
behaviors make them immune to persuasion? Do they have systems and rules in place for
resisting bogus offers? Or, are they never solicited in the first place? Findings on how non-

101

victims defend themselves can inform researchers and practitioners on developing materials and
trainings for those who are at-risk.
The studies presented in this dissertation focus primarily on older victims of fraud with
very little attention given to perpetrators—who they are, the tactics they use, and the
rationalizations they provide for engaging in fraud. Research on elder abuse indicates that
perpetrator characteristics are more important than victim characteristics in predicting abuse
outcomes (Pillemer, 1985; Reis & Nahamish, 1998). Given how important perpetrators are in
elder abuse, more research is needed to explore this side of the equation for fraud, particularly
how scam artists seek out victims and earn their confidence. Other studies are needed to survey
the range of perpetrators—from scam artists who work independently to large criminal
organizations that operate under the guise of a legitimate business. The scams described in
Chapter IV involved distant telemarketers operating out of Jamaica, predatory females promising
romance, bogus real estate investors, and scheming caregivers. Each of these offenders planned
and organized their schemes differently. Through a better understanding of how perpetrators
think, strategize, and organize their actions, we can mount more effective fraud prevention and
detection policies.

102

REFERENCES
AARP. (1996). Telemarketing fraud victimization of older Americans: an AARP survey.
Washington, DC: AARP.
AARP. (2003). Off the hook: Reducing participation in telemarketing fraud. Retrieved from
assets.aarp.org/rgcenter/consume/d17812_fraud.pdf
Acierno, R., Hernandez, M. A., Amstadter, A. B., Resnick, H. S., Steve, K., Muzzy, W., &
Kilpatrick, D. G. (2010). Prevalence and correlates of emotional, physical, sexual, and
financial abuse and potential neglect in the United States: The national elder mistreatment
study. American Journal of Public Health, 100, 292–297.
Agarwal, S., Driscoll, J. C., Gabaix, X., & Laibson, D. (2009). The age of reason: Financial
decisions over the life cycle and implications for regulation. Brookings Papers on
Economic Activity, 2, 51-117.
Akaike, H. (1987). Factor analysis and AIC. Psychometrika, 52, 317–332.
Alves, L. M., & Wilson, S. R. (2008). The effects of loneliness on telemarketing fraud
vulnerability among older adults. Journal of elder abuse & neglect, 20(1), 63-85.
Anderson, K. B. (2004). Consumer fraud in the United States: An FTC survey. The Federal
Trade Commission. Retrieved from http://www.ftc.gov/reports/consumer-fraud-us-ftc-
survey-august-2004

103

Anderson, K. B. (2007). Consumer fraud in the United States: The second FTC survey. The
Federal Trade Commission. Retrieved from http://www.ftc.gov/reports/consumer-fraud-
united-states-second-federal-trade-commission-survey
Anderson, K. B. (2013). Consumer fraud in the United States, 2011: The third FTC survey. The
Federal Trade Commission. Retrieved from
http://www.ftc.gov/os/2013/04/130419fraudsurvey.pdf
Asp, E., Manzel, K., Koestner, B., Cole, C. A., Denburg, N. L., & Tranel, D. (2012). A
neuropsychological test of belief and doubt: Damage to ventromedial prefrontal cortex
increases credulity for misleading advertising. Frontiers in Neuroscience, 6(100), 2-9.
Attanasio, O. P., Banks, J., Meghir, C. & Weber, G. (1999). Humps and bumps in lifetime
consumption. Journal of Business & Economic Statistics, 17, 22-35.
Baumeister, R. F., DeWall, C. N., Ciarocco, N. J., & Twenge, J. M. (2005). Social Exclusion
Impairs Self-Regulation. Journal of Personality and Social Psychology, 88(4), 589–604.
Baumeister, R. F., Twenge, J. M, & Nuss, C. K. (2002). Effects of social exclusion on cognitive
processes: Anticipated aloneness reduces intelligent thought. Journal of Personality and
Social Psychology, 83(4), 817–27.
Bernatz, S. (2008). Undue influence. In Assessment of older adults with diminished capacity: A
handbook for psychologists. American Bar Association Commission on Law and Aging &
American Psychological Association (pp 114-117).

