University of Southern California Dissertations and Theses

Observed and self-reported childrearing in mothers with a history of sexual abuse

(USC Thesis Other)

Observed and self-reported childrearing in mothers with a history of sexual abuse

PDF

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content OBSERVED AND SELF-REPORTED CHILDREARING IN MOTHERS
WITH A HISTORY OF SEXUAL ABUSE

by

Christine Payne Bryson

__________________________________________________________________

A Dissertation Presented to the
FACULTY OF THE GRADUATE SCHOOL
UNIVERSITY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Requirements for the Degree
DOCTOR OF PHILOSOPHY
(SOCIAL WORK)

August 2007

Copyright 2007 Christine Payne Bryson
ii
ACKNOWLEDGMENTS

I would like to acknowledge those who have contributed to my personal and
professional development during this process. First, I am most grateful to my
husband, Scott, who tirelessly supported me with minimal eye-rolling when I decided
over and over again to ―hang out and do fun things‖ instead of being a ―real grad
student‖; with more than his share of childcare and household duties; and most
importantly with humor, encouragement, and perspective when I struggled. He
always makes life fun and meaningful. Next, my boys—Ben, Luke, and J.P.—have
fueled my desire to learn more, to play harder, and to pay attention to what matters
most. All three of them have evoked instincts in me that have driven me to
understand the science of childrearing and attachment, and it is a privilege to be their
mom. I am also deeply appreciative of my family—Deborah and Galen Buckwalter,
O.J. Bryson, Judy and Bill Ramsey, Gary Payne, and Jen, both Chrises, and
Katinka—who are always on my side. Special thanks to Mom and Galen, who held
my hand through the trenches of this dissertation process. Additionally, I am grateful
to Wes and Emilie Cunningham for Sunday nights, which kept life fun.
I also want to thank Dr. Daniel Siegel for his mentorship and investment in me
academically, personally, and professionally. It is a pleasure to learn from and work
with him. Some of the best things I know, I have learned from him. I am appreciative
to Dr. Penelope Trickett who allowed me to use this rich longitudinal data for this
study. Lastly, thanks are in order to Dr. Bruce Jansson who encouraged me when I
most needed it.
iii
TABLE OF CONTENTS
ACKNOWLEDGEMENTS ...................................................................................... ii
LIST OF TABLES .................................................................................................... v
ABSTRACT .............................................................................................................. vi
CHAPTER 1: INTRODUCTION ......................................................................... 1
Overview of Study ......................................................................................... 1
Child Sexual Abuse .............................................................................. 2
Sexual Abuse History and Childrearing ........................................................ 4
Childrearing Dimensions: Nurture and Control ........................................... 7
Figure 1: Baumrind‘s Orthogonal Domains and Typology of
Childrearing .......................................................................................... 32
Sexual Abuse History and Dimensions of Warmth and Control ................... 10
Conclusion ..................................................................................................... 15
Specific Aims ................................................................................................ 17
Study Hypotheses .......................................................................................... 18

CHAPTER 2: METHODS ..................................................................................... 19
Participants .................................................................................................... 19
Measures ........................................................................................................ 21
Procedures ..................................................................................................... 23
Data Collection ..................................................................................... 23
Procedures for Developing the Observational Coding System ............ 23
Procedures for Implementing the Observational Coding System ........ 28

CHAPTER 3: RESULTS ....................................................................................... 31
General Preliminary Analysis ........................................................................ 31
Data Preparation and Missing Data ...................................................... 31
Observational Childrearing Data: Data Reduction and Measurement
Development ......................................................................................... 32
Item-level Analysis ............................................................................... 33
Behavioral Frequency: Factor Analysis and Reliability ...................... 34
Global Ratings: Factor Analysis and Reliability Analysis .................. 38
Self-report Childrearing Data: Data Reduction and Measurement
Development ......................................................................................... 39
Item-level Analysis ............................................................................... 39
Self-report Discipline Items: Factor Analysis and Reliability ............ 40
Self-report Global Parenting Style Items: Factor Analysis
and Reliability ............................................................................ 42
iv
Relationships between Observational and Self-report Data .......................... 43
Summary ........................................................................................................ 46
Demographic Analysis .................................................................................. 46
Statistical Analysis of Composites ................................................................ 48
Sexual Abuse History and Childrearing ............................................... 48

CHAPTER 4: DISCUSSION ................................................................................. 51
Limitations ..................................................................................................... 59
Approaches to Measuring Parenting .................................................... 59
Measurement: Sexual Abuse Status .................................................... 60
Racial and Cultural Issues .................................................................... 61
Generalizing Results ............................................................................. 62
Future Research ............................................................................................. 63
Sexual Abuse ........................................................................................ 63
Childrearing .......................................................................................... 64
Conclusion ..................................................................................................... 65

REFERENCES .......................................................................................................... 66

APPENDICES ........................................................................................................... 75
Appendix A: Sexual Abuse History Interview ............................................. 76
Appendix B: Self-reported Childrearing Types of Discipline Used ............ 77
Appendix C: Self-reported Childrearing—Parenting Style .......................... 78
Appendix D: Etch-a-Sketch Parenting Style Code System Manual ............. 79
Appendix E: Intra-Class Correlations Scores for Coding System Items ...... 99

v
LIST OF TABLES

Table 1: Maccoby and Martin‘s Parenting Styles Classified by Two
Dimensions ............................................................................................. 10

Table 2: Empirical Studies on Child Sexual Abuse and Dimensions of
Warmth and Control as Parenting Outcomes .......................................... 14

Table 3: Rotated Component Factor Loadings for Behavioral Items .................... 37

Table 4: Rotated Component Factor Loadings for Global Items .......................... 39

Table 5: Rotated Component Factor Loadings for Discipline Items ..................... 41
Table 6: Rotated Component Factor Loadings for Parenting Style Items ............. 42
Table 7: Inter-correlations of Variables ................................................................. 45

Table 8: Standardized Beta Coefficients of Sexual Abuse History with
Observed and Self-Reported Childrearing Behaviors and Global
Parenting Style ......................................................................................... 49

Table 9: Sample Means (Standard Deviations, and Ranges) of Childrearing
Variables .................................................................................................. 50

vi
ABSTRACT

The aims of this study were a) to develop a coding system for parenting
behaviors based on the domains of warmth and control, b) to extract meaningful
constructs from both observational and self-report data, and to examine the
correspondence between them, and c) to examine the impact of childhood sexual
abuse on aspects of parental warmth and control. The sample was comprised of one
hundred and twenty-seven mothers whose daughters were subjects in a longitudinal
study in which half of the daughters, aged 6-16, were victims of familial sexual abuse,
and a demographically similar comparison group.
Principal component factor analyses and reliability analyses yielded four
observational and three self-reported constructs related to the parenting dimensions of
warmth and control. Using hierarchical regression analyses, the impact of child
sexual abuse on the extracted warmth and control constructs was examined. Overall,
in terms of warmth, findings indicated that mothers with a history of sexual abuse
were not different from non-abused mothers in the amount of self-reported warmth,
but were observed to be significantly warmer towards their daughters. In terms of the
control dimension, mothers with a history of child sexual abuse reported difficulty
providing positive structure/control in their parenting.
This study contributes to the emerging body of research that suggests that
parenting may indeed be impacted by childhood sexual abuse, by providing a fuller
understanding of the possible ways that past sexual abuse experiences may impact
different aspects of parental warmth and control.
1
CHAPTER 1: INTRODUCTION
Overview of Study
This study involved the analysis of existing data from an 18-year-long study on
girls, aged 6-16 at the beginning of the study, half of whom were victims of familial
sexual abuse, and half of whom comprised a comparison group. This longitudinal
study, in which the 6
th
phase of data collection has just been completed, has received
federal support from NIMH and ACYF for numerous years, and is the longest running
study on sexual abuse in the history of the United States.
1
To date, almost without
exception, the analyses and publications from this data have focused on the girls in the
study. The present study used observational and self-report data obtained from the
mothers of these girls to investigate relationships between childhood sexual abuse and
childrearing.
While child sexual abuse is a significant problem in our society, research on the
association between child sexual abuse and parenting is in its infancy. Unfortunately,
the majority of studies examining relationships between child sexual abuse and
parenting are characterized by a variety of methodological problems which have
resulted in a contradictory body of literature with limited generalizability. Furthermore,
a review of the literature revealed that only four studies have used observational data to
measure parental behavior.

1
Numerous publications have resulted from this study (Trickett, Noll, Reifman & Putnam,
2001).
2
Using observational and self-report data, this study built on existing empirical
work to better understand the childrearing perceptions and behaviors of the mothers in
this invaluable community-based sample of women and their daughters. Further,
including an observational component of data collection and analysis in this study
allowed for the identification of more discrete observational constructs by which
parenting, and the potentially important influence of sexual abuse history on
childrearing, can be more rigorously evaluated. Thus the development of an
observational coding instrument, measuring parental warmth and control, was also a
major objective of this study.
This introduction will establish the widespread problems of child sexual abuse
and the repeated patterns of sexual victimization seen from one generation to the next.
Then, a discussion of limited theoretical understanding and limited empirical data on
how child sexual abuse impacts parental warmth and control will be used to establish
the need for descriptive exploratory analyses. Finally, the specific aims and hypotheses
of this study will be presented.
Child Sexual Abuse
Child sexual abuse is a widespread and harmful problem in our society. With
estimations that as many as 1 out of 4 women have been victims, (Finkelhor, Moore, L.,
& Strauss, 1997), it is unfortunate that the majority of research on child sexual abuse
has only begun to emerge in the last few decades. Like other forms of child abuse,
sexual abuse is an intergenerational phenomenon (Kreklewetz & Piotrowski, 1998;
Oates, Tebbutt, Swanston, Lynch, & O'Toole, 1998; Trickett, Everett, & Putnam, 1995;
3
Trickett, Noll, Reiffman, & Putnam, 2001). Theories about the causes of
intergenerational transmission of maltreatment include social learning theory, impaired
parental functioning as a result of an abuse history, enmeshed family boundaries, family
and social environments that make children vulnerable, and insecure attachment, to
name a few (Burkett, 1991; Maker & Buttenheim, 2000; Morton & Browne, 1998;
Newcomb & Locke, 2001; Sroufe & Fleeson, 1986). These theories attempt to explain
this complicated association. While some male victims do become perpetrators of
sexual abuse, it is rare for female victims to become perpetrators. Therefore, these
theories attempting to explain the cycle of abuse may be particularly limited in their
application to sexual abuse because they primarily address how a victim becomes a
perpetrator, not how a victim has a child who then becomes a victim, which is more
likely the case for females. Thus, given the general paucity of literature and
explanatory models, it is difficult to determine whether and how these theories apply
specifically to intergenerational sexual victimization.
This intergenerational pattern has been shown to be evident in the data
collected for this study as well; mothers whose daughters were sexually abused reported
significantly more sexual abuse themselves (49.2%) than the mothers of girls who had
not been sexually abused (16.4%) (Trickett et al., 1995). Other studies similarly find
that sexually abused children have significantly disproportionate percentages of mothers
with sexual abuse histories (Briere & Runtz, 1988; Browne & Finkelhor, 1986; Burkett,
1991; Finkelhor et al., 1997; Kreklewetz & Piotrowski, 1998; Oates et al., 1998). Such
findings prompt questions regarding why and how this repeated victimization happens,
4
and whether parenting plays a role. Yet, there is currently no conceptual model
explaining the mechanisms of this intergenerational phenomenon. First, more must be
understood about which, if any, aspects of parenting are affected by child sexual abuse,
since the childrearing environment in these families is likely to be different in many
ways (Trickett et al., 1995). A comprehensive theoretical or empirical model for the
mechanism of transmission was outside the scope of this study. Rather, the focus of
this study was to examine the potential impact of child sexual abuse on two important
aspects of parenting—warmth and control. If the degree of warmth or control differed
for mothers with a history of sexual abuse, these differences may be important
contributors to this frequently seen association.
Sexual Abuse History and Childrearing
It is well established that child sexual abuse is associated with numerous
difficulties in adulthood, such as a higher risk for revictimization (Noll, Horowitz,
Bonanno, Trickett, & Putnam, 2003), and for psychological (Briere & Runtz, 1993;
Trickett & Putnam, 1993), physical (Sickel, Noll, Moore, Putnam, & Trickett, 2002),
and interpersonal problems (DiLillo, 2001; Rumstein-McKean & Hunsley, 2001;
Trickett, 1997).
2
For mothers with a history of child sexual abuse, psychological and
interpersonal difficulties, in particular, may interfere with the ability to parent
optimally.

2
For a review of the sequelae of CSA, see (Trickett & Putnam, 1998).
5
The first literature documenting concerns regarding parenting in incest survivors
began to emerge from clinical reports in the 1980s (Fitzgerald, Shipman, Jackson,
McMahon, & Hanley, 2005). Empirical research conducted since then has provided
important information regarding the effects of sexual abuse on parenting (for a review,
see Ruscio, 2001; DiLillo, 2003). Many studies have found that child sexual abuse has
a negative effect on various aspects of parenting. For example, mothers with a history
of childhood sexual abuse have been found to act more aggressively (Newcomb &
Locke, 2001), focus less on their child and more on themselves (Burkett, 1991),
experience more parenting stress (Douglas, 2000), communicate poorly, and see their
parental role more negatively (Banyard, 1997; Cohen, 1995). Other studies found that
mothers who were victims of child sexual abuse show more role reversal with their
child (Alexander, 2000; Burkett, 1991), report less parenting competence (Zuravin &
Fontanella, 1999), and experience less satisfaction with their childrearing role (Herman,
1981; Herman & Hirschman, 1981).
However, most studies that address parenting characteristics of women with
sexual abuse histories are plagued by a variety of significant limitations.
Methodological problems—such as using only clinical samples, not using a control or
comparison group, and using small samples—are further compounded by the fact that
most studies use varying definitions or criteria for child sexual abuse (CSA) (DiLillo,
2001; Trickett, 2005, 2006). Even as recently as 2001, one article stated, ―With regard
to maternal functioning, one study examined childrearing attitudes and behavior but did
not include a non-CSA control group, thus making it difficult to interpret the meaning
6
of the results. No research has examined the actual maternal functioning of CSA
survivors in a community sample‖ (Rumstein-McKean & Hunsley, 2001).
Additionally, some researchers point out that many existing studies may not
distinguish between sexual abuse and other forms of child abuse (Everett, Trickett, &
Putnam, 1998; Kaufman & Zigler, 1987; Trickett, 2006). This lack of differentiation is
problematic because of the unique nature of sexual abuse. Unfortunately, then, much of
the scant research that has been done on the parenting of sexual abuse survivors is
opinion-based (though regarded as if it is research-based), vague, and methodologically
problematic, leaving scholars with little empirically valid information on these
populations (Rumstein-McKean & Hunsley, 2001; Tamraz, 1996).
3

It is important to note, in addition to the methodological limitations in many of
these studies, that findings rely almost exclusively on self-report data. In fact, only four
studies to date have used observational approaches to examine any aspect of parenting
among mothers with a sexual abuse history (Burkett, 1991; DiLillo, Tremblay, &
Peterson, 2000; Fitzgerald et al., 2005; Lyons-Ruth & Block, 1996). Self-report
instruments are of great value for gathering data on subjects‘ beliefs, perceptions, and
ideas about parenting. However, measuring behaviors and traits of parenting using
observation by trained coders combined with self-report instruments provide more

3
Though the childrearing beliefs and behaviors of mothers of children who have been physically
abused have been studied well, the research on the childrearing perceptions and behaviors of mothers in
families where sexual abuse occurs has just barely begun to emerge (Trickett, 1998).

7
stable data and avoid the biases that are inherent in self-reported measures alone
(Gardner, 2000).
4

Childrearing Dimensions: Warmth and Control
There are a variety of ways to think about and measure parenting. In fact,
dimensions of parenting have been characterized in many different ways throughout the
history of childrearing theory. Childrearing theory is an enormous field,
comprehensively covering how parents impact and socialize their children; this includes
parenting styles (the emotional background for parenting behaviors, or how parents do
things), parenting practices (specific parenting behaviors, or what parents do), and
parenting roles (such as nurturers or disciplinarians) (Cowan, Powell, & Cowan, 1998;
Locke & Prinz, 2002). In fact, parenting experts Cowan et al. (1998) state, ―There are
almost as many ways of conceptualizing and assessing parent-child relationships as
there are researchers and authors‖ (Locke & Prinz, 2002, p. 6).
5

Despite the variety of ways to look at parenting, there are two dimensions—
warmth and control—which are the most frequently used constructs to examine
parenting, and have been derived from many years of research. Historically, the

4
Gardner (2000) notes that while some studies report significant associations between self-
report and observational measures of the same construct during parent-child interactions, many of these
studies find Pearson correlations near r = .3, which suggests that both approaches separately contribute
uniquely. There is still much to be learned about the correspondence between self-report data and
observational data when measuring parenting. According to Gardner (2000), it is not clear whether self-
report or observational approaches best measure various dimensions of parenting.

5
Parenting practices are typically measured in terms of frequency or content of behaviors (such
as spanking or hugging or ignoring), whereas parenting styles and roles are usually measured more
globally on a continuum (such as strict-permissive, friend-authoritarian, serious-playful, consistent-
inconsistent) (Locke & Prinz, 2002).