104

Blum, B. (2003). The “IDEAL” Protocol for Undue Influence – An Introduction, NCPJ Life &
Times, 2(7), 1-4.
Bonnie, R., & Wallace, R. (2003). Elder Mistreatment: Abuse, Neglect, and Exploitation in an
Aging America. National Research Council (U.S.). Washington D.C.: National Academies
Press.
Bowen, G. A. (2008). Naturalistic inquiry and the saturation concept: a research note. Qualitative
Research, 8(1), 137-152.
Boyle, P. A., Yu, L., Wilson, R. S., Gamble, K., & Buchman, A. S. (2012). Poor decision
making is a consequence of cognitive decline among older persons without Alzheimer’s
disease or mild cognitive impairment. PLoS ONE, 7(8), e43647.
Brandt, J., Spencer, M., & Folstein, M. (1988). The telephone interview for cognitive status.
Neuropsychiatry, Neuropsychology, & Behavioral Neurology, 1, 111–117.
Butrica, B. A., Smith, K. E., & Iams, H. M. (2012). This is not your parents’ retirement:
Comparing retirement income across generations. Social Security Bulletin, 72(1), 37-58.
Button, M., Gee, J., Lewis, C., & Tapley, J. (2010). The human cost of fraud: A vox populi.
Centre for Counter Fraud Studies & MacIntyre Hudson. Retrieved from
http://www.port.ac.uk/departments/academic/icjs/centreforcounterfraudstudies/documents/
filetodownload,116542,en.pdf

105

Cacioppo, J. T., Hughes, M. E., Waite, L. J., Hawkley, L. C., & Thisted, R. A. (2006).
Loneliness as a specific risk factor for depressive symptoms: cross-sectional and
longitudinal analyses. Psychology and aging, 21(1), 140-151.
Carstensen, L. L. (1992). Social and emotional patterns in adulthood: Support for socioemotional
selectivity theory. Psychology and Aging, 7(3), 331–338.
Carstensen, L. L., & Mikels, J. A. (2005). At the intersection of emotion and cognition aging and
the positivity effect. Current Directions in Psychological Science, 14(3), 117-121.
Castle, E., Eisenberger, N. I., Seeman, T. E., Moons, W. G., Boggero, I. A., Grinblatt, M. S., &
Taylor, S. E. (2012). Neural and behavioral bases of age differences in perceptions of trust.
Proceedings of the National Academy of Sciences, 109(51), 20848-20852.
Charles, S. T., & Carstensen, L. L. (2008). Unpleasant situations elicit different emotional
responses in younger and older adults. Psychology and Aging, 23(3), 495-504.
Choi, N. G., & Mayer, J. (2000). Elder abuse, neglect, and exploitation: Risk factors and
prevention strategies. Journal of gerontological social work, 33(2), 5-25.
Chou, K. L., Lee, T., & Ho, A. H. (2007). Does mood state change risk taking tendency in older
adults?. Psychology and aging, 22(2), 310-318.
Cialdini R. B. (2007). Influence: The psychology of persuasion. New York: HarperCollins.
Cohen, C. A. (2006). Consumer fraud and the elderly. Journal of Gerontological Social Work
46(3–4): 137–144.

106

Cohen, L. E., & Felson, M. (1979). Social change and crime rate trends: A routine activity
approach. American Sociological Review, 52, 170–183.
Coleman, J. W. (1987). Toward an integrated theory of white-collar crime. American Journal of
Sociology, 93, 406-439.
Connolly, M. T. (2010). Where elder abuse and the justice system collide: Police power, parens
patriae, and 12 recommendations. Journal of elder abuse & neglect, 22(1-2), 37-93.
Conrad, K. J., Iris, M., Ridings, J. W., Langley, K., & Wilber, K. H. (2010). Self-report measure
of financial exploitation of older adults. The Gerontologist, 50(6), 758-773.
Copes, H., Kerley, K. R., Mason, K. A., & Van Wyk, J. (2001). Reporting behavior of fraud
victims and Black's theory of law: An empirical assessment. Justice Quarterly, 18(2), 343-
363.
Cornwell, E. Y., & Waite, L. J., (2009). Social disconnectedness, perceived isolation, and health
among older adults. Journal of Health and Social Behavior, 50(1), 31–48.
Craik, F. I. M. (1986). A functional account of age differences in memory. In F. Klix & H.
Hangendorf (Eds.), Human memory and cognitive capabilities: Mechanisms and
performances (pp. 409-422). Amsterdam: Elsevier-North-Holland.
Craik, F. I. M., & Jennings, J. M. (1992). Human memory. In F. I. M. Craik and T. A. Salthouse
(eds.), The Handbook of Aging and Cognition. (pp. 51–110). Hillsdale, NJ: Lawrence
Erlbaum Associates.

107

Creswell, J. W., & Clark, V. L. P. (2007). Designing and conducting mixed methods research.
Australian and New Zealand Journal of Public Health, 31(4), 388-389.
Deem, D. (2000). Notes from the field: Observations in working with the forgotten victims of
personal financial crimes. Journal of Elder Abuse and Neglect, 12(2), 33-48.
Deevy, M., & Beals, M. (2013). The scope of the problem: An overview of fraud prevalence
measurement. Financial Fraud Research Center: Stanford Center on Longevity. Retrieved
from http://fraudresearchcenter.org/wp-content/uploads/2013/11/Scope-of-the-Problem-
FINAL_corrected2.pdf
Deevy, M. Lucich, S. & Beals, M. (2012). Scams, schemes and swindles: A review of consumer
financial fraud research. Financial Fraud Research Center: Stanford Center on Longevity.
Retrieved from http://longevity3.stanford.edu/ blog/2012/11/19/scams-schemes-and-
swindles-a-review-of-consumer-financial-fraud-research/
Delazer, M., Domahs, F., Bartha, L., Brenneis, C., Lochy, A., Trieb, T., & Benke, T. (2003).
Learning complex arithmetic—an fMRI study. Cognitive Brain Research, 18(1), 76-88.
DeLiema, M., Yon, Y., & Wilber, K. H. (2014). Tricks of the trade: Motivating sales agents to
con older adults. The Gerontologist, gnu039.
Denburg, N. L., Cole, C. A., Hernandez, M., Yamada, T. H., Tranel, D., Bechara, A., & Wallace,
R. B. (2007). The orbitofrontal cortex, real-world decision making, and normal aging.”
Annals of the New York Academy of Sciences, 1121(1), 480–498.
Doerner, W. G., & Lab, S. P. (2008). Victimology. Newark, NJ: Lexis Nexis Matthew Bender.

108

Dong, X., & Simon, M. A. (2008). Is greater social support a protective factor against elder
mistreatment?. Gerontology, 54(6), 381-388.
Dong, X., & Simon, M. A. (2013). Association between elder abuse and use of ED: Findings
from the Chicago health and aging project. The American Journal of Emergency Medicine,
31(4), 693–698.
Dowd, J. J. (1975). Aging as exchange: A preface to theory. Journal of Gerontology, 30(5), 584–
594.
Duclos, R., Wan, E. W., & Jiang, Y. (2013). Show me the honey! Effects of social exclusion on
financial risk-taking. Journal of Consumer Research, 40(1), 122–135.
Federal Bureau of Investigation (2014). Fraud Target: Senior Citizens. [internet]. Retrieved from
http://www.fbi.gov/scams-safety/fraud/seniors
Federal Trade Commission. (2013). Consumer Sentinel Network Data Book for January -
December 2012. Washington, DC.
Financial Industry Regulatory Authority (FINRA). (2006, May). Investor fraud study: Final
report. NASD Investor Education Foundation: Author. Retrieved from
http://www.finra.org/web/groups/foundation/@foundation/documents/foundation/p118422.
pdf
FINRA Investor Education Foundation. (September, 2013). Financial Fraud and Fraud
Susceptibility in the United States. Research Report from a 2012 National Survey. Applied
Research and Consulting: New York, NY.

109

Folstein, M. F., Folstein, S. E., & McHugh, P. R. (1975). Mini-mental state: A practical method
for grading the cognitive state of patients for the clinician. Journal of Psychiatric
Research, 12, 189–198.
Friedman, M. (1992). Confidence swindles of older consumers. The Journal of Consumer Affairs
26(1): 20-46.
Ganzini, L., McFarland, B., & Bloom, J. (1990). Victims of fraud: Comparing victims of white
collar and violent crime. Bulletin of the American Academy of Psychiatric Law, 18(1), 55-
63.
Gassoumis, Z. D., Navarro, A. E., & Wilber, K. H. (2014). Protecting victims of elder financial
exploitation: the role of an Elder Abuse Forensic Center in referring victims for
conservatorship. Aging & mental health, (ahead-of-print), 1-9.
Gentry, J. W., Kennedy, P. F., Paul, K., & Hill, R. P. (1995). The vulnerability of those grieving
the death of a loved one: Implications for public policy. Journal of Public Policy &
Marketing, 128-142.
Goergen, T., & Beaulieu, M. (2010). Criminological theory and elder abuse research—Fruitful
relationship or worlds apart?. Ageing International, 35(3), 185-201.
Gold, J. (2013, April 22). Seniors Get Hung Up In Health Care Scams. Kaiser Health News.
Available from http://www.everydayhealth.com/senior-health/seniors-get-hung-up-in-
health-care-scams.aspx