8
childrearing literature first suggested parenting styles could be viewed as two
orthogonal variables of warmth/hostility and control/autonomy (Schafer, 1959). Then,
parenting styles were commonly examined as warmth/acceptance versus hostility as one
dimension, and restrictive versus permissive as the other (Becker, 1964). Following
these approaches, Diana Baumrind‘s well-known typology came to emphasize two
similar dimensions (Maccoby & Martin, 1983). Based on parent-child interactions,
Baumrind (1972) suggested a classification of parenting style based upon two parenting
style dimensions, which are depicted in Figure 1: (a) warmth/responsiveness, and (b)
control—which includes discipline, structure and limit setting, as well as
communicating expectations for behavior (Baumrind, 1972; Cowan et al., 1998).
6

Using these two dimensions, Baumrind categorized three types of parenting styles:
authoritative parents are highly responsive and warm and provide structure and limits
that are communicated to the child; permissive parents are similarly warm and
responsive, but exercise little control in terms of limit setting or expectations;
authoritarian parents demonstrate high control, but are largely cold and are frequently
unresponsive to their children (Baumrind, 1972; Cowan et al., 1998; Maccoby &

6
Discipline, an ever-popular aspect of childrearing, has often been thought of in terms of the
control or restrictiveness dimension. Maccoby and Martin discuss how discipline was conceptualized
more specifically beyond just degree of control, with more of a focus on the type of discipline, "In these
early studies, techniques of discipline tended to be classified under two major headings: (1) power-
assertive discipline, including physical punishment, yelling, shouting, forceful commands, and threats;
and (2) 'love-oriented' discipline, including showing disappointment, isolation, withdrawal of love, praise,
contingent giving of affection, and reasoning."; however, these dimensions that came from "early factor
analyses were only weakly related to the theories that had guided the formulation of the studies" (1983,
p. 37).

9
Martin, 1983). Research has demonstrated that for the most part, though not entirely
across the board, authoritative parenting is most successful in terms of child outcomes
(Baumrind, 1991; Maccoby & Martin, 1983).

Figure 1. Baumrind’s orthogonal domains and typology of childrearing.
7

High Control

Authoritative Authoritarian

Warm/ Cold/
Responsive Unresponsive

Permissive

Low Control

Following Baumrind‘s (1972) typology, Maccoby and Martin (1983) observed
that these three parenting styles might not encompass the full range of parenting
behaviors, and proposed that parenting styles be viewed as two orthogonal dimensions:
responsiveness and demandingness (Cowan et al., 1998; Maccoby & Martin, 1983). As
seen in Table 1, they combined these dimensions to derive four parenting types—the
same three as Baumrind, with a fourth category termed ―disengaged.‖

7
Adapted from (Baumrind, 1972).
10
Table 1. Maccoby and Martin’s Parenting Styles Classified by Two Dimensions
8

Responsive Unresponsive

Demanding/high control

Authoritative

Authoritarian

Undemanding/low control

Permissive

Disengaged

The responsiveness/warmth and demandingness/control dimensions are the
―two most heavily referenced constructs in the parenting research literature‖
(Kuczynski, 2003; Locke & Prinz, 2002, p. 895).
9
These two dimensions were helpful
constructs to begin with in this study, and were used to guide which items were
measured in the observational coding system, and which self-report data were selected
for analyses.
Sexual Abuse History and Dimensions of Warmth and Control
When it comes to parental warmth and control in women with a history of
sexual abuse, it is apparent that while subject profiles, methodology, definitions of
sexual abuse, and parenting dimensions vary, an examination of the literature reveals
conflicting results. Some research suggests that mothers with a history of CSA may be
giving inconsistent/disorganized guidance or discipline, and may not be making
developmentally appropriate demands on their children (Burkett, 1991; Cole & Putnam,

8
Adapted from Maccoby & Martin (1983).

9
For a full review of the past 20 years of measurement of parental discipline and nurturance (see
Locke & Prinz, 2002).
11
1992; Cole & Woolger, 1989; Cole, Woolger, Power, & Smith, 1992; Gelinas, 1983;
Herman, 1981; Main & Goldwyn, 1984; Ruscio, 2001). Additionally, mothers with a
sexual abuse past have been found to be more permissive than the general population
and less authoritarian (Cohen, 1995; Cole & Woolger, 1989; Cole et al., 1992; Ruscio,
2001), demonstrate less sensitivity and support toward their children (Cohen, 1995), and
engage less with their infants (Lyons-Ruth & Block, 1996). Even after controlling for
the quality of relationships in the family of origin, one study found that women who had
a history of sexual abuse had more negative perceptions of themselves as parents and
used more physical punishment (Banyard, 1997).
On the other hand, some studies do not find empirical evidence indicating that
child sexual abuse predicts problems with aspects of parental warmth and control.
Several studies report that mothers with a history of child sexual abuse are not different
from nonabused mothers in terms of authoritarianism, or how they use
discipline/control (Cole & Woolger, 1989; Cole et al., 1992; Schuetze & Eiden, 2005;
Trickett et al., 1995). Other studies have not found significant associations among child
sexual abuse and other parenting variables, such as communicating in a controlling way
(Burkett, 1991), parenting with sensitivity, nurturance and involvement (Cole &
Woolger, 1989), expressing attitudes about nurturing or discipline (Cole & Putnam,
1992), having developmentally appropriate expectations, and being punitive or
nurturing (DiLillo et al., 2000).
Thus, it is difficult to draw conclusions regarding the parenting dimensions of
warmth and control in sexually abused mothers because while both clinical and
12
empirical literatures indicate that mothers with a history of child sexual abuse may
struggle with certain aspects of parenting, other studies have not found evidence in
support of such difficulties. In fact, DiLillo (2003, p. 319) states that ―Despite
associations between CSA and parenting difficulties, the limited research addressing
specific aspects of parenting, and limitations in study design, preclude causal inferences
and make conclusions tentative at the present time.‖
The contradictory results in this body of literature may in part be the result of
differing approaches and numerous limitations in research methodology. In order to
better understand the findings of these studies and make hypotheses about these
mothers, a thorough review of studies investigating the impact of child sexual abuse on
the dimensions of parental warmth and control was conducted. Although a number of
studies discussed previously examined the impact of child sexual abuse on parenting,
few studies have focused on the theoretical domains of warmth and control. In fact,
only twelve studies were identified in which child sexual abuse was examined in
relation to aspects of warmth and/or control. To evaluate the findings of these twelve
studies, it was necessary to set standards for which studies would be considered. In
order to say with any confidence that it was a history of child sexual abuse that was
accounting for particular parenting outcomes, sexually abused mothers must have been
compared to mothers with no such history, but, who were otherwise comparable.
13
Therefore, only studies that met four criteria were included.
10
First, child
sexual abuse was not combined with other potentially confounding characteristics. For
example, the sample in one study was comprised of mothers whose fathers were
incestuous and alcoholic; another study‘s abused group was also being treated for
substance abuse. Studies such as these were not included. Second, the study must have
a comparison or control group which was comprised of individuals with no history of
child sexual abuse. Third, because demographic characteristics are known to be
associated with parenting, the comparison group must have been similar to the CSA
group in terms of salient demographic characteristics, such as age, SES, and race, or
these differences were controlled for statistically. Fourth, the study must have used a
non-clinical sample, such as a community sample, a university sample, or another
sample pool that was not utilizing intervention or mental health services. An alphabetic
listing of empirical studies investigating the impact of child sexual abuse on the
parenting dimensions of warmth and/or control which met the four criteria listed above
can be found in Table 2. After applying the above criteria, only three of the twelve
studies remained, and only one of these studies used observational methods.
11

10
These standards follow the example of criteria given in Trickett & McBride-Chang, 1995.
11
Of the 9 studies which did not meet the criteria for inclusion, 5 studies used clinical samples, 2 studies
did not use a comparison group, and 2 studies combined child sexual abuse with other potentially
confounding characteristics.

14
Table 2. Empirical Studies on Child Sexual Abuse and Dimensions of Warmth and Control as Parenting Outcomes

Study Data Collection
Method
Sample N and
Characteristics
Definition of Sexual
Abuse
Findings
Banyard, V.,
1997
Self-report N=518 low-income
mothers (237 were
involved with CPS for
physical abuse or neglect
& 281 control mothers on
welfare with no CPS
record); (note: missing
data skewed sample
towards women with CSA
and whose children had
been abused)
Fondling or intercourse
against her wishes < age
18.
Warmth—not examined

Control--CSA associated with
greater use of physical punishment

Fitzgerald, M.,
Shipman, K.,
Jackson, J.
McMahon, R.,
& Hanley, H.,
2005
Observational &
self-report
N=35 (17 incest survivors
& 18 nonabused women
and their 3-6 year old
children).
Recruited from a
community sample.

Sexual victimization < age
17 by male family
member in caregiving
role. All survivors
indicated that they were
unwilling participants and
self-identified their
experiences as sexual
abuse.
Warmth—CSA and no CSA were
comparably supportive, affectionate
and interactional styles with their
children were positive

Control—not examined
Schuetze, P. &
Das Eiden, R.,
2005
Self-report N=263 (107 CSA & 156
no CSA) recruited from a
community prenatal clinic
At least one contact (e.g.,
intercourse, genital
contact) or noncontact
(e.g., sexual invitations,
exhibitionism) episode by
either a family or
nonfamily member < age
18 & perpetrator was > 4
yrs. older than the victim
at the time of the incident
OR force was used.
Warmth—not examined

Control—CSA associated with
higher punitive discipline, but once
maternal depression and partner
violence were entered into the
model, CSA and parenting outcomes
had no relationship

CSA=mothers with a history of child sexual abuse. No CSA=mothers did not have a history of child sexual abuse.

15
Note that only one study investigating warmth, and only two studies
investigating control, could be confidently considered. In terms of warmth, despite the
clinical and empirical perceptions that child sexual abuse is linked to problems with
parenting, this study, with 35 subjects, used observational and self-report data and
found that incest survivors were comparable to nonabused mothers by demonstrating
moderate to high levels of support, assistance, and affection with their children
(Fitzgerald et al., 2005).
12

The two studies examining aspects of control in mothers with a history of child
sexual abuse had different results. The Banyard (1997) study found that child sexual
abuse was associated with greater use of physical punishment. However, the Schuetze
& Das Eiden (2005) study indicated that although CSA was associated with higher
punitive discipline, once maternal depression and partner violence were added into the
model, child sexual abuse and punitive discipline had no significant relationship.
In sum, there is a great paucity of competent research on the most prevalent parenting
dimensions—warmth and control—for mothers with a history of sexual abuse.
Conclusion
The existing maltreatment literature that examines parenting indicates that
mothers who have been sexually abused in childhood may or may not struggle with
aspects of the dimensions of warmth and control. Although previous research on the

12
A semi-structured lab session, which included free play, clean-up, and ten minute sets of
parent-child interactions, was used for the observational component of this study. Interactions consisted
of problem-solving tasks and developmentally challenging puzzles, in which the child needed parental
assistance. The variables of interest for this study were supportive presence, quality of assistance,
maternal confidence, and child‘s affection towards mother.

16
impact of child sexual abuse on parental warmth and control has provided important
information, there are three main problems. First of all, the literature is riddled with
methodological problems that make it difficult to draw conclusions about how, if at all,
mothers‘ parenting is affected by childhood sexual abuse. In fact, only three studies can
be confidently considered. Secondly, literature to date has almost exclusively relied on
self-report data; studies that include the observing of actual parenting behaviors of
mothers with a history of sexual abuse are almost nonexistent (Fitzgerald et al., 2005).
Lastly, the literature is conflicting. While some studies have demonstrated that child
sexual abuse history is associated with a variety of parenting difficulties in both
dimensions of warmth and control, other studies have found no evidence of this nature.
Put simply, then, one question that has not been sufficiently explored in the
literature is: To what degree and in what ways does a sexual abuse history impact
parental warmth and control? Because it was uncertain whether mothers who have
experienced sexual abuse have more difficulty parenting than their nonabused
counterparts, it was important that this study build upon the scant and problematic
existing studies, and explore not only whether this was the case in this sample, but also
whether any problems mothers experienced tended to be related to warmth, to control,
or to both.
Therefore, one main goal of this study was to examine whether the warmth and
control dimensions of parenting were different for mothers who have been sexually
abused. In this sample, if mothers who experienced child sexual abuse parent
differently from their nonabused counterparts, this finding could yield important

17
information about the impact of child sexual abuse on parenting and differences in the
childrearing environment. Thus, using a community sample with a demographically
controlled comparison group, both self-report and observational variables reflecting
childrearing behaviors and perceptions that conceptually fall under both dimensions
were examined in this study.
Specific Aims
Expanding the knowledge base regarding mothers who had a sexual abuse
history was accomplished by examining observational and self-reported parenting
dimensions of warmth and control in a community sample. To that end, the specific
research objectives of this study were the following:
1. To develop an observational coding system focused on childrearing
dimensions of warmth and control in both parental style (global ratings) and practices
(behavioral frequencies).
2. To reduce the observational data into meaningful parenting dimensions
focused on warmth and control through exploratory measurement development and
psychometric analysis.
3. To reduce self-report data into meaningful parenting dimensions focused
on warmth and control through exploratory measurement development and
psychometric analysis, and to examine the relationships between observational and self-
report data in this sample.
4. To examine whether sexual abuse history is related to childrearing
dimensions of warmth and control, using both observational and self-report data.

18
The rationale for these specific aims was to obtain observational data and allow
future studies to measure parental warmth and control with observational data, to
develop a better understanding of the relevance of the warmth and control dimensions,
to assure that measures were psychometrically valid, and to provide information on the
impact that child sexual abuse has on these meaningful parenting constructs.
Study Hypotheses
As stated, research on the impact of child sexual abuse on parenting is in its
infancy. Further, because of the conflicting nature of existing empirical studies in terms
of how child sexual abuse impacts parenting, and because of the methodological issues
that often limit interpretation of results, it was difficult to formulate hypotheses about
the relationships between a history of child sexual abuse and the dimensions of warmth
and control in this exploratory study. However, attachment and childrearing theories,
coupled with empirical studies on the long-term impacts of sexual abuse, suggest that
survivors of childhood sexual abuse may have interpersonal difficulties, and therefore
may struggle with aspects of intimate relationships. Possibly, the ability to consistently
demonstrate warm parenting may be a struggle for mothers with such a history. There
was not much reason to assume that sexually abused mothers would be more
authoritarian; in fact, if these mothers are less warm, according to Maccoby and
Martin‘s (1983) typology, they may be more permissive or disengaged as parents.
Thus, it was hypothesized that sexually abused mothers in this sample would
demonstrate and self-report less warmth and less control than their nonabused
counterparts.

19
CHAPTER 2: METHODS
Participants
Participants were part of an ongoing longitudinal study focused on child sexual
abuse and its psychobiological impact on development (Putnam & Trickett, 1997). The
original sample was comprised of 84 sexually abused girls and 82 comparison girls who
were matched demographically. Criteria for inclusion in the abused group included the
following: 1) the victim was a 6-16 year old female, 2) sexual abuse had to have
consisted of genital contact and/or penetration, 3) the perpetrator was a family member,
4) the abuse had to have been disclosed within six months prior to participation in the
study, and 5) a nonabusing parent/guardian was willing to participate in the study as
well.
13
Participants in the abused group were referred by protective service agencies in
the metropolitan Washington D.C. area. Comparison subjects were recruited by posting
requests for participants in facilities and newspapers of the same neighborhoods in
which the abused girls lived.
The sample for the current study was comprised of the non-abusing mothers
who were primary caregivers of these girls. Primary caregiver status was defined as
having lived with the target child most of the child‘s life including the two years prior to
the study; all biological mothers were assumed to be primary caregivers in this sample.
In cases where subjects of the longitudinal study were siblings, the mother‘s data was
used only once (n = 127).

13
Mother‘s live-in boyfriend was considered a family member in this study.

20
Approximately half of the mothers had a daughter who had experienced
confirmed familial sexual abuse; the other half were mothers of a demographically
similar comparison group. The subjects in this study ranged in age from 24-49, (mean
= 35.25, sd = 5.51) and had socio-economic status scores, as measured by the
Hollingshead, ranging from 11-66, (mean = 35.55, sd = 12.36), indicating, on average,
working-class-status.
14
Racially, the sample was made up of 52.3% mothers who were
white, and 47.7% mothers who were of color.
15

Overall, one-third of the sample reported sexual abuse history (33.9%, n = 39).
However, of the mothers whose daughters had been sexually abused, almost half
(49.2%, n = 29) reported some type of sexual abuse history, in contrast to the mothers
of the comparison group, in which only 16.4% (n = 9) reported a history of sexual
abuse. Of the thirty-nine mothers who reported a history of child sexual abuse, the
average age of the onset of abuse was 8.5 years (sd = 3.9). The vast majority of
perpetrators were family members (85%). Specifically, 41% percent of the mothers
reported that they had been abused by their fathers or father figures (i.e., stepfathers,
adoptive fathers, or foster fathers), followed by other relatives (29%; cousins, uncles, or
grandfathers), siblings (15%), and non-family members (15%) during childhood.
Measures
The measures used for this study fall into four main categories.