110

Greenspan, S., Loughlin, G., & Black, R. S. (2001). Credulity and gullibility in people with
developmental disorders: A framework for future research. International review of
research in mental retardation, 24, 101-135.
Havighurst, R. J., & Albrecht, R. E. (1953). Older people. Longmans, Green & Co: New York,
NY.
Hedden, T., & Gabrieli, J. D. E. (2004). Insights into the ageing mind: A view from cognitive
neuroscience. Nature Reviews Neuroscience, 5, 87–96.
Hidalgo, B., & Goodman, M. (2013). Multivariate or multivariable regression? American
Journal of Public Health, 103(1), 1-3.
Holtfreter, K., Reisig, M. D., & Pratt, T. C. (2008). Low self‐control, routine activities, and fraud
victimization. Criminology, 46(1), 189-220.
Jackman, M. R. (2002). Violence in social life. Annual Review of Sociology, 28, 387-415.
Jackson, S. L., & Hafemeister, T. L. (2011). Financial abuse of elderly people vs. other forms of
elder abuse: Assessing their dynamics, risk factors, and society’s response. Final Report to
U.S. Department of Justice (Document. No. 233613).
Jackson, S. L., & Hafemeister, T. L. (2012). Pure financial exploitation vs. hybrid financial
exploitation co-occurring with physical abuse and/or neglect of elderly persons. Psychology
of Violence, 2(3), 285–96.
Karp, N. (2012). Protecting older investors: The challenge of diminished capacity. Generations,
36(2), 33–38.

111

Kass, R. E. & Wasserman, L. A. (1995). A reference Bayesian test for nested hypotheses and its
relationship to the Schwarz criterion. Journal of the American Statistical Association, 90,
928-934.
Kemp, B., & Mosqueda, L. (2005). Elder financial abuse: An evaluation framework and
supporting evidence. Journal of American Geriatric Society, 53, 1123-1112.
Knight, A. J., & Marson, D. C. (2012). The Emerging Neuroscience of Financial Capacity.
Generations, 36(2), 46-52.
Langenderfer, J., & Shrimp, T. A. (2001). Consumer vulnerability to scams, swindles, and fraud:
A new theory on visceral influences on persuasion. Psychology & Marketing, 18(7), 763–
783.
Lee, J., & Geistfeld, L. V. (1999). Elderly consumers’ receptiveness to telemarketing fraud.
Journal of Public Policy & Marketing, 18(2), 208–217.
Lee, J., & Soberon-Ferrer, H. (1997). Consumer vulnerability to fraud: Influencing factors. The
Journal of Consumer Affairs, 31(1), 70-89.
Lerner, J. S., Li, Y., & Weber, E. U. (2013). The financial costs of sadness. Psychological
Science, 24(1), 72–79.
Lewis, R. C. (2012, August 16). Why are elderly duped? Area in brain where doubt arises
changes with age. ScienceDaily. Retrieved from
http://www.sciencedaily.com/releases/2012/08/120816121836.htm

112

Lichtenberg, P. A., Stickney, L., & Paulson, D. (2013). Is psychological vulnerability related to
the experience of fraud in older adults? Clinical Gerontologist, 36(2), 132–46.
Lloyd, J. (2012, December 8). Why are elderly more vulnerable to scams? Brain changes. USA
Today. Retrieved from http://www.usatoday.com/story/news/nation/2012/12/07/elderly-
financial-exploitation-holidays/1746911/
Lo, Y. T., Mendell, N. R., & Rubin, D. B. (2001). Testing the number of components in a normal
mixture. Biometrika, 88, 767–778.
Loeb, P. A. (1996). Independent living scales (ILS) manual. San Antonia, TX: Psychological
Corporation.
Lowenstein, A. (2009). Elder abuse and neglect—“old phenomenon:” New directions for
research, legislation and service developments. Journal of Elder Abuse & Neglect, 21(3),
278–287.
Magidson, J., & Vermunt, J. K. (Eds.). (2004). Latent class models. The Sage handbook of
quantitative methodology for the social sciences, pp. 175-198.
Marson, D. C., Martin, R. C., Wadley, V., Griffith, R. H., Snyder, S., Goode, P. S., Kinney, F.
C., et al. (2009). Clinical interview assessment of financial capacity in older adults with
mild cognitive impairment and Alzheimer’s disease. Journal of the American Geriatrics
Society, 57(5), 806–814.
Martin, N. (2009). Consumer scams and the elderly: Preserving independence through shifting
default rules. Elder Law Journal, 17, 1.

113

Martin, R. C., Okonkwo, O. C., Hill, J., Griffith, H. R., Triebel, K., Bartolucci, A., ... & Marson,
D. C. (2008). Medical decision‐making capacity in cognitively impaired Parkinson's
disease patients without dementia. Movement Disorders, 23(13), 1867-1874.
McCutcheon, A. L. (Ed.). (1987). Latent class analysis (No. 64). Sage.
Mead, N. L., Baumeister, R. F., Stillman, T. F., Rawn, C. D., & Vohs, K. D. (2011). Social
exclusion causes people to spend and consume strategically in the service of affiliation.
Journal of Consumer Research, 37(5), 902–919.
MetLife Mature Market Institute. (October, 2009). Market survey of long-term care costs: The
2009 MetLife market survey of nursing home, assisted living, adult day services and home
care costs. Retrieved
from http://www.metlife.com/assets/cao/mmi/publications/studies/mmi-market-survey-
nursing-home-assisted-living.pdf
Miles, M. B., Huberman, A. M., & Saldaña, J. (2013). Qualitative data analysis: A methods
sourcebook. SAGE Publications, Incorporated.
Montgomery, S. A., & Asberg, M. (1979). A new depression scale designed to be sensitive to
change. The British journal of psychiatry, 134(4), 382-389.
Moos, R. H., Schutte, K. K., Brennan, P. L., & Moos, B. S. (2005). The interplay between life
stressors and depressive symptoms among older adults. The Journals of Gerontology
Series B: Psychological Sciences and Social Sciences, 60(4), P199-P206.
Morse, J. M. (1995). The significance of saturation. Qualitative health research, 5(2), 147-149.

114

Moye, J., & Marson, D. C. (2007). Assessment of decision-making capacity in older adults: an
emerging area of practice and research. The Journals of Gerontology Series B:
Psychological Sciences and Social Sciences, 62(1), P3-P11.
Muthén, L. K., & Muthén, B. (Eds.) (2006). Mplus. Los Angeles, CA: Muthén & Muthén.
Naimark, D. (2001). Financial exploitation of the elderly: The evaluation of mental capacity and
undue influence. American Journal of Forensic Psychiatry, 22(3), 5-20.
Navarro, A. E., Gassoumis, Z. D., & Wilber, K. H. (2013). Holding abusers accountable: an
elder abuse forensic center increases criminal prosecution of financial exploitation. The
Gerontologist, 53(2), 303-312.
Navarro, A. E., Wilber, K. H., Yonashiro, J., & Homeier, D. C. (2010). Do we really need
another meeting? Lessons from the Los Angeles County Elder Abuse Forensic Center. The
Gerontologist 50(5), 702–711.
Nerenberg, L. (2006). Communities respond to elder abuse. Journal of gerontological social
work, 46(3-4), 5-33.
Pak, K. & Shadel, D. (2011). AARP Foundation national fraud victim study. Washington DC.
Retrieved from assets.aarp.org/rgcenter/econ/fraud-victims-11.pdf
Park, D. C., Smith, A. D., Lautenschlager, G., Earles, J. L., Frieske, D., Zwahr, M., & Gaines, C.
L. (1996). Mediators of long-term memory performance across the life span. Psychology
and aging, 11(4), 621-637.