14
Hollingshead scores can range from 8-66, with the lowest number being the lowest in terms of
status.

15
This group is primarily African-American, with the exception of two Hispanic subjects and
one Asian subject.

21
1. Sexual Abuse Status. Because of known difficulty in obtaining reliability
and validity from retrospective reports of sexual abuse, questions regarding sexual
victimization were embedded in an interview in which subjects were asked about their
childhood, their relationships with their parents, and other personal history.
16
While
subjects were asked a series of questions about their knowledge of familial experiences
with sexual abuse, they were asked whether they had any such experiences; if they
indicated a sexual abuse history, careful probing was done to obtain details regarding
the age and duration of victimization, the relationship to the perpetrator, and the nature
of the abuse (Appendix A). Sexual abuse was defined as unwanted sexual contact with
someone older, before the age of 18. Individual case reviews were conducted in order
to determine the mothers‘ sexual abuse status. Due to the small subset of subjects who
indicated that they had been sexually abused (n = 39), characteristics of the abuse
experiences—such as severity, frequency, duration, and relationship to the
perpetrator—were not able to be used in analyses. Thus, following the lead of similar
studies, a history of child sexual abuse was dichotomized as ―yes‖ and ―no‖ (Alexander,
2000; Banyard, 1997; Cohen, 1995; DiLillo et al., 2000; Douglas, 2000).
2. Observed Childrearing Domains: Etch-A-Sketch Parenting Style Coding
System. The main purpose of the coding system, ―Etch-A-Sketch Parenting Style
Coding System,‖ was to behaviorally assess childrearing behaviors and overall style of
parenting. The coding system, which was an adapted version of a previously used

16
For a thorough discussion of methodological issues regarding sexual abuse see (Goldman &
Padayachi, 2000).

22
coding system developed by Trickett and colleagues, used both behavioral frequency
counts and global ratings in order to focus on parenting behaviors and the style of
interaction. Videotaped interactions between mother-daughter dyads were used to
obtain observational data on parenting. Interactions lasting 10-20 minutes featured the
pair engaging in a timed problem-solving task in which they were asked to work
together to draw a series of five increasingly difficult shapes or pictures on an Etch-A-
Sketch. Dyads were given three minutes to complete each of five tasks and were
instructed to each control one dial, and to avoid touching each other‘s dials. Verbal and
nonverbal parenting behaviors and styles were chosen to reflect the original Trickett
coding system, and were guided by the childrearing literature on salient dimensions of
parenting, particularly warmth and control. (The detailed coding manual with item-
level information can be found in Appendix D.)
3. Self-Report Parenting: Types of Discipline Used. Information on
subjects‘ disciplinary behaviors was obtained during a semistructured interview focused
on the developmental history of the child; this interview was conducted the first time
the subject was seen. The subjects were asked about the ―Types of Discipline Used.‖
Using a 5-point scale, subjects were instructed to select a number ranging from 1 for
―never‖ to 5 for ―often‖ indicating how often they used each of nine discipline
techniques, as can be seen in Appendix B.
4. Self-Report Parenting: Parenting Style. Information on subjects‘
―Parenting Style‖ was obtained in a series of eight questions asked during the same
semistructured interview (Appendix C); this interview was also conducted the first time

23
the subject was seen. The subjects were asked how they would describe themselves as
parents. The interviewer asked subjects to choose where they best fit on a five-point
scale, which was a spectrum of parenting characteristics.
Procedures
Data Collection
The childrearing and demographic data for this study was collected at either the
first or second time subjects were seen (Time 1 or Time 2 of data collection). These
sessions were typically a year apart and included a 3-hour protocol of semistructured
interviews and standardized measures regarding their own family of origin, details on
current family constellation and functioning, a developmental history of the daughter in
the study, and childrearing beliefs and behaviors, as well as a videotaped mother-
daughter interaction (Putnam & Trickett, 1997). Subjects‘ interviews occurred in a
separate room while their daughters were participating in the study. Subjects were
compensated monetarily at each time of data collection.
Procedures for Developing the Observational Coding System
Specific Aim 1: To develop an observational coding system focusing on childrearing
dimensions of warmth and control in both parental style and practices.
Obtaining the observational data used for this study required several iterative
phases in order to develop the coding system and extract meaningful constructs for
analyses. In a thorough review of existing measures of parental nurturance and
discipline, 27 standardized observational systems were compared (Locke & Prinz,
2002). Of these, only six were designed for the age range of the children in this study,

24
and of those six, only three had time requirements which were appropriate to the
existing videotaped interactions. However, these three applicable coding systems were
ruled out because of the specific constructs, the need for extensive training, or the
complexity of the coding system which prevented feasible usage within the scope of
this study. Thus, childrearing behaviors and styles were assessed by refining an
existing observational coding system.
17
The original Trickett coding system, which
was highly informed by childrearing theory, specifically by the dimensions of control
and warmth, was modified to be macroanalytic, event based, and focused only on the
mothers. The revised system also included a few new codes and two additional global
rating segments.
The coding system itself, as well as the implementation of the system, was
refined so that knowledge obtained from each stage of implementation informed other
stages of the system (Margolin et al., 1998). Using the ―feedback mechanism‖
framework that Margolin and colleagues present, which encourages using information
from one phase to influence other phases, the tasks of development of the coding
system, unitization, selecting and training coders, reliability checks, and data
transformation all guided each other.

17
The existing coding system was used by Dr. Penelope Trickett (2005) and colleagues to assess
parent-child interactions doing the same Etch-A-Sketch task for a NIH-funded study on physical abuse.
The system was microanalytic with sequential time-interval coding, where both parent and child
behaviors were coded.

25
In terms of specific tasks to be accomplished, two main steps were taken.
18
The
first task was refining the coding system and deciding on unitization. Codes were
chosen based on the parenting dimensions of interest for this study, namely warmth and
control. Many of the codes were chosen to echo items from self-report parenting
instruments that subjects completed. Coders rated subjects in three main components,
which included frequency counts of parenting behaviors for each of five tasks, global
ratings of mother‘s style of interaction for each of five tasks, and overall global ratings
for subjects‘ parenting style. Unitization of frequency counts for the parenting
behaviors of interest (for example, positive and negative character attribution, positive
and negative reinforcement, reasoning, attention directing, laughter, physical affection,
and aggression), were chosen in order to yield information on how often the mothers
demonstrated specific behaviors. A high frequency or an absence of behaviors (such as
instruction giving or negative character attribution of the child) was hypothesized to
provide more information than time-interval coding. Also, because the task became
increasingly difficult, it was anticipated that mothers‘ behaviors might change as
frustration was potentially increased; thus, both the frequency counts of behaviors and
the global style of interactions for each of the five tasks were coded separately for each
subject. However, as will be discussed in the preliminary analysis section, though all
five Etch-A-Sketch tasks were coded separately, they were ultimately summed to yield
more meaningful data. Then, overall global ratings which were reflective of the overall

18
These steps and protocol suggestions are adapted from two sources (Lindahl & Malik, 2000;
Margolin et al., 1998).

26
approach of the mother were documented after coders watched all five tasks to indicate
an overall impression of the mother‘s approach to parenting during the entire
interaction. Coders rated mothers on the following codes, which are explained in detail
in the coding manual found in Appendix D:
1. Parenting behaviors (frequency counts):

a. Verbal

i. Control/motivation

1. Positive character attribution of child
2. Negative character attribution of child
3. Positive reinforcement/approval
4. Negative reinforcement/disapproval
5. Explanations/reasoning
6. Prohibitions
7. Bargain/bribe
8. Threaten

ii. Teaching

1. Attention directing
2. Instruction/information giving
3. Instruction/information seeking
4. Other
5. Positive self/dyad attribution
6. Negative self/dyad attribution
7. Positive task attribution
8. Negative task attribution
9. Noninstructional task related conversation
10. Off-task conversation which doesn‘t interfere with task
11. Diverting/avoiding
12. Refusal/delay
13. Responding to child

b. Nonverbal

i. Control/Motivation

27
1. Restraining
2. Physical punishment/physical aggression
3. Nurturance/warmth

ii. Other

1. Diverting/off-task
2. Demonstrating annoyance
3. Laughter

2. Predominant style of task interaction. Considering both verbal and
nonverbal behavior of the mother, coders rated the affect/emotional tone that was most
dominant during each of the five tasks. Coders rated each dimension from very low to
high (1-5) for the following dimensions:
a. Light-hearted/approaches with humor/playful
b. Pleasant/happy/joyful
c. Apathetic/passive/withdrawn
d. Anxious/afraid
e. Irritated/angry/hostile
f. Who was leading the task (mother—child)

3. Overall parenting style. After watching all five tasks, using macro-
analytic descriptors, coders scored each mother based on the overall impression of the
mother‘s parenting style throughout all tasks. Most of the items were selected because
they were similar to items on which mothers rated themselves. Using a 5-point scale to
indicate the style of relating, coders rated mothers on the following parental styles:
a. Authority figure--friend
b. Lenient--strict
c. Irritable--easy-going
d. Inconsistent--consistent
e. Affectionate--unaffectionate
f. Unfair--fair
g. Serious--playful
h. Enjoys child--doesn‘t enjoy child

28
i. Encourages autonomy--doesn‘t encourage autonomy
j. Reasons/explains--makes rules/gives commands
k. Warm toward child--withholding/chilly
l. Shy/timid--outgoing/not shy

Procedures for Implementing the Observational Coding System
After the first step of refining the coding system and deciding on unitization, the
second step in the process of developing the coding system was to select and train
coders. A description of the coding project was posted on a database intended for
psychology students who were interested in volunteering or taking course credit to
assist with research projects. After a selection process based on availability,
commitment to the project, experience, and interest, which was determined via email
and then by phone interviews, four undergraduate psychology students were selected as
coders. Coders were asked to commit to at least one semester of training, meetings, and
completing coding their assigned subjects.
Each subject‘s videotaped interaction was coded by two different raters;
therefore, each of the four coders was assigned 48 subjects, with two coders rating one
additional dyad (total videotaped subjects N = 97). Subjects were assigned carefully to
pair each coder with the other three coders the same number of times; measures were
taken to ensure that each combination of coders did not receive a disproportionate
amount of subjects with an abuse history. On two occasions, subject numbers had to be
reassigned to a different coder because only one coder spoke Spanish and two dyads
communicated primarily in Spanish. These steps were taken in order to achieve optimal
reliability and reduce systematic error.

29
The coder training protocol involved the following: (a) coders thoroughly read
the coding manual that contained full descriptions of the coding system and definitions
of all codes; (b) coders began their training with a thorough review of the distinctions
between codes in a group setting; (c) in a group setting, using videotapes of interactions
which were not eligible for this study,
19
coders individually coded the training tapes,
presented their results, and then discussed applicable distinctions and discrepancies in
coding among themselves; (d) aspects of the coding system and several of the actual
codes were then refined based on feedback from the above steps, which included adding
codes and procedures, deleting codes, and collapsing codes, which then required a
repeat of steps a-c in a group setting.
When coders were thought to be sufficiently trained, all four coders were given
DVDs that contained the Etch-a-Sketch interactions of five randomly selected subjects
who had not been previously seen and who would not be used for analyses. Coders
were asked to take home the DVDs and independently watch and carefully code all five
tasks of each of the five subjects and complete the global ratings for them as well.
Reliability analyses were then conducted to assess all four coders‘ reliability to ensure
that coders were ready to proceed with independent coding.
20
Coders were further

19
Though there were not any training tapes available for this data set, there were several subjects
who completed the activity whose data could not be used in analyses because the adult in the interactions
was not the primary caregiver. Thus, the videotaped interactions of subjects who did not meet criteria for
these analyses were used for piloting the coding system and then to train coders.

20
Intra-class correlations (ICC), using a two-way mixed model for average measures of
reliability were conducted on each item across all 4 coders (Gordis, 2005; Mc Graw & Wong, 1996).
With the exception of several items which yielded an ICC of zero or below because the behaviors were

30
trained by individual coding in a group setting on all items that did not reach acceptable
levels until they reached an inter-rater reliability of .70.
21
Once acceptable reliability
was reached, ongoing meetings and trainings continued throughout the coding process
to ensure high reliability and to prevent coder drift. Coders completed coding the
videotaped interactions over a period of four months.

not seen in these pilot DVDs, most items reached acceptable levels of inter-rater reliability. For instance,
for the behavioral ratings which used frequency counts, ICC ranged from .206 to .980. Of 26 behavioral
items, 10 items had ICC of .677 and above, 12 items with no variability which had ICC scores of zero or
below (ten of these items were dropped because of this distribution problem), and four items which
ranged from .206 to .370. With the exception of the zero distribution items, the average ICC was .691.

21
For the 18 global items, two items had ICC of zero, which were ultimately dropped in later
analyses due to distribution problems. With the exception of the two items with a zero score, the mean
ICC of these items was .776.

31
CHAPTER 3: RESULTS
Analyses included a variety of statistical procedures to examine the data
obtained from the observational coding system and the self-report measures in order to
examine possible influences of sexual abuse history on childrearing approaches. The
aims of the study at hand were to explore and extract meaningful psychometric
constructs from the observational data and the self-report instruments, and then to
investigate whether there were significant relationships among sexual abuse history and
both observed and self-reported childrearing.
General Preliminary Analyses
Data Preparation and Missing Data
The highest priority was to locate missing items through an examination of hard-
copy data. When this was not possible, missing data was handled in a variety of ways,
including imputation, data substitution from another time point, and as a last resort,
exclusion of cases. In cases when entering the actual data from the hard-copy or a
duplicated but more complete database was not possible, missing data was imputed at
both the first-order (when one or two items are missing in a scale and the other items in
the scale can inform a probable score for the missing items), and at the second-order
(where other highly correlated items from other measures can be used to impute items
and total scores), when appropriate. Using a multiple imputation software program
called NORM, interval-level data was imputed using the EM algorithm which generates
start values, followed by multiple imputations of simulations of the missing data
(Darmawan, 2002; Rubin, 1987).

32
In cases where imputation was not appropriate, if the data was collected at
multiple time points, and the variable was considered a stable variable, substitution
from the next closest time point was used. For instance, for three cases in which socio-
economic status (SES) was missing, the Hollingshead SES score from the next closest
time point (on average 1-2 years) was substituted. As a last resort, cases that were
missing were excluded from analyses. Using Pearson correlations, chi-square, and t-
tests, as appropriate, no differences in terms of subjects with missing data and subjects
with complete data were detected.
22

Observational Childrearing Data:
Data Reduction and Measurement Development

Specific Aim 2: To reduce the observational data into meaningful parenting dimensions
focused on warmth and control through exploratory measurement development and
psychometric analysis.
This phase of analysis was aimed at taking the multiple measures related to
childrearing and employing a variety of statistical approaches including descriptive
analysis, factor analysis, and reliability analysis in order to take multiple items and
aggregate them to form separate reliable indicators of parenting along the dimensions of
warmth and control. This process resulted in more than just a measurement model, but
yielded actual findings that were useful and informative beyond just the preparation for

22
Using a chi-square analysis, there were no significant differences based on race in terms of
missing data (X
2
(1, N = 127) =.989, p = .348).

33
hypothesis testing, and can be used for further development and application of the
coding system in future studies.
Item-level analysis
Of 127 subjects, Etch-a-Sketch videotaped data was available on 84, but in order
to increase the number of cases for the development of the measurement model, all
dyadic interactions with primary mothers were included in the coding and psychometric
analyses. By including 13 videotaped interactions of mothers who completed the Etch-
a-Sketch with more than one of their daughters, the number of cases included totaled
97. The data obtained from these 13 duplicated subjects was only included in
psychometric testing of the coding system to extract meaningful composite factors. No
other analyses included the same mother twice. A t-test was conducted to compare
demographic characteristics of subjects who were missing the videotaped interaction to
those who were not. Analyses indicated no significant difference between the two
groups in terms of race (F = .001, p = .972), age (F = 1.787, p = .184), socioeconomic
status (F = .082, p = .774), or sexual abuse status (F = 1.166, p = .282).
After coders finished rating all videotaped subjects, descriptive analyses were
employed to examine missingness, variance, distribution, and skewness of each item
within each task rated in the coding system. Of the five tasks in which coders rated the
subjects, over 30% of the subjects were missing the first task, often referred to on the
videotapes as a ―practice task‖ by the task administrator, and there were six cases where
task 5 was missing for the dyad. A further complication was that on most of the 26
items being observed by frequency count, some of the behaviors were rarely

34
demonstrated, and thus, the separation of the coding by tasks proved to be unhelpful.
Thus, task 1 was dropped except in cases where the subject was missing another task, in
which case, task 1 frequency counts were substituted for the missing task. Then, each
item was summed across tasks 2 through 5 for each subject. As a result, each of the two
coders had a total score for each behavioral item for each subject as well as global
scores. Both the total behavioral frequency count of each item and the global scores
were averaged among both coders to yield one score per item for each subject, which
was used for item analysis (see Appendix E).
Of the 26 behavioral frequency count items, ten items were dropped from
further analyses based on variance, skewness, and distribution (Smith & McCarthy,
1995). More specifically, when examining a behavioral item, if 95% of cases
documented the behavior one or zero times (the skewness was also extreme at 4.0 or
greater), then the item was dropped from analysis (see Appendix E). Two global items
were dropped from analysis for similar reasons. For global ratings on the mothers‘ style
of interaction, 99% of subjects received a score of 1 on a 5-point scale, indicating very
little or no demonstration of ―apathetic‖ and ―anxious‖ interactions with their daughters;
thus, these items were dropped from subsequent analyses.
Behavioral Frequency: Factor Analysis and Reliability Analysis

While it is common practice to determine the number of factors based on
Kaiser‘s (1960) rule in which all factors with an eigenvalue greater than 1.0 are
retained, some argue that this common criterion often inaccurately results in an
overestimated number of factors (Cliff, 1988; Franklin, Gibson, Robertson, Pohlmann,

35
& Fralish, 1995; Hayton, Allen, & Scarpello, 2004; Horn, 1965; Kaiser, 1960;
Lautenschlager, 1989). Parallel analysis, developed by Horn (1965), is a more accurate
and conservative alternative for estimating the number of factors by using a more
stringent criterion. The premise of parallel analysis is that meaningful factors in the
data being tested should have higher eigenvalues than would be obtained from sampling
variability in random data with the same sample size and number of variables (Franklin
et al., 1995; Horn, 1965; Lautenschlager, 1989). Lautenschlager (1989) simulated
multiple datasets and used the mean of the eigenvalues obtained from random data for
numerous combinations of variable and sample sizes, and created charts from which
actual data eigenvalues can be compared; each factor eigenvalue obtained in actual
factor analysis can then be compared to the random data eigenvalues, such that any
eigenvalues of the factors from a dataset which fall below the random factor‘s
eigenvalue should be considered spurious (Franklin et al., 1995; Lautenschlager, 1989).
This simple comparison allows researchers to more accurately know how many factors
to limit in each analysis.
Separate exploratory factor analyses were conducted for the behavioral
frequency items and the global rating items due to their differing scales and conceptual
approaches to measurement. Thus, a principal component analysis on the remaining 16
behavioral frequency count parenting items was conducted using Varimax rotation
(Clark & Watson, 1995; Comrey, 1988; Floyd & Widaman, 1995; Gorsuch, 1983;
Stumpf & Parker, 2000). This first analysis yielded five factors with eigenvalues over
1.0; then, parallel analysis was conducted by using the approach detailed by

36
Lautenschlager (1989), which indicated limiting the number of factors to three.
Principal component factor analysis was conducted again using a Varimax rotation with
a restricted three factor structure (eigenvalues of 3.48, 2.69, 1.59). As detailed in Table
3, Factor one accounted for 21.76% of variance, with factor loadings ranging from .648
to .786; factor two explained 16.82% of the variance, with factor loadings that ranged
from .365 to .683; and factor three explained .10% of the variance, with factor loadings
of .473 to .851.
Using the items extracted within each factor, a series of reliability analyses were
then conducted, testing all items that loaded the highest within each factor.
Additionally, any items that had factor loadings of .3 or greater were included in
reliability analyses to assure the inclusion of all useful items in constructing the most
psychometrically and conceptually meaningful composite variables. To make the most
reliable scale, the statistical tool within reliability analysis, which shows Cronbach‘s
Alpha if an item is deleted, was used. As a general rule, if Cronbach‘s Alpha would
only slightly increase with the removal of the item, the item was retained in order to use
more available data. This procedure was followed for each of the three factors.