115

Phillips, D. M., & Stanton, J. L. (2004). Age-related differences in advertising: Recall and
persuasion. Journal of Targeting, Measurement and Analysis for Marketing, 13(1), 7-20.
Pillemer, K. (1985). The dangers of dependency: New findings on domestic violence against the
elderly. Social Problems, 33(2), 146-158.
Pinsker, D. M., McFarland, K., & Pachana, N. A. (2010). Exploitation in older adults: Social
vulnerability and personal competence factors. Journal of Applied Gerontology, 29(6),
740-761.
Quinn, M. J. (2000). Undoing undue influence. Generations, 24(2), 65-69.
Rabiner, D. J., O'Keeffe, J., & Brown, D. (2005). A conceptual framework of financial
exploitation of older persons. Journal of Elder Abuse & Neglect, 16(2), 53-73.
Rainwater, L. (1974). What money buys: Inequality and the social meanings of income. New
York: Basic Books.
Ramaswamy, V., DeSarbo, W. S., Reibstein, D. J., & Robinson, W. T. (1993). An empirical
pooling approach for estimating marketing mix elasticities with PIMS data. Marketing
Science, 12(1), 103−124.
Randolph, C., Tierney, M. C., Mohr, E., & Chase, T. N. (1998). The Repeatable Battery for the
Assessment of Neuropsychological Status (RBANS): Preliminary clinical validity. Journal
of clinical and experimental neuropsychology, 20(3), 310-319.
Reis, M., & Nahmaish, D. (1998). Validation of the indicators of abuse (IOA) screen. The
Gerontologist, 38(4), 471–480.

116

Ross, M., Grossmann, I., & Schryer, E. (2014). Contrary to psychological and popular opinion,
there is no compelling evidence that older adults are disproportionately victimized by
consumer fraud. Perspectives on Psychological Science, 9(4), 427–442.
Ruffman, T., Murray, J., Halberstadt J., & Vater, T. (2012). Age-related differences in deception.
Psychology and Aging, 27(3), 543-549.
Saldaña, J. (2013). The coding manual for qualitative researchers (No. 14). Sage.
Salthouse, T. A. (1991). Mediation of adult age differences in cognition by reductions in working
memory and speed of processing. Psychological Science, 2(3), 179–83.
Schoepfer, A., & Piquero, N. L. (2009). Studying the correlates of fraud victimization and
reporting. Journal of Criminal Justice, 37, 209–215.
Schonfeld, L., Larsen, R. G., & Stiles, P. G. (2006). Behavioral health services utilization among
older adults identified within a state abuse hotline database. The Gerontologist, 46(2), 193–
199.
Spencer, W. D., & Raz, N. (1995). Differential effects of aging on memory for content and
context: A meta-analysis. Psychology and aging, 10(4), 527-539.
Strauss, A. & Corbin, J. (1990). Basics of qualitative research: Grounded theory procedures and
techniques. Newbury Park, CA: Sage Publications.
Titus, R. (2001). Personal fraud and its victims. In N. Shover & J. P. Wright (Eds.). Crimes of
privilege: Readings in white-collar crime. (pp. 57-66). Oxford University Press.

117

Titus, R. M., Heinzelmann, F. & Boyle, J. M. (1995). Victimization of persons by fraud. Crime
& Delinquency, 41, 54-72.
Triebel, K. L., & Marson, D. C. (2012). The warning signs of diminished financial capacity in
older adults. Generations, 36(2), 39–45.
Tueth, M. J. (2000). Exposing financial exploitation of impaired elderly persons. The American
Journal of Geriatric Psychiatry, 8(2), 104-111.
Turner, R. J., Wheaton, B., & Lloyd, D. A. (1995). The epidemiology of social stress. American
Sociological Review, 60, 104-125.
U.S. Government Accountability Office (GAO). (2012, November). Elder justice: National
strategy needed to effectively combat elder financial exploitation. (Publication No. GAO-
13-110). Retrieved from http://www.gao.gov/products/GAO-13-110
U.S. Senate Special Committee on Aging. (2013, November). Senate aging committee launches
new anti-fraud hotline, enhanced website to assist seniors. (Press Release). Washington,
DC: Author. Retrieved from http://www.aging.senate.gov/
Van Wyk, J. & Benson, M. J. (1997). Fraud Victimization: Risky Business or just bad luck?
American Journal of Criminal Justice, 21(2), 163–79.
Van Wyk, J., & Mason, K. A. (2001). Investigating vulnerability and reporting behavior for
consumer fraud victimization: Opportunity as a social aspect of age. Journal of
Contemporary Criminal Justice, 17, 328–345.

118

Wallace, H. (2007). Victimology: Legal, psychological, and social perspectives. Boston: Pearson
A&B.
Welsh, K. A., Breitner, J. C. S., & Mgruder-Habib, K. M. (1993). Detection of dementia in the
elderly using telephone screening of cognitive status. Neuropsychiatry, Neuropsychology
& Behavioral Neurology, 6(2), 103–110.
Wilber, K. W. (1990). Material abuse of the elderly: When is guardianship a solution?. In E.
Dejowski (Ed.). Protecting judgment-impaired adults: Issues, interventions and policies,
(pp. 89-104). Haworth Press.
Wilber, K. H., & Reynolds, S. L. (1997). Introducing a framework for defining financial abuse of
the elderly. Journal of Elder Abuse & Neglect, 8(2), 61-80.
Wood, S., Rakela, B., Liu, P. J., Navarro, A. E., Bernatz, S., Wilber, K. H., ... & Schneider, D.
(2014). Neuropsychological Profiles of Victims of Financial Elder Exploitation at the Los
Angeles County Elder Abuse Forensic Center (LACEAFC). Journal of elder abuse &
neglect, 26(4), 414-423.
Yin, R. K. (2009). How to do better case studies. The SAGE handbook of applied social research
methods, pp. 254-282

119

Appendix A. Consort flow diagram of HRS sample selection for Chapters II and III

No (n=11,524)
Combined 2008 and 2010 Leave-
Behind (LB) Samples (N=13,074)
Victim of fraud in
the past 5-years?
Yes (n=665) No (n=12,409)
Valid year reported
for fraud?
Yes (n=506)
No (n=159)
 No yr. (n=120 )
 Invalid yr. (n=39)
Age 50+ in the 2008
or 2010 LB Survey?
Yes (n=497)
No (n=460)
Age 50+ in the 2008
or 2010 LB Survey?
Yes (n=12,206) No (n=203)
D
R
O
P
FRAUD SAMPLE
NON-VICTIM
CONTROL SAMPLE
Analysis weight = 0?
No (n=9)
Analysis weight = 0?
Yes (n=37)
 Proxy interview (n=11)
 Nursing home (n=1)
Yes (n=682)
 Proxy interview (n=369)
 Nursing home (n=127)

Abstract (if available)

Abstract In 2011, 7.3% of U.S. adults ages 65-74 and 6.5% of adults ages 75 and older were victims of financial fraud (Anderson, 2013). In addition to the billions of dollars lost annually to scams, indirect societal costs include paying for the care and support of elders who lost their life savings, and the expense of investigating cases and prosecuting offenders. One of the first steps in stopping fraud is determining who is most vulnerable, yet the research literature on risk factors presents a conflicting narrative. Some research studies and consumer protection agencies report that older adults are the most vulnerable age group due to greater social isolation and impairments in financial decision‐making (Ganzini, McFarland, & Bloom, 1990

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Forensic markers of physical elder abuse in medical and community contexts: implications for criminal justice interventions

PDF

Developing a better understanding of elder abuse multidisciplinary teams: addressing gaps for research, policy, and practice

PDF

Palliative care: breaking the knowledge barrier among community-based older adults

PDF

Fall related injuries among older adults in the Los Angeles region

PDF

Intergenerational support between grandparents and grandchildren in rural China and its effect on the psychological well-being of older adults

PDF

The dollars and sense of caregiving: exploring the financial financial realities of providing informal care to an older adult

Asset Metadata

Creator DeLiema, Marguerite Irene (author)

Core Title Using mixed methods to identify the characteristics of older fraud victims

School Leonard Davis School of Gerontology

Degree Doctor of Philosophy

Degree Program Gerontology

Publication Date 03/03/2015

Defense Date 02/19/2015

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

Tag cognitive functioning,elder,elder financial abuse,elder fraud,financial exploitation,fraud,fraud target,health and retirement study,latent class analysis,Loneliness,OAI-PMH Harvest,older adult,perpetrators,routine activity theory,scams,stressful life events,victim

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Wilber, Kathleen H. (committee chair), Crimmins, Eileen M. (committee member), Enguidanos, Susan M. (committee member)

Creator Email m.deliema@gmail.com,riparett@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-537449

Unique identifier UC11297795

Identifier etd-DeLiemaMar-3216.pdf (filename),usctheses-c3-537449 (legacy record id)

Legacy Identifier etd-DeLiemaMar-3216.pdf

Dmrecord 537449

Document Type Dissertation

Format application/pdf (imt)

Rights DeLiema, Marguerite Irene

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