37
Table 3. Rotated Component Factor Loadings for Behavioral Items
Behavioral Frequency Items Factor 1 Factor 2 Factor 3
Negative reinforcement .786 -.096 .148
Ignoring/annoyance .755 -.082 -.186
Prohibitions .733 -.213 .341
Instruction/information giving .719 .028 .070
Negative self-dyad attribution .648 .274 .102
Laughter -.085 .683 .028
Noninstructional task-related conversation .441 .636 -.302
Nurturance/warmth -.118 .617 -.288
Responding to child -.069 .616 .059
Information seeking/questioning .166 .539 .056
Negative task attribution .334 .473 .017
Positive task attribution -.135 .365 .186
Attention directing .161 -.166 -.146
Explanations/reasoning .141 -.158 .851
Positive reinforcement .424 .155 .634
Positive self/dyad attribution -.103 .473 .473
% of Variance Explained 21.760 16.820 9.97
Eigenvalue 3.482 2.692 1.595

For factor one, the most reliable aggregate was created with four items,
including Negative Self/Dyad Attribution, Negative Reinforcement, Ignoring/
Annoyance, and Negative Task Attribution, which conceptually was termed ―ES
Behavior: Negative Displeasure‖ (Cronbach‘s alpha = .702, ICC = .724). Factor 2,
ultimately named ―ES Behavior: Warm Responsiveness‖ was created with the
following items: Laughter, Noninstructional Task-Related Conversation,
Nurturance/Warmth, Responding to Child, Information Seeking/Questioning, and
Positive Self/Dyad Attribution (Cronbach‘s Alpha = .657, ICC = .599). Finally, factor
three was dropped due to the number of items and low reliability (Cronbach‘s Alpha =
.409).

38
Global Ratings: Factor Analysis and Reliability Analysis
Once again conducting a principal component factor analysis, four factors were
found. The principal component analysis was conducted again with a limited two factor
structure, as suggested by Lautenschlager‘s (1989) parallel analysis approach, with a
Varimax rotation. As can be seen in Table 4, the first factor explained 39.73% of the
variance, with factor loadings ranging from .365 to .872 (eigenvalue = 6.37), and the
second factor explained 14.26% of the variance, with factor loadings ranging from .650
to .831 (eigenvalue = 2.28).
As with the observed behavioral data, in order to combine the most conceptually
and psychometrically meaningful variables from this global observational data,
consideration of the above factor structure, the childrearing literature, and many
reliability analyses determined which items to aggregate into two composite variables.
Reliability analyses were conducted based on the above factor loadings and included
any items that loaded above a .3, regardless of whether they loaded the highest on a
particular factor. Again, the statistical tool within reliability analysis, which shows
Cronbach‘s Alpha if an item is deleted, was used to determine the most reliable item
combinations that were suggested by factor analysis. The first factor, called ―ES
Global: High Control/Authoritarian,‖ ultimately included the following items: Mother
Leads Task, Authoritarian, Strict, Restricts Child‘s Autonomy, Irritable, Gives
Commands, and Unfair (Cronbach‘s Alpha = .901, ICC = .721). The second factor, ―ES
Global: High Warmth/Responsiveness‖ included the items Playful, Warm,
Affectionate, Light-Hearted/Humor, Happy (Cronbach‘s Alpha = .805, ICC = .550).

39

Table 4. Rotated Component Factor Loadings for Global Items
Global Observational Items Factor 1 Factor 2
Irritable style .872 .183
Strict .825 .294
Irritable during task .791 -.116
Authority figure .758 .437
Unfair .755 .168
Inconsistent .720 -.056
Makes rules without reasoning .550 .294
Warm .452 .049
Mother lead task .391 .106
Restricting child‘s autonomy .387 -.229
Affectionate .365 -.001
Light-hearted/humorous .315 .831
Happy .488 .762
Outgoing -.234 .715
Playful .599 .706
Enjoys child .066 .650
% of variance explained 39.730 14.260
Eigenvalue 6.370 2.280

Self-Report Childrearing Data:
Data Reduction and Measurement Development

Specific Aim 3: To reduce self-report data into meaningful parenting dimensions
focused on warmth and control through exploratory measurement development and
psychometric analysis, and to examine the relationships between observational and self-
report data in this sample.
Item analysis
As detailed in the Measurement section, the self-report childrearing data used
for these analyses came from a semi-structured interview in which mothers were asked
to rate themselves on a 5-point scale. One set of questions focused on the frequency of

40
use of nine discipline approaches and the other set of eight questions focused on
parenting style. Of 127 subjects, data was available for 122 subjects for the discipline
items and for 121 subjects on the parenting style items. Basic descriptive analyses
confirmed that this data was well distributed, cleaned, and ready for further analyses.
All items were retained for factor analysis.
Self-report Discipline Items: Factor Analysis and Reliability Analysis
Similar to the preliminary analyses on the observational data, the goal was to
reduce the self-report measures on childrearing dimensions by constructing sound
composite variables of salient parenting dimensions. Thus, separate exploratory factor
and reliability analyses were conducted on the discipline items and the parenting style
items due to their radically different constructs.
A principal component factor analysis was conducted using Varimax rotation on
the nine items focused on discipline practices (Clark & Watson, 1995; Comrey, 1988;
Floyd & Widaman, 1995; Gorsuch, 1983; Stumpf & Parker, 2000). The first analysis
yielded four factors with eigenvalues over 1.0, ranging from 1.03 to 2.175. Then,
parallel analysis was conducted using the Lautenschlager‘s (1989) random eigenvalue
charts, which indicated that the factor analysis should be limited to two factors.
Principal component factor analysis was conducted again using a Varimax rotation with
a restricted two factor structure (eigenvalues of 2.175, 1.36). As detailed in Table 5,
Factor one accounted for 24.17% of variance, with factor loadings ranging from .536 to
.687; factor two explained 15.15% of the variance, with factor loadings that ranged
from .330 to .780.

41
Taking into account the results of these factor analysis and conceptual
guidelines from the discipline literature, numerous reliability analyses were conducted.
Because there were no items that had factor loadings above .3 that did not load onto one
of the two factors, only items that loaded on each factor were entered into
Table 5. Rotated Component Factor Loadings for Discipline Items
Behavioral frequency items Factor 1 Factor 2
Use of guilt/silent treatment .678 .237
Use of verbal punishment .650 .188
Use of threatening punishment .602 .050
Use of explanation/teaching -.569 .233
Use of ignoring bad behavior .536 .140
Use of isolation .059 .780
Use of taking away privileges -.071 .774
Use of physical punishment .240 .340
Use of prohibitions .091 .330
% of variance explained 24.167 15.154
Eigenvalue 2.175 1.364

reliability analysis. Entering all five items that loaded on the first factor yielded a
Cronbach‘s alpha of .604; the analysis indicated that the alpha would not improve with
the removal of any additional items. Thus, these five items were combined to create a
conceptually appropriate variable, which was named ―SR: Punitive Control/Negative
Structure.‖ The second factor, consisting of four items, yielded a Cronbach‘s alpha of
.431 and was dropped due to low reliability. The best alpha could have been achieved
by dropping two items; however, the highest the alpha achieved would be .450 and it
would consist of only two items.

42
Self-report Global Parenting Style Items: Factor Analysis and Reliability Analysis
As with the discipline items, a principal component factor analysis was
conducted using Varimax rotation on the eight items focused on parenting styles. The
first analysis yielded two factors (eigenvalues 2.30, 1.492). Parallel analysis indicated
that the factor analysis should be limited to two factors, thus further factor analyses
were not needed. As detailed in Table 6, factor one accounted for 28.755% of variance,
with factor loadings ranging from .453 to .800; factor two explained 18.65% of the
variance, with factor loadings that ranged from .511 to .713.
Table 6. Rotated Component Factor Loadings for Parenting Style Items
Behavioral frequency items Factor 1 Factor 2
Easy-going (not irritable) .800 .192
Consistent .733 -.073
Fair .707 -.006
Playful .453 .411
Lenient -.031 .713
Friend (not authority figure) .017 .688
Affectionate .190 .646
Reasons (not making rules/giving commands) .018 .511
% of variance explained 28.755 18.650
Eigenvalue 2.300 1.492

Combining the structure of these factors and conceptual guidelines from the
parenting literature, a series of reliability analyses were executed. The four items that
loaded on the first factor were entered into reliability analysis, where Cronbach‘s alpha
indicated that dropping the ―Playful‖ item would increase the alpha of the scale from
.641 to .642. While the deletion of this item did not really improve the alpha, to be
more conservative, this item was dropped because this item loaded almost equally onto

43
both factors and the item was not well distributed. The resulting 3-item composite was
termed ―SR: Positive Structure/Control‖. The four items that loaded on the second
factor could not be improved by removing any items, and the four items were combined
to create a ―SR: Warmth‖ composite scale (Cronbach‘s alpha = .557).
Relationships between Observational and Self-report Data
Using both observational and self-report data provided a fuller picture of
parenting styles and behaviors. Because it is uncertain which approach is better for
different aspects of parenting, and because observational data on this population is
almost nonexistent, it was helpful to include both approaches to examine warmth and
control in this study. Though it is not unusual for observational and self-report data to
be poorly associated, it was still surprising that these two approaches measuring similar
constructs would be so lowly correlated with each other. In this sample, a brief glance
at Table 7 reveals that (a) the observational variables are highly correlated with other
observational variables, (b) the self-report variables are rarely correlated with other self-
report variables, but (c) with one exception, the observational and self-report variables
are not correlated with each other.
A more in depth examination indicated that with the exception of the ES
Behavior: Negative Displeasure variable, (which was only significantly correlated with
one variable--ES Global: High Control, r= .48), all of the other observational variables
were significantly correlated with each other. Whereas, only one significant correlation
was found among the three self-reported variables, which was between SR: Punitive
Control/negative structure and SR: Positive Control/structure (r = -.26). Overall, the

44
only significant relationship between the observed and self-report data was between ES
Global: Warmth and SR: Positive Control/structure (r = .33). Thus, even though the
dimensions of warmth and control were measured by observational and self-reported
approaches, these dimensions were rarely significantly associated with one another. It
was also interesting that child sexual abuse was significantly correlated with observed
warmth (ES Behavior: Warm Responsiveness, r = .36, and ES Global: Warmth, r = .35),
but not with self-reported warmth (SR: Warmth, r = .04). In terms of relationships
between child sexual abuse and the control dimension variables, only one significant
correlation was found—a negative association with self-reported positive structure and
control (SR: Positive Structure/Control, r = -.23)

45
Table 7. Inter-correlations of Variables

1 2 3 4 5 6 7 8 9 10 11
1. Sexual Abuse Status
2. ES Behavior: Negative
displeasure
.15
3. ES Behavior: Warm
Responsiveness
.36** .14
4. ES Global: High
control/authoritarian
-.20 .48** -.34**
5. ES Global:
High warmth
.35** -.12 .42** -.64**
6. SR: Pun. Control/
negative structure
-.08 -.01 -.04 .18 -.08
7. SR: Positive
control/structure
-.23* .00 -.11 .17 -.33** -.26**
8. SR: Warmth

.04 .01 .07 -.01 .05 .01 .16
9. Age

.11 -.19 .09 -.42** .15 -.13 -.01 .01
10. Race

-.29** -.03 -.23* .26* -.38** -.01 .27** .06 -.08
11. Socioeconomic status

.00 -.11 .08 -.15 .02 -.18 .05 -.18 .30** -.04

Note: Sexual Abuse Status: ―1‖ if sexually abused, ―0‖ if no history of sexual abuse. Race: ―1‖ if African American, Hispanic, or Asian,
―0‖if Caucasian.
ES = Etch-a-Sketch Observational Task, SR = Self-Report
* Correlation significant at p < .05
** Correlation significant at p < .01
46

Summary
Based on the above preliminary analyses on both the observational and self-
report data, the following composite variables were extracted: two observed parenting
behavioral frequency composites (ES: Negative Displeasure and ES: Warm
Responsiveness), two observed global parenting style composites (ES: High
Control/Authoritarian and ES: High Warmth), and three self-report parenting
approaches (SR: Punitive Control/Negative Structure, SR: Positive Control, and SR:
Warmth). The observed behavioral frequencies, the observed global ratings, and the
self-report data yielded a total of seven meaningful variables; note that these three
differing approaches of measuring parenting each produced a negative/controlling type
of variable and each produced a warmth type of variable. Some of the inter-correlations
of these distinct variables, which approach measurement differently, are related to one
another, but still offer their own unique contribution in measuring parenting, (Table 7).
Demographic Analyses
As the childrearing literature continually indicates, demographic characteristics
are often significantly related to many parenting constructs. Therefore, preliminary
analyses were conducted to determine whether race, age, and socioeconomic status
needed to be controlled in later analyses. Race was significantly correlated with the
following observed composite variables: ES Behavior: Warm Responsiveness
(r = -.234), ES Global: Warmth (r = -.381), and ES Global: Authoritarian control
(r = .264). Age was significantly related to authoritarian control (r = -.419), and
Hollingshead scores were not significantly correlated with any of the outcome variables
47

for this study. Because race, age, and socioeconomic status are frequently related to
parenting variables, these demographic variables were considered carefully. Even
though each of these demographic variables was not significantly related to every
outcome variable, as a conservative attempt to control for any explanative power, all
three demographic variables were included as part of the regression models in order to
parcel out their contribution to explaining differences in the outcome variables.
In this sample, non-Caucasian mothers were observed to be more authoritarian
and less warm than Caucasian mothers. Additionally, older subjects were observed to
be less controlling of their daughters. While socioeconomic status (SES) was not
significantly related to any outcome variables, because of the common relation between
SES and parenting, and because subjects had such a large range of SES, to be
conservative, it was controlled in analyses as well. Thus, hierarchical regression
analyses were chosen in order to remove the effects of potential salient demographics in
order to more accurately see the effect of sexual abuse history upon childrearing
(Cohen, Cohen, Aiken, & West, 2002). Before beginning regression analyses, Pearson
correlations were conducted on the independent variables—child sexual abuse, the
seven childrearing outcome variables, and the relevant demographic variables (Table
7).
23

23
Pearson correlations were also conducted by sexual abuse status, and were almost identical
and not significantly different from the correlations of the combined groups listed in Table 7.
48

Statistical Analyses of Composites
Specific Aim 4: To examine whether sexual abuse history is related to childrearing
dimensions of warmth and control, using both observational and self-report data.
Sexual Abuse History and Childrearing
Guided by the findings of the above three steps, previous research, theory, and
the proposed research objectives, hierarchical multiple regression analysis was
conducted to test the fourth specific aim of this study. Using hierarchical multiple
regression for each of the seven childrearing dependent variables, control variables of
race, age, and SES were entered as a block in the first step. At step 2, sexual abuse
history was added into the model in order to test the relationships of sexual abuse
history with a series of childrearing variables (see Table 9 for means, standard
deviations, and ranges of all seven variables by group status). Results, as displayed in
Table 8, indicate that sexual abuse status is significantly associated with variables in
both observed and self-reported parenting. In terms of observed parenting, mothers
with a history of sexual abuse demonstrated significantly more warmth in both their
behavior (df = 1, p = .005), and in their global parenting style (df = 1, p = .018) than
mothers with no such history. However, a sexual abuse history did not significantly
differentiate between mothers in terms of either global authoritarian/control (df = 1, p =
.347), nor in terms of negative disapproving behavior (df = 1, p = .130).

49

Table 8. Standardized Beta Coefficients of Sexual Abuse History with Observed and Self-Reported Childrearing
Behaviors and Global Parenting Style

Demo-
graphic
variables
ES:
Negative
displeasure
behavior

N = 83
ES:
Warm
responsive
behavior

N = 83
ES:
High
control
global

N = 83
ES:
High
warmth
global

N = 83
SR:
Punitive
control or
negative
structure
N = 105
SR:
Positive
control or
structure

N = 103
SR:
Warmth

N = 115
Race .003 -.138 .202* -.299** -.030 .271** .059
Age -.197 .018 -.381** .098 -.062 .040 .096
SES -.053 .071 -.030 -.018 -.231* -.020 -.117
Sexual
abuse
.178 .318** -.099 .258** -.057 -.231** .085
Adjusted
R
2

.022 .110** .198 .178** .030 .127** -.012

*p < or = .05, **p < or = .01

Similar parenting dimensions obtained from self-report data indicated that
positive structure/control was significantly related to having a history of sexual abuse,
with sexually abused mothers reporting less fairness, less consistency, and more
irritability (df = 4, p = .019). However, self-reported punitive control (df = 1, p = .580),
and self-reported warmth (df = 1, p = .396) were not significantly associated with a
history of sexual abuse.
Demographic variables were often significant contributors within the tested
models; therefore, beta coefficients in Table 8 indicate the adjusted R
2
after adding
sexual abuse status into the model, in order to clearly see its contribution to outcome
variables, parceling out the influence of the demographic variables. In sum, three of the
seven childrearing outcome variables were significantly associated with a history of
child sexual abuse, with mothers who had been sexually abused in childhood
50

demonstrating more observed warmth (from both behavioral frequencies and global
ratings), and less self-reported use of positive control or structure with their daughters.
Table 9. Sample Means, Standard Deviations, and Ranges of Childrearing Variables

Mean
_____________________
Standard Deviation
_____________________
Range
_____________________
Childrearing Variables Total CSA
d
Comp
arison
Total CSA Comp
arison
Total CSA Comp
arison
ESB: Negative Displeasure
a
5.9 7.1 5.2 6.0 6.8 5.5 0-35 0-35 0-24
ESB: Warm Responsiveness 36.3 45.1 31.5 18.1 19.9 15.2 6-105 16-105 6-66
ESG: Control/Authoritarian
b
18.3 17.1 18.9 4.5 4.2 4.6 11-31 11-29 11-31
ESG: Warmth 33.9 37.6 31.9 7.8 6.9 7.5 16-49 26-49 16-48
SR: Punitive Control/ Structure
c
14.0 13.7 14.2 2.8 2.7 2.9 8-22 8-19 9-22
SR: Positive Control/Structure 10.9 10.2 11.2 1.9 2.0 1.7 5-15 5-14 7-15
SR: Warmth

13.7 13.8 13.7 2.1 2.3 2.0 8-18 9-18 8-18

a
ESB= Etch-a-Sketch Behavior,
b
ESG= Etch-a-Sketch Global,
c
SR= Self-report,
d
CSA= Child Sexual Abuse Group

51

CHAPTER 4: DISCUSSION
The goals of this study were a) to develop a coding system for observational
data obtained during a mother-daughter interactive task, based on the parental domains
of warmth and control; b) to extract meaningful constructs from both observational and
self-report data, and to examine the correspondence between them; and c) to examine
the impact of childhood sexual abuse on aspects of parental warmth and control.
Results from the developed observational coding system and the self-report data
included both the confirmatory and contradictory. Overall, in terms of warmth, findings
indicated that mothers with a history of sexual abuse were not different from non-
abused mothers in the amount of self-reported warmth, but were observed to be
significantly warmer towards their daughters. In terms of the control dimension,
mothers with a history of child sexual abuse reported difficulty providing positive
structure/control in their parenting.
Given the lack of reliable studies on the impact of child sexual abuse on parental
warmth and control, and the inconsistencies in them, this study is a contribution to the
maltreatment and childrearing fields in that it examined the relationships between child
sexual abuse and the dimensions of parental warmth and control using a community
sample with a comparison group. An additional strength of this study was the
development of an observational coding system which made it possible to include both
observational and self-report data. Further, this study included the examination of the
correspondence between these approaches. This study has the potential to begin to shed
light on the conflicting body of research on the impact of child sexual abuse on parental
52

warmth and control, and to begin the process of building a credible body of empirical
research on these mothers. These findings may be useful in the further development of
observational approaches in measuring parenting, and may help with the development
of a conceptual model to guide future research on the intergenerational transmission of
sexual victimization. In order to further understand the impact of child sexual abuse on
parenting, and ultimately to understand the mechanisms of the intergenerational
vulnerability to victimization, additional rigorous studies need to be conducted that
conceptually and empirically separate out the effects of abusive experiences, different
aspects of parenting, and the larger context of the parent-child relationship, as well as
the family system at large.
The first and second aims of the study were to develop an observational coding
system, and then reduce the observational data into meaningful parenting dimensions
focused on warmth and control. Principal component analyses along with reliability
analyses ultimately resulted in four observational composites. While the global items
yielded two composites along the dimensions of interest—one warmth variable and one
control variable—the items from the behavioral frequencies did not as neatly fit these
dimensions. The behavioral frequency items yielded a warmth variable, but not a
control variable. However, the other factor found in the behavioral items, ES Behavior:
Negative Displeasure, was highly correlated with the global high control variable (r =
.48). After careful data reduction, the observational coding system provided useful
information about important differences in the sample that the self-report data alone
would not have revealed.
53

The third aim was to reduce self-report data into meaningful parenting
dimensions focused on warmth and control, and to examine the relationships between
observational and self-report data in this sample. The self-report data produced three
factors—two that reflected different aspects of the control dimension, and one warmth
factor. Even if it can be assumed that observed interactions and self-reported
perceptions are fairly accurate, and that coders were able to document those behaviors
accurately, item and factor analyses of the data was as much art as it was science.
While previous studies, theory, and statistical conventions guided these analyses,
several different item-level variables could have been combined to reach different
composite variables, leading to potentially different results.
In this study, the observational variables tended to be related to other
observational variables, but this was not as much the case for the self-reported
variables. While self-reported warmth did not correlate significantly with any other
variables, the two observed warmth variables were associated with one another, and
with many other of the variables in this study. That the observed and self-reported
constructs which were similar to each other were not significantly correlated, reveals
that each type of data offered a unique contribution to understanding this sample. In
fact, the observational data only revealed group differences in warmth, and the self-
report data only revealed group differences in control, despite the fact that both
measurement approaches had both dimensions.
One thing to consider is that from a multitrait-multimethod standpoint, measures
of the same trait should correlate higher with each other than they do with measures of
54

different traits involving separate methods (Bagozzi, Phillips, & Yi, 1991; Campbell &
Fiske, 1959). However, for both the warmth and control variables, the observational
variables are highly correlated with each other, as are the self-report variables,
regardless of whether they are measuring warmth or control. This lack of correlation
among the warmth variables and among the control variables is problematic. It is
uncertain whether these issues of variance are due to the measurement method or to the
constructs of warmth and control, and it is uncertain which approach captures the most
reliable and valid information. Thus, conducting convergent and discriminant
validation using the multitrait-multimethod matrix approach of Campbell and Fiske
(1959), or conducting confirmatory factor analyses to examine the trait and method
variance among the dimensions of parental warmth and control, would be helpful for
further understanding of these parenting dimensions and this coding system.
The fourth aim of the study was to examine whether sexual abuse history was
related to the childrearing dimensions of warmth and control, using both observational
and self-report data. Previous research indicated that mothers with a history of sexual
abuse may have more difficulty with certain aspects of parenting than non-abused
mothers. Thus, the last step of this study was to take the seven factors harvested from
the observational and self-report data to examine the impact of child sexual abuse on
different aspects of childrearing. It was hypothesized that sexually abused mothers in
this sample would rank significantly lower in terms of both warmth and control. In
terms of control, then, it was not surprising that the mothers in this sample with a sexual
abuse history reported themselves as struggling with appropriate amounts of positive
55

structure and control. Sexually abused mothers in this study indicated that they
perceived themselves to be less fair, less consistent, and more irritable with their
children than the nonsexually abused mothers. This finding is consistent with some
literature, which suggests that sexually abused mothers tend to be more permissive and
have difficulty setting limits. However, given the conflicting empirical research on this
topic, it was uncertain whether the mothers in this sample would be more controlling,
less controlling, or not differ from the nonsexually abused mothers.
While sexually abused mothers in this sample had difficulty with the control
dimension, they did not seem to struggle with the warmth/responsiveness dimension. In
fact, it was surprising, and counter to what was hypothesized, that mothers with a
history of sexual abuse were significantly warmer toward their daughters than the non-
abused mothers in the observed interactions. It is interesting to note that ―warmth" was
not significantly differentiated by abuse status in the self-report measures. There are
several possible explanations for the above findings.
Because self-reported warmth did not discriminate between the groups, but
observed warmth did, it is important to consider the nature of observational data.
Specifically, one idea to consider is that when they were being videotaped, the mothers
who had a sexual abuse history were not as confident in their natural parenting and were
intentionally warmer and more affectionate towards their daughters, knowing that they
were being observed. Further, mothers may have been more comfortable being warm
towards their daughters, but not as comfortable demonstrating control. Or it may be
56

that the task itself didn‘t warrant the need for parental control, thus it was not
demonstrated.
An additional explanation for the group differences in terms of observed
maternal warmth is the very nature of being in this study, where many developmental
questions were asked about the daughters, including asking the mothers to reflect back
to pregnancy and early infancy. This experience may have had an intervening effect
upon the amount of warmth mothers with differing backgrounds, such as a history of
child sexual abuse, may have demonstrated. It must also be considered that the self-
reported perceptions or displayed behavior of the mothers in this study may not reflect
their actual beliefs and behaviors.
Another possibility is that mothers who were abused might over-compensate by
setting fewer limits or boundaries and playing more of a ―friend‖ role with their
daughters, where warm affection and conversations abound. As a result, minimal
structure or boundaries would be present; this overcompensation might be particularly
the case if the daughter had also been abused. This theory might also explain why in
this study there were not differences between abused and non-abused mothers in terms
of punitive control, showing negative disapproving behavior, or authoritarianism, but
there were differences for mothers with a history of abuse who reported that they
struggle with setting fair and consistent boundaries. They may believe that
―responsiveness‖ is the most important part of parenting and that ―demandingness‖ or
any type of consistent control might be perceived by their daughters as unloving.
57

Furthermore, more permissive parenting practices in mothers with a history of
child sexual abuse may be a result of role reversal or a hypersensitive aversion to being
an authority figure, and risking that their daughters might feel at all powerless. Though
theory suggests that the parenting dimensions of warmth and control are orthogonal, or
independent, this has not been sufficiently proven, and as can be seen in this discussion,
there may be complicated inter-relationships among these related variables. In fact,
mothers with a history of sexual abuse may think of warmth and control as opposite
ends of one spectrum, and not as two separate dimensions of parenting. They may
perceive that they must be either warm and nurturing or punitive and controlling.
Because of their history, they may not comprehend that the use of appropriate
boundaries and control can be compatible with warm, nurturant parenting. This would
be interesting to examine in future studies.
Also, there may be other mediating or moderating factors that may influence
these relationships and account for these findings. For instance, because some mothers
in both groups had recently discovered that someone in the family had sexually
victimized their daughter, their parental perceptions and behaviors related to parenting
that daughter may be skewed. Certain aspects of parenting may be influenced by the
daughter‘s sexual abuse status. Particularly when considering the observational
findings, results might be confounded by the daughter‘s victimization. For example,
mothers, regardless of their own sexual abuse history, may parent and interact
58

differently with daughters who have recently been sexually abused.
24
The parenting of
mothers who have a history of abuse and have recently discovered their daughters‘
abuse may be particularly impacted. In this particular sample, in which there is an over-
representation of mothers with daughters who had just recently disclosed sexual abuse,
results may be confounded.
Finally, these findings also could be due to limited power to detect differences in
this sample. Using a larger sample, future analyses in examining whether parenting
differs among mothers with a child sexual abuse history when the child‘s sexual abuse
status is considered as well would be helpful in understanding these issues more fully.
In terms of the impact of child sexual abuse on parenting, difficulty with the
control dimension, such as setting limits and boundaries, may contribute to a
vulnerability to sexual victimization in the next generation. Specifically, if mothers are
not setting limits or boundaries with other adults or with their own daughters, their
daughters might be exposed to circumstances in which perpetrators have more access to
them. Furthermore, if daughters have not had a maternal role model teaching them how
and when to set appropriate limits and boundaries, they may be more vulnerable to
victimization.

24
It would be interesting to compare self-rated parenting to observed parenting in which a
mother is observed with one daughter who had been sexually abused and with another daughter who had
not been sexually abused.
59

Limitations
Approaches to Measuring Parenting
When measuring most aspects of parenting, issues arise that are potential
limitations. One problem that many studies deal with is the possibility that observed
behavior in a research setting is not necessarily representative of typical interactions and
parental behavior. Likewise, difficulty can arise with self-report measures for many
constructs of parenting, which are prone to strong social desirability. More research is
needed to examine the correspondence between observed parenting, self-reported
parental practices, and how parents actually parent (Locke & Prinz, 2002).
It is important to consider that the coding system developed for this study was
based on a previous coding system and on the childrearing literature; it was not an
established validated system and needs to be utilized in future studies for further
credibility. The constructs and specific behaviors chosen to represent the salient
dimensions during observation may be limited in what was both included and excluded
in the coding system. For instance, on a microanalytic level, the ―control‖ dimension of
interest may have been better represented by additional behaviors not examined, and on
a macroanalytic level, other dimensions of childrearing not included in this system
might be helpful in understanding parenting more fully.
While both observational and self-report data are vulnerable to inaccuracy for
different reasons, observational data may be more accurate than self-report for certain
aspects of parenting that can be measured behaviorally, since two ―objective‖ raters are
60

examining the demonstrated behavior. In fact, the observed behavioral warmth, the
observed global warmth, and the self-reported warmth variable all demonstrated that
mothers with a history of sexual abuse are warmer as parents. However, only the
observational warmth measures are significantly related to sexual abuse status. Thus,
the two observational warmth dimensions are validating of one another in that they both
were significantly associated with a history of sexual abuse. Because these two
variables are highly correlated with each other as well, (p=.42), future observational
studies may benefit from being able to use only global ratings which require much less
time and training than behavioral frequency counts.
25

Measurement: Sexual Abuse Status
One limitation of this study is the small sample size of mothers with a sexual
abuse history. While it is known that in this sample, 85% of those sexually abused had
a familial perpetrator, and that the mean age in which abuse began was age 8, specific
information about sexual abuse experiences was ultimately not able to be used in
analyses due to the small sample size. Not being able to include or control for the
degree of severity, the frequency of abuse, the relationship to the perpetrator, or
differing experiences of abuse, makes drawing conclusions about the impact of child
sexual abuse on childrearing difficult.

25
However, it may be that the global data may not be as effective without the behavioral ratings, since the
behavioral ratings force coders to pay attention to certain behaviors which likely influence their global
ratings.
61

Racial and Cultural Issues
It is known that there are often different approaches to childrearing by different
racial or ethnic groups. Though race was controlled for in analyses, how racially and
culturally different subjects interpreted the childrearing questions may have been a
limitation in this study as well. It is also important to note that three different races
comprised the four observational coders in this study. Global ratings of parenting are
obviously very subjective, and while there were clear definitions of what types of
actions constituted specific behaviors to be noted in frequency, some level of
interpretation is unavoidable even with event-based coding. Thus, the observational
variables may have been impacted by racial differences in parenting perceptions for
both the mothers in the sample and the coders.
Additionally, the sample recruited for the original study included more African-
Americans in the comparison group (60% African-Americans in the comparison group
where there was no history of sexual abuse, and only 37.1% in the abused group.) As
mentioned earlier in this study, daughters who have been sexually abused are
significantly more likely to have mothers with a sexual abuse history. Thus, because
there were more African-Americans in the comparison group, results regarding
childrearing practices and behaviors may be confounded, despite controlling for race in
regression analyses. Numerous significant correlations between race and the dependent
variables in this study indicate that these issues are important to consider in interpreting
results.

62

Generalizing Results
Though one of the strengths of this study is that it differs from many studies that
use clinical samples by using a community sample, the results still cannot be
generalized to all child sexual abuse victims. It is important to remember that half of
the mothers in this sample had daughters who had recently disclosed sexual abuse at the
hands of a perpetrator within the family. Thus, the mothers in this sample were not
homogenous in this regard. Regardless of whether or not these mothers had their own
abuse histories, mothers whose daughters have been abused may be depressed (Everett
et al., 1998; Trickett et al., 1995), or have other issues that could potentially skew their
responses to childrearing perceptions as well as how they interact with their daughters.
Furthermore, items related to warmth and control that are measured without the
context of specific parenting situations are limited and preclude generalizing. In
addition, the warmth dimension has been measured in many ways across studies,
including affection, responsiveness, warmth, encouragement, and support, to name a
few, and research has not determined whether it should be measured as a global
construct or as sub-dimensions (Locke & Prinz, 2002). Lastly, these results cannot be
generalized because of the relatively small sample size, which makes further
examinations within the subgroups of abused mothers problematic.
Other Explanations

Because credible research on the topic at hand is almost nonexistent, this study
was exploratory in that it began to look at relationships between certain variables that
are likely to be parts of a complex puzzle. In this regard, it is likely that other factors,
63

which may be viable explanations for either direct or mediating relationships with the
variables used in this study, have been omitted. For example, this study did not
examine substance abuse, maternal mental health, and other personal or relational
functioning variables, such as attachment, that are likely to influence parenting. There
are many other possible mediating and moderating variables that will need to be
examined in the future.
Future Research
Sexual Abuse
To begin, studies that address the aforementioned limitations of this study are
important areas for future research. Studies that give clear parameters of what
constitutes child sexual abuse and include details of the abuse experience are also
necessary contributions to the state of knowledge in this field. There is a great need for
more studies using community samples and observational data to examine the sequelae
of child sexual abuse on multiple aspects of parenting. Further, childrearing, maternal
reports on aspects of family of origin, and other family environmental variables, should
be taken into consideration when looking at child sexual abuse outcomes, as many of
these influences may mediate the effects of abuse upon outcomes. For example,
attachment theory and research suggests that it is not traumatic childhood events which
impact parenting, per se, but the degree with which those experiences have been
resolved, which has the biggest influence on parenting (Cohn, Cowan, Cowan, &
Pearson, 1992; Crandell, Fitzgerald, & Whipple, 1997; Crowell & Feldman, 1988,
1991; Fonagy, Steele, & Steele, 1991; Haft & Slade, 1989; Lichtenstein, Belsky, &
64

Crnic, 1998; Siegel, 1999; Sroufe, Egeland, Carlson, & Collins, 2005; van IJzendoorn
& Bakermans-Kranenburg, 1996; van IJzendoorn & Fox, 1995). Therefore, studies
which include adult attachment style, and which examine whether mothers with a
history of sexual abuse have coherently integrated abusive experiences, may be
enlightening. Studies that examine whether parenting is different after the discovery of
a child‘s victimization would be important contributions to the maltreatment field as
well.
Childrearing
In terms of parenting, more studies are needed that include observational data so
as not to rely solely on self-report perceptions. In this study, the observational variables
contributed heavily to gaining more insight into this sample. Further studies on how
sexual abuse history impacts other dimensions of childrearing, which also include child
outcomes, are needed. Also, guided by the literature, testing moderating and mediating
relationships among a variety of variables that may impact childrearing would greatly
increase knowledge. Finally, studies are needed in which there is an examination of the
profiles of the subjects with different outcomes. For instance, it could be enlightening
to examine the differences between sexually abused mothers who are parenting well
versus those who are not, or between sexually abused mothers whose daughters have
been victims and those whose daughters have not been victims. These findings may
give clues for conceptual models and future interventions. Studying cases where there
is a break in the intergenerational chain of victimization may yield important
information, particularly in terms of differences in childrearing.
65

Conclusion

Using both observational and self-report measures of the parenting dimensions
of warmth and control, this study examined specific childrearing perceptions and
behaviors in order to begin to address the lack of research on how parenting may differ
for mothers with a history of sexual abuse. These findings may be useful in
understanding ways the childrearing environment may vary for mothers with differing
histories, which may lead to better interventions for mothers who have a history of
sexual abuse. Additionally, because these questions have not been answered
sufficiently in the literature, results of this study may help with the development of
future hypotheses and ultimately, with the development of a conceptual model to guide
future research in parenting and intergenerational sexual abuse victimization. This
study contributes to the emerging body of research that suggests that parenting may
indeed be impacted by childhood sexual abuse, and it is just a beginning step in
exploring the ways that past sexual abuse experiences may impact parenting.
This study demonstrated that mothers with a history of child sexual abuse
provided warm interactions with their daughters, and that of the parenting variables
examined in this study, the only difficulty found for CSA mothers was providing a
consistent, positive structure. For mothers with a history of sexual abuse and
professionals who work with these populations, these findings may provide hope, and
aid in the development of more relevant and targeted interventions to assist mothers
with appropriate use of parental control.
66

REFERENCES

Alexander, P. C. (2000). Childhood sexual abuse history and role reversal in parenting.
Child Abuse & Neglect, 24(6), 829-838.
Bagozzi, R. P., Phillips, L. W., & Yi, Y. (1991). Assessing construct validity in
organizational research. Administrative Science Quarterly, 36.
Banyard, V. L. (1997). The impact of childhood sexual abuse and family functioning on
four dimensions of women's later parenting. Child Abuse & Neglect, 21(11),
1095-1107.
Baumrind, D. (1972). An exploratory study of socialization effects on black children:
Some black-white comparisons. Child Development, 43, 262-267.
Baumrind, D. (1991). Effective parenting during the early adolescent transition. In P. A.
Cowan & M. E. Hetherington (Eds.), Family transitions: Advances in family
research (Vol. 2, pp. 111-164). Hillsdale, NJ: Erlbaum.
Briere, J., & Runtz, M. A. (1988). Symptomatology associated with childhood sexual
victimization in a nonclinical adult sample. Child Abuse & Neglect, 12(51-59).
Briere, J., & Runtz, M. A. (1993). Childhood sexual abuse: Long-term sequelae and
implications for psychological assessment. Journal of Interpersonal Violence, 8,
312–330.
Browne, A., & Finkelhor, D. (1986). Impact of child sexual abuse: a review of the
research. Psychological Bulletin, 99, 66–77.
Burkett, L. P. (1991). Parenting behaviors of women who were sexually abused as
children in their families of origin. Family Process, 30, 421-434.
Campbell, D. T., & Fiske, D. W. (1959). Convergent and discriminant validation by the
multitrait-multimethod matrix. Psychological Bulletin, 56, 81-105.

67

Clark, L., & Watson, D. (1995). Constructing validity: Basic issues in objective scale
development. Psychological Assessment, 7(3), 309-319.
Cliff, N. (1988). The eigenvalues-greater-than-one rule and the reliability of
components. Psychological Bulletin, 103, 276-279.
Cohen, J., Cohen, P., Aiken, L. S., & West, S. G. (2002). Applied multiple regression:
Correlation analysis for the behavioral sciences (3rd ed.). Mahwah, NJ:
Lawrence Erlbaum.
Cohen, T. (1995). Motherhood among incest survivors. Child Abuse & Neglect, 19,
1423-1429.
Cohn, D., Cowan, C. P., Cowan, P. A., & Pearson, J. L. (1992). Mothers' and fathers'
working models of childhood attachment relationships, parenting styles, and
child behavior. Development & Psychopathology, 4(3), 417-431.
Cole, P. M., & Putnam, F. W. (1992). Effect of incest on self and social functioning: A
developmental psychopathology perspective. Journal of Consulting and Clinical
Psychology, 60(2), 174-184.
Cole, P. M., & Woolger, C. (1989, April). The role of emotion in the parenting
difficulties of incest victims. Paper presented at the SRCD, Kansas City, MO.
Cole, P. M., Woolger, C., Power, T. G., & Smith, K. D. (1992). Parenting difficulties
among adult survivors of father-daughter incest. Child Abuse & Neglect, 16,
239-249.
Comrey, A. L. (1988). Factor-analytic methods of scale development in personality and
clinical psychology. Journal of Consulting and Clinical Psychology, 56(5), 754-
761.
Cowan, P. A., Powell, D., & Cowan, C. P. (1998). Parenting interventions: A family
systems perspective. In W. Damon (Ed.), Handbook of child psychology, 5th
edition: Child psychology in practice (Vol. 4, pp. 3-72). United States: John
Wiley & Sons.
68

Crandell, L. E., Fitzgerald, H. E., & Whipple, E. E. (1997). Dyadic synchrony in parent-
child interactions: A link with maternal representations of attachment
relationships. Infant Mental Health Journal, 18(3), 247-264.
Crowell, J. A., & Feldman, S. (1988). Mothers' internal models of relationships and
children's behavioral and developmental status: A study of mother-child
interaction. Child Development, 59, 1273-1285.
Crowell, J. A., & Feldman, S. (1991). Mothers' working models of attachment
relationships and mother and child behavior during separation and reunion.
Developmental Psychology, 27, 597-605.
Darmawan, I. G. (2002). NORM software review: Handling missing values with
multiple imputation models. Evaluation Journal of Australia, 2(1), 51-57.
DiLillo, D. (2001). Interpersonal functioning among women reporting a history of
childhood sexual abuse: empirical findings and methodological issues. Clinical
Psychology Review, 21(4), 553-576.
DiLillo, D., Tremblay, G. C., & Peterson, L. (2000). Linking childhood sexual abuse
and abusive parenting: The mediating role of maternal anger. Child Abuse &
Neglect, 24(6), 767-779.
Douglas, A. R. (2000). Reported anxieties concerning intimate parenting in women
sexually abused as children. Child Abuse & Neglect, 24(3), 425-434.
Everett, B., Trickett, P. K., & Putnam, F. W. (1998, July). Depression symptoms in
mothers of sexually abused females. Paper presented at the Poster presented at
the meeting of the International Society for the Study of Behavioral
Development, Berne, Switzerland.
Finkelhor, D., Moore, D., L., H. S., & Strauss, M. A. (1997). Sexually abused children
in a national survey of parents: Methodological issues. Child Abuse & Neglect,
21(1), 1-9.
69

Fitzgerald, M. M., Shipman, K. L., Jackson, J. L., McMahon, R. J., & Hanley, H. M.
(2005). Perceptions of parenting versus parent-child interactions among incest
survivors. Child Abuse & Neglect, 29, 661-681.
Floyd, F. J., & Widaman, K. F. (1995). Factor analysis in the development and
refinement of clinical assessment instruments. Psychological Assessment, 7(3),
286-299.
Fonagy, P., Steele, H., & Steele, M. (1991). Maternal representations of attachment
during pregnancy predict the organization of infant-mother attachment at one
year of age. Child Development, 62, 891-905.
Franklin, S. B., Gibson, D. J., Robertson, P. A., Pohlmann, J. T., & Fralish, J. S. (1995).
Parallel Analysis: a method for determining significant principal components.
Journal of Vegitation Science, 6, 99-106.
Gardner, F. (2000). Methodological issues in the direct observation of parent-child
interaction: Do observational findings reflect the natural behavior of
participants? Clinical Child and Family Psychology Review, 3(3), 185-193.
Gelinas, D. J. (1983). The persisting negative effects of incest. Psychiatry, 46, 312–332.
Goldman, J. D. G., & Padayachi, U. K. ( 2000). Some methodological problems in
estimating incidence and prevalence in child sexual abuse research. The Journal
of Sex Research, 37(4), 305.
Gordis, E. B. (2005). Conducting intra-class correlations for observational coding data
when training coders (pp. Email and phone conversations).
Gorsuch, R. L. (1983). Factor Analysis. Hillsdale, NJ: Lawrence Erlbaum.
Haft, W. L., & Slade, A. (1989). Affect attunement and maternal attachment: A pilot
study. Infant Mental Health Journal, 3, 157-172.
70

Hayton, J. C., Allen, D. G., & Scarpello, V. (2004). Factor retention decisions in
exploratory factor analysis: A tutorial on parallel analysis. Organizational
Research Methods, 7(2), 191-106.
Herman, J. (1981). Father-daughter incest. Cambridge, MA: Harvard University Press.
Herman, J., & Hirschman, L. (1981). Families at risk for father-daughter incest. The
American Journal of Psychiatry, 138(7), 967-970.
Horn, J. L. (1965). A rationale and test for the number of factors in factor analysis.
Psychometrika, 32, 179-185.
Kaiser, H. F. (1960). The application of electronic computers to factor analysis.
Educational and Psychological Measurement, 20, 141-151.
Kaufman, J., & Zigler, E. (1987). Do abused children become abusive parents?
American Journal of Orthopsychiatry, 57, 186-192.
Kreklewetz, C. M., & Piotrowski, C. C. (1998). Incest survivor mothers: Protecting the
next generation. Child Abuse & Neglect, 22(12), 1305-1312.
Kuczynski, L. (2003). Beyond bidirectionality: Bilateral conceptual framework for
parent-child relations.Unpublished manuscript.
Lautenschlager, G. J. (1989). A comparison of alternatives to conducting monte carlo
analyses for determining parallel analysis criteria. Multivariate Behavioral
Research, 24(3), 365-395.
Lichtenstein, P. J., Belsky, J., & Crnic, K. (1998). Earned security, daily stress, and
parenting: A comparison of fie alternative models. Development and
Psychopathology, 10, 21-38.
Lindahl, K. M., & Malik, N. M. (2000). System for coding interactions and family
functioning (SCIFF): A coding system for family problem
discussions.Unpublished manuscript, University of Miami.
71

Locke, L. M., & Prinz, R. J. (2002). Measurement of parental discipline and nurturance.
Clinical Psychology Review, 22, 895-929.
Lyons-Ruth, K., & Block, D. (1996). The disturbed caregiving system: Relations
among childhood trauma, maternal caregiving, and infant affect and attachment.
Infant Mental Health Journal, 17(3), 257-275.
Maccoby, & Martin. (1983). Socialization in the context of the family: Parent-child
interaction. In E. M. Hetherington (Ed.), Handbook of child psychology: Vol. 4.
Socialization, personality, and social development (4th ed., Vol. 4, pp. 1-101).
New York: Wiley.
Main, M., & Goldwyn, R. (1984). Predicting rejection of her infant from mother's
representation of her own experience: Implications for the abused-abusing
intergenerational cycle. Child Abuse & Neglect, 8(203-217).
Maker, A. H., & Buttenheim, M. (2000). Parenting difficulties in sexual-abuse
survivors: A theoretical framework with dual psychodynamic and cognitive-
behavioral strategies for intervention. Psychotherapy, 37(2), 159-170.
Margolin, G., Oliver, P. H., Gordis, E. B., O'Hearn, H. G., Medina, A. M., Ghosh, C.
M., & Morland, L. (1998). The nuts and bolts of behavioral observation of
marital and family interaction. Clinical Child and Family Psychology Review,
1(4), 195-211.
Mc Graw, K. O., & Wong, S. P. (1996). Forming inferences about some intraclass
correlation coefficients. Psychological Methods, 1(1), 30-46.
Morton, N., & Browne, K. D. (1998). Theory and observation of attachment and its
relation to child maltreatment: A review. Child Abuse & Neglect, 22(11), 1093-
1104.
Newcomb, M. D., & Locke, T. F. (2001). Intergenerational cycle of maltreatment: A
popular concept obscured by methodological limitations. Child Abuse &
Neglect, 25, 1219-1240.
72

Noll, J. G., Horowitz, L. A., Bonanno, G. A., Trickett, P. K., & Putnam, F. W. (2003).
Revictimization and self-harm in females who experienced childhood sexual
abuse: Results from a prospective study. Journal of Interpersonal Violence,
18(12), 1452-1471.
Oates, R. K., Tebbutt, J., Swanston, H., Lynch, D. L., & O'Toole, B. I. (1998). Prior
childhood sexual abuse in mothers of sexually abused children. Child Abuse &
Neglect, 22(11), 1113-1118.
Putnam, F. W., & Trickett, P. K. (1997). Psychobiological effects of sexual abuse: A
longitudinal study. Annals of the New York Academy of Sciences, 821, 150-159.
Rubin, D. B. (1987). Multiple imputation for nonresponse in surveys. New York: J.
Wiley and Sons.
Rumstein-McKean, O., & Hunsley, J. (2001). Interpersonal and family functioning of
female survivors of childhood sexual abuse. Clinical Psychology Review, 21(3),
471-490.
Ruscio, A. M. (2001). Predicting the child-rearing practices of mothers sexually abused
in childhood. Child Abuse & Neglect, 25, 369-387.
Schafer, E. S. (1959). A circumplex model for maternal behavior. Journal of Abnormal
and Social Psychology, 59, 226-235.
Schuetze, P., & Eiden, R. D. (2005). The relationship between sexual abuse during
childhood and parenting outcomes: Modeling direct and indirect pathways.
Child Abuse & Neglect, 29(6), 645-659.
Sickel, A. E., Noll, J. G., Moore, P. J., Putnam, F. W., & Trickett, P. K. (2002). The
long-term physical health and healthcare utilization of women who were
sexually abused as children. Journal of Health Psychology, 7(5), 583-597.
Siegel, D. J. (1999). The developing mind. New York: Guilford Press.
73

Siegel, D. J., & Hartzell, M. (2003). Parenting from the inside out. New York: Putnam.
Smith, G. T., & McCarthy, D. M. (1995). Methodological considerations in the
refinement of clinical assessment instruments. Psychological Assessment, 7(3),
300-308.
Sroufe, L. A., Egeland, B., Carlson, E. A., & Collins, W. A. (2005). The development of
the person: The minnesota study of risk and adaptation from birth to adulthood.
New York: Guilford Press.
Sroufe, L. A., & Fleeson, J. (1986). Attachment and the construction of relationships. In
W. W. Hartup & Z. Rubin (Eds.), Relationships and development. Hillsdale,
N.J.: Lawrence Earlbaum Associates.
Stumpf, H., & Parker, W. D. (2000). A hierarchical structural analysis of perfectionism
and its relation to other personality characteristics. Personality and Individual
Differences, 28, 837-852.
Tamraz, D. N. (1996). Nonoffending mothers of sexually abused children:
Comparisons of opinions and research. Journal of Child Sexual Abuse, 5(4), 75-
104.
Trickett, P. K. (1997). Sexual and physical abuse and the development of social
competence. In S. S. Luthar, J. A. Burack, D. Cicchetti & J. R. Weisz (Eds.),
Developmental psychopathology: Perspectives on adjustment, risk, and
disorder (pp. 390-416): Cambridge University Press.
Trickett, P. K. (1998). Multiple maltreatment and the development of self and emotional
regulation. In B. Rossman & M. Rosenberg (Eds.), The multiple victimization of
children: Conceptual, developmental, research, and clinical issues.
Binghamton, NY: Hawarth Press.
Trickett, P. K. (2005). Child-rearing coding system.
74

Trickett, P. K. (2006). Defining Sexual Abuse. In M. M. Feerick, J. F. Knutson, P. K.
Trickett & S. Flanzer (Eds.), Child Abuse and Neglect: Definitions,
Classifications, and a Framework for Research. Baltimore: Brookes.
Trickett, P. K., Everett, B., & Putnam, F. W. (1995). Child rearing beliefs and practices
of mothers of sexually abused girls and female adolescents. Paper presented at
the Society for Research on Child Development, Indianapolis.
Trickett, P. K., Noll, J. G., Reiffman, A., & Putnam, F. W. (2001). Variants of
intrafamilial sexual abuse experience: Implications for long-term development.
Journal of Development and Psychopathology, 13, 1001-1019.
Trickett, P. K., & Putnam, F. W. (1993). Impact of child sexual abuse on females:
Toward a developmental psychobiological integration. Psychological Science,
4(2), 81-87.
Trickett, P. K., & Putnam, F. W. (1998). Developmental consequences of child sexual
abuse. Washington D.C.: American Psychological Association.
van IJzendoorn, M. H., & Bakermans-Kranenburg, M. J. (1996). Intergenerational
transmission of attachment: Towards a contextual approach. Polish Quarterly of
Developmental Psychology, 2(1), 1-16.
van IJzendoorn, M. H., & Fox, N. A. (1995). Adult attachment representations, parental
responsiveness, and infant attachment: A meta-analysis on the predictive
validity of the adult attachment interview. Psychological Bulletin, 117(3), 387.
Zuravin, S., & Fontanella, C. (1999). Parenting behaviors and perceived parenting
competence of child sexual abuse survivors. Child Abuse & Neglect, 23(7), 623-
632.

75

APPENDICES
76

Appendix A
Sexual Abuse History Interview

Sexual Abuse Status: The following questions were embedded in a developmental
history and were obtained during an interview of the mothers which focused on the
mother‘s history.

―As you know, one of the things we are looking at in this research is the effect
sexual abuse might have on a girl‘s development. As you probably know from the
newspapers and television, more and more cases of sexual abuse are being reported
everyday—it seems to be a very serious problem and unfortunately, a very common
problem in our society as well. Was anyone in your family to your knowledge, ever
touched in a sexual way when he or she was a child? _______________ (If yes,
probe for who and the details of the abuse.)
____________________________________________ (If mother did not mention
self)

Then you had no such experience as a child? (If yes, probe as above.)‖

77

Appendix B
Self-reported Childrearing--Types of Discipline Used

Items indicating Type of Discipline Used include a series of 8 questions which
were obtained during a semi-structured interview conducted the second session that the
subject was seen in the first wave of data collection, which focused on the
developmental history of the child. The subjects were asked how they would describe
themselves as a parent. The interviewer asked subjects to choose where they fit best on
a five-point scale, which is a spectrum of discipline tools.

―This is a list of different types of discipline that parents sometimes use when
they are trying to socialize their children or teach their children how they want
them to behave.‖ (Read through the list with the mother, giving examples of
any type of discipline that might not be immediately clear to her.)

―Could you please go down the list and tell me the number that best describes
how often you use each type of discipline?‖

Never Often
 Explanations, reasoning, teaching 1 2 3 4 5

 Verbal prohibitions, saying ―no‖ 1 2 3 4 5

 Physical punishment 1 2 3 4 5

 Verbal punishment 1 2 3 4 5

 Isolation, sending to room, ―time out‖ 1 2 3 4 5

 Taking away privileges 1 2 3 4 5

 Threaten punishment (not having to use it) 1 2 3 4 5

 Ignoring bad behavior 1 2 3 4 5

 Guilt, silent treatment 1 2 3 4 5

78

Appendix C
Self-reported Childrearing—Parenting Style

Items indicating Parenting Style include a series of 8 questions which were
obtained during a semi-structured interview conducted the second session that the
subject was seen in the first wave of data collection, which focused on the
developmental history of the child. The subjects were asked how they would describe
themselves as a parent. The interviewer asked subjects to choose where they fit best on
a five-point scale, which is a spectrum of parenting characteristics.

―The first thing I‘d like to know today is how you would describe yourself as a
parent. We‘re going to use a card we used once last time. You recall this card
has words on it that are often used to describe parents. Parents can be authority
figures to their children, for example, or they can be friends to their children. A
‗1‘ on this card would mean that a parent saw herself primarily in the role of an
‗authority figure‘ to her children, ‗5‘ would mean that she saw herself primarily
in the role of ‗friend‘. A number between ‗1‘ and ‗5‘ would mean that she saw
herself somewhere in between the two, a ‗2‘, for example, would mean she saw
herself mostly as an ‗authority figure‘ but also a little bit as a ‗friend‘, whereas a
‗4‘ would mean she saw herself mostly as a ‗friend‘ but also a little bit as an
‗authority figure.‘‖

―I would like you to show me on the number on the card that best describes the
way you see yourself as a parent. Let‘s start with the first pair of words: parents
can be authority figures for their children or they can be friends or something in
between. Where on the line between these two words is the number that you
think bests describes you as a parent?‖

authority figure friend 1 2 3 4 5
lenient strict 1 2 3 4 5
irritable easy-going 1 2 3 4 5
inconsistent consistent 1 2 3 4 5
affectionate unaffectionate 1 2 3 4 5
unfair fair 1 2 3 4 5
serious playful 1 2 3 4 5
reasons, explains makes rules 1 2 3 4 5

79

Appendix D
Etch-A-Sketch Parenting Style Coding System Manual

Etch-A-Sketch Parenting Style Coding System

AN OBSERVATIONAL CODING MANUAL:
INSTRUCTIONS AND DEFINITIONS

Tina Payne Bryson, MSW, PhD, ABD
University of Southern California
School of Social Work
Dissertation Project 2005-2006

80

INTRODUCTION

The main purpose of the Etch-A-Sketch Parenting Style Coding System is to
behaviorally assess parent-child interactions during an increasingly difficult exercise.
Verbal and non-verbal parenting behaviors or types of interactions, the style or those
interactions, and the overall parenting style of the mother will be assessed. The overall
purpose of observing this data is to assess childrearing domains.

The mother-daughter dyads that will be assessed via this coding system were
instructed to complete a series of tasks together, in which they were given five
increasing difficult pictures and asked to work together to draw the pictures. Dyads
were given three minutes to complete each of the five tasks. Interactions lasted an
average of 15 minutes and were videotaped. Your job will be to view the interactions
on the videotapes and code using the provided coding sheets.

Coders will be trained until acceptable reliability is reached; ongoing meetings
and trainings will take place throughout the coding process to ensure high reliability and
to prevent coder drift. Coders will complete one rating sheet for each of the five
increasingly difficult tasks for each parent-child dyad. Coders will be observing and
coding three different aspects of interaction. First, for each of the five picture tasks,
coders will be rating the nature or Types of Interactions that are present. These will
include verbal and non-verbal behaviors and will be coded based on frequency.
Second, for each of the five tasks, coders will rate five different Styles of Interaction on
a five point scale. Finally, coders will give a global rating at the end of all five picture
tasks which will be the coders‘ assessments of Parenting Style. Coders will then fill out
a checklist at the end of each coding session to ensure that missing data is minimal.

81

DEFINITIONS OF ETCH-A-SKETCH PARENTING STYLE CODING SYSTEM

Nature of Behavior or Interaction: These are coded in terms of frequency. Each
time one of the behaviors is observed, it should be marked on your coding sheet
next to the item. After watching all of the tasks, you will total the frequency for
each item.

Verbal:

Control/Motivation to do the design

1. Positive character attribution of child—Statements attributing to child a positive
trait or quality. This positive comment is focused on the CHILD, not on the
child‘s work. Examples: ―You are a hard worker,‖ ―You are a good girl.‖

2. Negative character attribution of child—Statements attributing to child a
negative trait or quality. This negative comment is focused on the CHILD, not
on the child‘s work. Examples: ―You are lazy,‖ ―You are difficult.‖

3. Positive reinforcement/approval—Verbalizations regard the good quality of the
child‘s work or behavior (other than character). This positive comment is
focused on the WORK, not on a characteristic of the child. Examples: ―Good
job,‖ ―You did that one quickly,‖ ―Great!,‖ ―There you go!‖

4. Negative reinforcement/disapproval—Verbalizations involving the poor quality
of the child‘s work (versus character) which may thwart motivation. This
negative comment is focused on the child‘s WORK, not on a characteristic of
the child. Examples: ―That‘s wrong‖, ―You didn‘t do that carefully,‖ ―Don‘t
shake it so hard.‖

5. Reasoning/Explanations—Verbalizations which give reasons or explanations
which can motivate the child. Examples: explaining the need for cooperation
and the effects it will have, reasoning with cause and effect, referencing social
norms, like ―If we work together it will be easier,‖ ―If you hurry, we can get
home sooner,‖ ―If you shake the Etch-A-Sketch, it erases and we can start
again.‖ These are not reasons/explanations regarding the task directions, which
should be coded as instruction/information giving.

6. Prohibitions—Verbal prohibitions unelaborated by reasons. Examples: ―No‖,
―Stop‖, ―Whoa‖, ―You can‘t. . . ―. Do not code this if it is part of an instruction,
such as ―No, turn your knob the other way.‖ If the prohibition also includes an
instruction, code this as well as the instruction/information code.
82

7. Bargain/bribe—Promise of positive consequences later in return for immediate
cooperation. Example: ―If you will finish this, we will have time to stop at Mc
Donald‘s on the way home.‖

8. Threaten—Promise of negative consequences if the child does not cooperate.
Example: ―If you don‘t do this, you can‘t watch TV when we get home.‖

Teaching

1. Attention directing—Attempts by mother to focus child‘s attention on the task.
Examples: ―Watch where my line is going,‖ ―Look at the picture,‖ ―Come here
and start working.‖

2. Instruction/information giving—Commands and requests focused on completing
the task or requests and suggestions on how something should be done.
Examples: ―We should start turning the knobs together‖, ―Start turning toward
me when you get here‖, ―Turn the knob to the right,‖ ―Go back a bit,‖ ―Wait; ok,
stop,‖ ―Come on, come on, keep going.‖ These instructions and information are
task oriented.

3. Information seeking/questioning—Statements or questions designed to elicit
directions for proceeding with the task; requests for information. Examples:
―What should we do next?,‖ ―How do I make this turn?‖

Other

1. Positive self/dyad attribution—Statements attributing to self or dyad a positive
trait or ability. Examples: ―I am so smart‖, ―We are good at this.‖

2. Negative self/dyad attribution—Statements attributing to self/dyad a negative
trait or ability. Examples: ―I can‘t do anything right‖, ―We messed up‖,
includes ―I don‘t know how to do this‖ if there is no qualifying statement about
the task being too difficult.

3. Positive task attribution—Statements attributing positive characteristics to the
task. Examples: ―This is easy‖, ―This is fun‖, ―Wow.‖

4. Negative task attribution—Statements attributing negative characteristics to the
task. Examples: ―This is too hard‖, ―This is stupid.‖

5. Non-instructional task-related conversation. Examples: ―We‘re done,‖ ―This
line is a little short.‖
83

6. Off-task conversation which doesn‘t interfere with task completion. The task is
able to continue while they talk. Example: ―How was school?‖

7. Diverting/Avoiding—Mother changes focus to something off task, or if already
off task, focuses on something else off task. Interferes with completion of task.
Examples: ―Is that a mirror or a window over there?,‖ ―I wonder when we will
be able to go home.‖

8. Refusal/Delay—Any verbal non-compliance including delaying statements.
Examples: ―I won‘t do anymore,‖ ―Let‘s do this later.‖

9. Responding to Child—this is anytime the mother is responding to the child‘s
question or statement. For example, it can be reflective, such as when the child
says ―Go this way‖ and the mother responds, ―You want me to turn it toward
you?‖ or it can be a clarifying question. ―This way?‖ or it can be answering the
child in another way. The important thing is that it is only in direct response to
the child‘s verbal prompting. If she responds and then says something that fits
into another coding category, code that as well.

Nonverbal

Control/Motivation (includes grunts, sighs etc.)

1. Restraining--Physical movement controlling movement of child‘s body.
Behavior that prevents or prohibits action. Examples: holding the child‘s hand
to prevent turning knobs, holding child in chair to prevent them from running
around room.

2. Physical punishment/physical aggression. Examples: hitting, slapping, pushing,
includes impulsively grabbing the Etch-A-Sketch or pushing away the child‘s
hand.

3. Nurturance/warmth--Affectionate physical contact or demonstrating enjoyment
of the child. Examples: kissing, hugging, holding, patting, caressing, smiling at
child, or any behavior which shows that the mother is enjoying the child.

Other

1. Diverting/off-task--Behavior that is inconsistent with working successfully on
the task. Examples: day dreaming, pacing in the room, not responding
immediately to the request or suggestion of the child, making a design on the
84

Etch-A-Sketch that is not related to the task. Each time this happens, it should
be coded.

2. Demonstrating annoyance—This is behavior which may come in the form of
ignoring, inducing guilt, silent treatment, or physical manifestations of
annoyance with the CHILD. This is specifically behavior which communicates
that the mother is not pleased with or is irritated with the child. Examples:
ignoring child, eye-rolling, or sighs directed at child. Each time this happens, it
should be coded.

3. Laughter—Mother laughs. Each time this happens, it should be coded. If the
mother laughs AT the child in a way that is demeaning, this should not be
coded.
85

Definitions of Etch-A-Sketch Parenting Style Coding System

Predominant Style of Task Interaction/Emotional Tone: The affective style that is most
predominant in the mother will be chosen by the coder at the end of each of the five
Etch-A-Sketch picture tasks.

STYLE OF TASK INTERACTION/EMOTIONAL TONE: Global Ratings—Please
circle the choice which best captures the style of the mother‘s interaction/her emotional
tone at the end of each task. Please consider both frequency and intensity of behaviors.
If the style/tone occurs infrequently and is of low intensity, it should be rated at the low
end of the scale. If the style/tone is of more moderate frequency and intensity, it should
be rated in the middle range of the scale. Predominant or intense style/tone should be
rated on the high end of the scale.

1 is lowest (observe very little of it) and 5 is high (observe it predominantly)

1. Light-hearted/approaches with humor/flexible 1 2 3 4 5
2. Pleasant/happy/joyful 1 2 3 4 5
3. Apathetic/passive/flat affect 1 2 3 4 5
4. Anxious/afraid 1 2 3 4 5
5. Irritated/anxious/hostile 1 2 3 4 5
6. Who was leading the task: mom 1 2 3 4 5 child

86

DEFINITIONS OF PREDOMINANT STYLE
OF INTERACTION/EMOTIONAL TONE
26

Rate the style/affect/emotional tone/behavior that is most dominant. All scales in this
instrument should consider fundamental non-verbal elements, such as eye contact, facial
expression, tone of voice, posture, gestures, timing of response, and intensity of
response, as well as verbalizations (Siegel & Hartzell, 2003). The style does not have to
necessarily have to be directed at the child, but it can be. Consider what the mother
says and does and how she expresses herself. In other words, parents may express
happiness or fear or any of the possible styles mentioned below through verbalizations
(e.g.: loud sighs, saying ―I don‘t like this.‖), body language (e.g. rolling eyes, rigid
posture, forceful contact), or emotional tone (sounding irritated, depressed, or
impatient).

1. Light-hearted/approaches with humor/playful

This code assesses the overall level of light-heartedness, humorous approach
that these mothers show in their interactions with their daughters during the task.
This style will be demonstrated through both verbal and non-verbal cues listed
above.

Characteristics of this style:
a. Facial expressions: bright-eyed, relaxed, easy-going, usually smiling
b. Vocalizations: laughing, giggling, joking, enjoying herself and child, and
usually not interrupting the flow of ongoing stream of behavior. This may
even include fun exaggeration, imitating, or using fantasy content, such as ―I
am talking to the Etch-A-Sketch‖., content is positive and encouraging
c. Behavior: being appropriately silly and fun, ability to facilitate a fun or
funny experience for child, warm and affectionate in nature, playful

The lower end of this scale is characterized by an absence of light-hearted/easy-
going-ness and humor. The higher end of the scale is characterized by an
approach to the tasks and child in the manner(s) listed in ―Characteristics of this
Style‖.

1—Very Low: The parent shows very little light-hearted-
ness/humor/playfulness and demonstrates very few of the Characteristics of this
Style. The mother is not light-hearted or playful/humorous and remains quite
serious.

26
Descriptions are based on previous coding system designed by Penelope Trickett and by
Kristin Lindahl (SCIPD).
87

2—Low: The parent generally does not demonstrate the Characteristics of this
Style and is only very minimally light-hearted or playful during the interaction.

3—Moderate: The parent shows some of the Characteristics of this Style and
there are occasions when the mother seems somewhat light-hearted, humorous,
or playful, but this is clearly less than half of the time. The parent for the most
part is fun-loving, but there are definite parts of the interaction in which the
parent seems not to be. When the child attempts to have fun with the task or
with her mother, she generally responds appropriately.

4—Moderately High: The parent shows many of the Characteristics of this
Style in a fairly consistent manner. For about half the time, the parent is
actively light-hearted, playful, or humorous. The parent is clearly enjoying the
child. Light-heartedness/humor is the main undercurrent of this mother‘s style.

5—High: For at least half of the time the mother is actively light-hearted,
humorous, and playful in body language, tone, facial expressions and attitude.
The Characteristics of this Style are predominant nearly the entire interaction.
Light-heartedness/Approaching with Humor is present in almost all of the
interactions throughout the task. This style is predominant, obvious, and easy to
identify.

2. Pleasant/happy/joyful
This code reflects the overall emotional tone of showing happiness and being
pleasant in the affect and behavior of the mother during the interactions in this
task.

Characteristics of this Style:

a. Facial Expression: forehead smooth, eyes bright, often seen partially closed
with wrinkles in outer corners, teeth may be seen, eye contact with child
while smiling or laughing

b. Body and Behavior: physical expression of this affect may be manifest in an
overall increase in body tension, or in an animated fast rate of movement
such as jumping up and down, enthusiastic play including motoric
activation, may include approach behaviors, body posture open, oriented
physically toward the child

c. Vocalizations: audible laughter, excited vocalizations, rise in voice
intonation, squeals, animated language, expressions of pleasure with
achievements or play

88

1—Very Low: The parent shows no or very little happiness and demonstrates
very few of the Characteristics of this Style. Characteristics of sadness,
withdrawn behavior, lack of engagement with environment are predominant and
unhappiness, sadness, or distress is notable most or all of the time.

2—Low: The parent generally does not demonstrate the Characteristics of this
Style and is only minimally happy or pleasant during the interaction. This
mother generally demonstrates unhappiness, sadness, or distress. There may be
moments where she briefly demonstrates happiness during the interaction, but
after a short time she resumes demonstrating sadness. This rating should be
given if the parent is mostly unhappy, but at times shows moments of happiness.

3—Moderate: The parent shows some of the Characteristics of this Style and
there are occasions when the mother seems somewhat happy or pleasant, but this
is clearly less than half of the time.

4—Moderately High: The parent shows many of the Characteristics of this
Style in a fairly consistent manner. Pleasantness/happiness is the main
undercurrent of this mother‘s style. For about half the time, the parent is happy,
pleasant, and joyful. There may be one or two moments where the mother
seems to be unhappy, but these moments are very brief and then the mother is
happy quickly.

5—High: For at least half of the time the mother is happy and joyful in
verbalizations, tone, facial expressions and behavior. The Characteristics of this
Style are predominant.

3. Apathetic/passive/withdrawn

This code assesses the degree to which the mother withdraws or disengages
herself from the child or the task. The mother may seem detached, backed off,
or shut down physically or emotionally. Be sure NOT to code when a mother is
just shy, reserved, or quiet, unless she is clearly uninvolved or removed
(disengaged and apathetic); this is not an indication of personality or culturally
different behaviors (such as eye gaze.)

Characteristics of this style:

d. Facial Expression: avoidance of eye contact, looking down, drooping
eyelids, shapeless mouth, unexpressive and non-attending, flat facial
expression
e. Body and Behavior: ignoring child or task, turning body away, slumpy
posture, changing body position to create distance, crossing arms,
89

disinterested fidgeting, unrelated body movements, slow/lethargic
movements
f. Vocalizations: mumbled, content may be inconsistent or inappropriate for
interaction, tone of voice is flat/tired/bored/disinterested when speaking,
says things like ―Do whatever you want‖ or ―You go ahead, I don‘t feel like
it‖ or ―It doesn‘t matter‖

1—Very Low: The parent shows none or very few of the Characteristics of this
Style. The mother is not withdrawn/apathetic and remains actively engaged,
interested, and involved with the child and task.

2—Low: The parent generally DOES NOT demonstrate the Characteristics of
this Style and is only minimally withdrawn/apathetic from the interaction.
There may be moments where she briefly ignores the child, disengages or shuts
down during the interaction, (like fidgeting for a few moments and looks away
for a few moments), but after a short time she resumes active interest and
involvement. This rating should be given if the parent is engaged and involved,
but at times has a bit of indifference in tone or ignores the child on occasion.

3—Moderate: The parent shows some of the Characteristics of this Style and
there are occasions when the mother seems somewhat withdrawn or apathetic,
but this is clearly less than half of the time. The parent for the most part is
involved, but there are definite parts of the interaction in which the parent seems
detached, disinterested, or avoidant. When attempts are made to re-engage the
parent, she generally responds appropriately.

4—Moderately High: The parent shows many of the Characteristics of this
Style in a fairly consistent manner. For about half the time, the parent is
actively withdrawn or disengaged. The parent is clearly ignoring, uninvolved,
disinterested. When attempts are made to re-engage the parent, the parent
generally responds, but there are likely to be times in which the parent is
unresponsive or responds inappropriately (ignores a touch, doesn‘t laugh at a
joke, doesn‘t answer child).

5—High: For at least half of the time the mother is actively withdrawn or
disengaged in body language, tone, or attitude. When attempts are made to re-
engage the parent, she may often not respond or may respond inappropriately.
The Characteristics of this Style are predominant.

4. Anxious/Afraid
This code reflects the degree of anxiety and fear that the mother demonstrates.

90

Characteristics of this style:

g. Facial Expression: eyebrows raised in flattened appearance, forehead may
show horizontal or vertical wrinkles, eyes widely open with dilated pupils,
staring, frozen watchfulness, fearful or nervous looking
h. Body and Behavior: visual/postural/or physical avoidance, freezing in
activity, comfort-seeking, visual scanning, over-vigilant watchfulness,
nervous movements (biting lip, ringing hands, chewing on fingers), restless
anxious movements
i. Vocalizations: none, distress or appeals for help, verbalizing anxiousness or
fear

1—Very Low: The parent shows none or very few of the Characteristics of this
Style. The mother does not appear to be anxious or afraid.

2—Low: The parent generally DOES NOT demonstrate the Characteristics of
this Style and is only minimally anxious/nervous/fearful. There may be
moments where she demonstrates anxiousness or fear during the interaction, but
they are brief and she is able to relax and be present to the situation.

3—Moderate: The parent shows some of the Characteristics of this Style and
there are occasions when the mother seems somewhat anxious or fearful, but
this is clearly less than half of the time. The parent for the most part is calm, but
there are definite parts of the interaction in which the parent seems nervous or
afraid.

4—Moderately High: The parent shows many of the Characteristics of this
Style in a fairly consistent manner. For about half the time, the parent is
demonstrating fear and anxiety. The parent is clearly acting worried, nervous,
or fearful.

5—High: For at least half of the time the mother is anxious or fearful in body
language, behavior, or vocalizations. This mother is predominantly anxious or
fearful.

5. Irritated/Angry/Hostile
This code reflects the degree of irritability, anger, and hostility that the mother
demonstrates in her interactions during the task. They can be directed toward
the daughter or the task, but do not have to be.

91

Characteristics of this Style:

j. Facial Expression: eyes may be narrow or squinted, cheeks pushed up, jaw
rigid, mouth closed up, eyes pressed together, clinched teeth, frowning could
include smirk, expression of sarcasm or annoyance to anger and rage
k. Body and Behavior: tension present, not relaxed, directed or undirected
physical or verbal aggression, interruptive of ongoing behavior, aggression
could be with the purpose of causing harm, aggression with no goal but
causes harm, aggression toward child, self, objects, other people, high
negative intensity
l. Vocalizations: tone of voice is angry or annoyed, angry noises of varying
intensity, antagonizing, teasing, demands stated in a negative way without
redirection, (e.g. ―stop that noise!‖), abrupt tone of voice, using child‘s
formal name

1—Very Low: The parent shows none or very few of the Characteristics of this
Style. The mother is not irritated, angry or hostile and demonstrates patience
toward child and task.

2—Low: The parent generally DOES NOT demonstrate the Characteristics of
this Style and is only minimally irritated or frustrated with the interaction.
There may be moments where she shows irritation or anger during the
interaction, but these moments are brief and they are followed by reconnection
with the child in a loving manner.

3—Moderate: The parent shows some of the Characteristics of this Style and
there are occasions when the mother seems somewhat irritated, angry, or hostile,
but this is clearly less than half of the time. The parent for the most part is
patient and warm, but there are definite parts of the interaction in which the
parent seems irritated or angry or hostile.

4—Moderately High: The parent shows many of the Characteristics of this
Style in a fairly consistent manner. For about half the time, the parent is angry,
irritated, or hostile.

5—High: For at least half of the time the mother is actively irritated, angry, or
hostile in body language, tone, or verbalizations and/or behavior. The
Characteristics of this Style are predominant.

6. Who Was Leading the Task?
This rating scale, where you choose 1-5, is designed to reflect whether the
mother, the daughter, or both are leading or taking the initiative for the task.
92

1-- A 1 will be given in circumstances where the mother is clearly taking the
leadership role and the daughter is only involved as a passive participant who is
carrying out the instructions of the mother. The daughter exhibits no initiative,
and the mother is the only one in charge.

2—A 2 should be given when the mother is definitely leading the task and
initiating getting the task done, but involves the daughter, and the daughter is
taking some initiative as well.

3—A 3 should be given if there are equally shared amounts of time when the
mother and the daughter are leading the task, or they are collaborative the whole
time in making decisions together. There is no CLEAR leader between them—
they are equally initiating.

4—A 4 should be given when the daughter is definitely leading the task and
initiating getting the task done, but the mother sometimes takes some initiative
as well.

5—A 5 will be given in circumstances where the daughter is clearly taking the
leadership role, and the mother is only involved as a passive participant who is
carrying out the instructions of the daughter. The mother exhibits no initiative
and the child is the only one in charge.
93

Definitions of Etch-A-Sketch Parenting Style Coding System

Parenting Style: Macro-analytic descriptors reflect the global impression that the coder
feels best describes the mother‘s parenting style. After watching all five picture tasks,
coders will choose one descriptor of each set of eight which will reflect the coders
overall impression of the mother‘s predominant parenting style. You will fill out one of
these for each dyad.

Instructions: Using your overall impression of how the mother on this tape
interacted with her daughter, please choose one number which best describes the
mother in terms of style of parenting. For example, parents can be authority
figures to their children or they can be friends to their children; often these styles
of parenting are seen on a spectrum. Please choose the number which best fits
this mother‘s approach towards her daughter.

Choosing a ―1‖ would mean that the mother ONLY acted as an authority figure;
a number score of ―2‖ would mean that you observed the mother acting
primarily in the role of an ―authority figure‖ toward her daughter; a number ―3‖
would mean that you equally saw both authority figure and friend behavior; a
number ―4‖ would indicate that the mother acted primarily in the role of friend,
and a ―5‖ would indicate that the mother ONLY acted as a friend.

Choose one number for each item below that you think best describes this mother as a
parent. Please circle your choice.

1. Authority figure 1 2 3 4 5 friend

2. Lenient 1 2 3 4 5 strict

3. Irritable 1 2 3 4 5 easy-going

4. Inconsistent 1 2 3 4 5 consistent

5. Affectionate 1 2 3 4 5 unaffectionate

6. Unfair 1 2 3 4 5 fair

7. Serious 1 2 3 4 5 playful

8. Enjoys child 1 2 3 4 5 doesn‘t enjoy child

9. Encourages autonomy 1 2 3 4 5 doesn‘t encourage autonomy

94

10. Reasons/explains 1 2 3 4 5 makes rules/gives commands

11. Warm toward child 1 2 3 4 5 withholding/chilly

12. Shy/timid 1 2 3 4 5 outgoing/not shy

95

PROCEDURES AND CONVENTION RULES
Etch-A-Sketch Parenting Style Coding System

1. Coders will sign a written policy of confidentiality.
2. The interaction to be coded begins as soon as the tester places the design down
in front of subjects.
3. Code through all interval periods until the task ends.
4. Ignore non-verbal, non-related behaviors.
5. Codes are not mutually exclusive, meaning that you may code two behaviors at
once.
6. You are required to fill out the coder checklist at the end of each coding session.
7. Indicate when there has not been anything to code ―NOT CODEABLE‖ so that I
know that there is not missing data or that it just hasn‘t been filled out.
8. The interaction begins as soon as the tester places the first sheet down in front of
the subjects. Keep coding as the same coding task until the next sheet of paper
is set down.

96

Coder Checklist
Please check EVERY ITEM, EVERY TIME

1. Do you have 6 total sheets for this dyad?

2. Is your name on every sheet?

3. Is the subject number on every sheet?

4. Check to make sure you have one sheet for each task and that they are numbered
correctly 1-5 in the ―Design #‖ line.

5. Make sure you have tallied your frequency counts on each line.

6. Did you fill out the five Style of Interaction rating scales at the bottom of each
task sheet?

7. Make sure you have filled out the global Parenting Style sheet at the end.

8. Did you check TWICE that every line/question is complete?

9. Put code sheets back in your folder until our next meeting.

97

Coder Reference Sheet: Outline Overview of
Etch-A-Sketch Parenting Style Coding System

A. Nature of Behavior or Interaction: These are coded as present or not and are
counted by frequency.

 Verbal: Control/Motivation to motivate to do the design
o Positive character attribution of child (―You are a hard worker‖)
o Negative character attribution of child (―You are lazy‖)
o Positive reinforcement/approval (―Good job‖)
o Negative reinforcement/disapproval (thwarts motivation--―That‘s
wrong‖)
o Reasoning/Explanations (―If we work together it will be easier‖)
o Prohibitions (―No, stop‖)
o Bargain/bribe (―If you hurry we will have time to stop at Mc Donald‘s
on the way home‖)
o Threaten (―If you don‘t sit still, you will be punished‖)
 Verbal: Teaching
o Attention directing (―Watch where my line is going‖)
o Instruction/information giving—future (―We should start turning the
knobs together‖)
o Instruction/information giving—concurrent (―Turn the knob to the
right‖)
o Information seeking/questioning (―What should you do next?‖)
 Verbal: Other
o Positive self/dyad attribution (―I am so smart‖ or ―We are good at this‖)
o Negative self/dyad attribution (―I can‘t do anything right‖ or ―We
messed up‖)
o Positive task attribution (―This is easy‖)
o Negative task attribution (―This is too hard‖)
o Non-instructional task-related conversation (―We‘re done‖)
o Off-task conversation which doesn‘t interfere with task (―How was
school?‖)
o Diverting/Avoiding (―Is that a mirror or a window over there?‖)
o Refusal/Delay (―I won‘t do anymore‖)

 Nonverbal: Control/Motivation
o Facilitating (physical movement to help complete task, like moving
closer)
o Restraining (physical movement to impede completion like holding the
others‘ hand to prevent turning knobs)
98

o Physical punishment/physical aggression (impulsively grabbing the etch-
a-sketch or pushing away the child‘s hand)
o Nurturance (affectionate physical contact)
 Nonverbal: Other
o Working on task (turning knob, studying picture)
o Diverting/Off-task (behavior not consistent with task completion)
o Ignoring/Guilt/Silent treatment (ignoring child, eye rolling or sighs
directed at child)
o Interactive laughter

B. Style of Behavior or Interaction: These macroanalytic global ratings are intended
capture the affect that is most dominant and most related to the interaction. Coders will
choose which one best describes the mother‘s style of interaction. See Attached
definition sheet.

1. Light-hearted/humorous/joking
2. Pleasant/happy/joyful
3. Supportive/affectionate/loving
4. Apathetic/passive/depressed
5. Timid/anxious/afraid
6. Whining/clinging/demanding
7. Unhappy/sad/distressed
8. Irritated/angry/hostile

C. Parenting Style: Macroanalytic bi-polar descriptors reflect the global impression
that the coder feels best describes the mother‘s parenting style. Coders will choose one
descriptor for each set of eight.

1. Authority figure vs. friend
2. Lenient vs. strict
3. Irritable vs. easy-going
4. Inconsistent vs. consistent
5. Affectionate vs. unaffectionate
6. Unfair vs. fair
7. Serious vs. playful
8. Reasons/explains vs. makes rules/gives commands

99

Appendix E
Intra-class Correlation Scores for Coding System Items

The event-based items are listed with their intra-class correlations, using a two-way
mixed model of average measures. Items which were deleted due to extreme low
variance and skewness are indicated.

VERBAL 1 ICC
Item
Selection
Control/Motivation 1
1 Positive Character Attribution of Child 0 deleted
2 Negative Character Attribution of Child 0.747 deleted
3 Positive Reinforcement/Approval 0.738
4 Negative Reinforcement/Disapproval 0.98
5 Explanations/Reasoning 0.876
6 Prohibitions 0.957
7 Bargain/Bribe 0 deleted
8 Threaten 0 deleted
Teaching 2
1 Attention Directing -0.053
2 Instruction/Information Giving--concurrent 0.934
3 Information Seeking/Questioning 0.904
Other 3
1 Positive Self/Dyad Attribution 0.25
2 Negative Self/Dyad Attribution 0.206
3 Positive Task Attribution 0
4 Negative Task Attribution 0.821
5 Non-Instructional Task-Related Conversation 0.677
6 Off-Task Conversation which doesn't Interfere with Task -1.25 deleted
7 Diverting/Avoiding 0 deleted
8 Refusal/Delay 0 deleted
9 Responding to Child -1.034
NONVERBAL 2
Control/Motivation 2
1 Restraining 0 deleted
2 Physical Punishment/Physical Aggression 0.37 deleted
3 Nurturance 0.292
Other 3
1 Diverting/Off-task 0 deleted
2 Ignoring/Guilt/Silent Treatment -0.519
3 Laughter 0.919

100

The global items are listed with their intra-class correlations, using a two-way mixed
model of average measures. Items which were deleted due to extreme low variance and
skewness are indicated.

Style of Interaction (1-5)
1. Light-hearted/approaches with humor/flexible 0.800
2. Pleasant/happy/joyful 0.907
3. Apathetic/passive/flat affect 0 deleted
4. Timid/anxious/afraid/preoccupied 0 deleted
5. Irritated/anxious/hostile 0.876
6. Who was leading the task:
0.966
Parenting Style: (1-5)
1. Authority figure vs. friend 0.856
2. Lenient vs. strict 0.900
3. Irritable vs. easy-going 0.893
4. Inconsistent vs. consistent -0.400
5. Affectionate vs. unaffectionate 0.659
6. Unfair vs. fair 0.690
7. Serious vs. playful 0.768
8. Enjoys child vs. doesn't enjoy child 0.286
9. Encourages autonomy vs. not encourage autonomy 0.564
10. Reasons/explains vs. makes rules/gives commands 0.916
11. Warm toward child vs. withholding/chilly 0.636
12. Shy/timid vs. outgoing/not shy 0.653

Abstract (if available)

Abstract The aims of this study were a) to develop a coding system for parenting behaviors based on the domains of warmth and control, b) to extract meaningful constructs from both observational and self-report data, and to examine the correspondence between them, and c) to examine the impact of childhood sexual abuse on aspects of parental warmth and control. The sample was comprised of one hundred and twenty-seven mothers whose daughters were subjects in a longitudinal study in which half of the daughters, aged 6-16, were victims of familial sexual abuse, and a demographically similar comparison group.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Parenting practices among non-offending mothers of sexually abused girls and its impact on the abused girls' behavioral adjustment: perspectives from a multigenerational, longitudinal study

PDF

Male reports of abuse in romantic relationships on the rise

PDF

Trajectories of internalizing problems among maltreated girls and boys: differences by maltreatment type and developmental timing

PDF

The impact of child maltreatment on the mental health of adolescents: a longitudinal study of social anxiety and self-perception

PDF

Examining the longitudinal relationships between community violence exposure and aggressive behavior among a sample of maltreated and non-maltreated adolescents

PDF

Untangling the developmental relations between depression and externalizing behavior among maltreated adolescents

PDF

Gendered issues in alcohol abuse and dependence among HIV-positive African Americans

PDF

Mediators and moderators in the link between maternal psychological control and peer victimization for Hong Kong Chinese boys

PDF

The relationship among the nurturance and monitoring dimensions of parenting, academic self-concept, and acculturation, in the academic achievement of Latino college students

PDF

Trauma-related treatment gains among women with histories of interpersonal violence and co-occurring mental health and substance abuse disorders

PDF

Marital conflict and parenting: moderating effects of ethnicity and parent sex

PDF

Family aggression, prosocial friends, and the risk of dating and friend victimization in late adolescence

PDF

The link between maternal depression and adolescent daughters' risk behavior: the mediating and moderating role of family

PDF

Spouse aggression, depression, and physical health: a multivariate longitudinal study of midlife couples

PDF

The development of electrodermal fear conditioning from ages 3 to 8 years and relationships with antisocial behavior at age 8 years

PDF

Social status, perceived social reputations, and perceived dyadic relationships in early adolescence

PDF

Complicated relationships: the prevalence and correlates of sexual assault, dating violence victimization, and minority stress among sexual minority adolescents throughout the United States

PDF

Prenatal predictors of parental sensitivity in first-time mothers and fathers

PDF

Understanding allegations of childhood neglect using structured and unstructured administrative data

PDF

Dynamic analyses of the interrelationship between mothers and daughters on a measure of depressive symptoms

Asset Metadata

Creator Bryson, Christine Payne (author)

Core Title Observed and self-reported childrearing in mothers with a history of sexual abuse

School School of Social Work

Degree Doctor of Philosophy

Degree Program Social Work

Publication Date 07/25/2007

Defense Date 05/29/2007

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

Tag childrearing,Control,OAI-PMH Harvest,observational,Parenting,Sexual Abuse,warmth

Language English

Advisor Trickett, Penelope K. (committee chair), Margolin, Gayla (committee member), Mennen, Ferol E. (committee member)

Creator Email cbryson@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m676

Unique identifier UC1127244

Identifier etd-Bryson-20070725 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-521484 (legacy record id),usctheses-m676 (legacy record id)

Legacy Identifier etd-Bryson-20070725.pdf

Dmrecord 521484

Document Type Dissertation

Rights Bryson, Christine Payne

Type texts

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

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu