Childhood emotional abuse and disordered eating among undergraduate females: Mediating influence of alexithymia and distress☆
Introduction
A plethora of research within the last two decades has contributed to the understanding of the etiology and maintenance of eating disorders. Both anorexia and bulimia are thought to emerge from multiple risk factors, including biological, sociological, and psychological indices (Striegel-Moore & Cachelin, 2001). The role of negative family environments has received considerable attention as an important risk factor in the development of eating disorders among adolescents and young women (Leung, Schwartzman, & Steiger, 1996). Specifically, child abuse has been of interest to scholars and clinicians; however, research has focused predominantly on childhood sexual and physical abuse and has virtually ignored the potential influence of childhood emotional abuse (CEA). This current study explored the association between CEA and disordered eating (DE) among a sample of individuals at risk of developing eating disorders—female undergraduates. Drawing upon a stress-vulnerability theoretical and trauma theory framework (e.g., Rorty & Yager, 1996), the connection between CEA and DE was evaluated in a comprehensive multivariate model (Figure 1) using structural equation modeling in which depression, anxiety, and alexithymia were evaluated as mediators. In this study, CEA was defined as “verbal assaults on a child's sense of worth or well-being, or any humiliating, demeaning, or threatening behavior directed toward a child by an older person” (Bernstein & Fink, 1998, p. 2).
Eating disorders scholars have not assessed childhood emotional abuse directly because of their tendency to treat adverse family environment as a proxy of abuse (Kent & Waller, 2000). For example, early research on family factors of women with eating disorders described these environments as intrusive, overprotective, controlling for anorexics, and chaotic and emotionally cold for bulimics (Bruch, 1973). However, Kent, Waller, and Dagnan (1999) found that the primary predictor of disordered eating symptoms was CEA, after controlling for other forms of abuse. Given these findings, CEA should be examined more completely.
From a stress-vulnerability model of DE and trauma theory (e.g., Rorty & Yager, 1996), the experience of any form of child abuse is thought to lead to boundary violations and trust issues. As the abuse becomes persistent and recurrent, it is then associated with a decrease in a child's self-concept and self-esteem. According to this theory, this diminished sense of self then leads to difficulty in managing strong affect (e.g., alexithymia), contributing to the risk of general distress (GD) (e.g., depression, anxiety) and maladaptive coping strategies in adulthood (e.g., eating disorders; Follette, Ruzek, & Abueg, 1998).
Mazzeo and Espelage (2002) assessed CEA directly and tested aspects of this theory. One component these authors examined was alexithymia, which has been associated at a bivariate level with eating disorders (e.g., Taylor, Parker, Bagby, & Bourke, 1996) and child abuse (e.g., Berenbaum, 1996). Alexithymia has been defined as cognitive deficits in identifying and verbalizing emotions and an inability to distinguish between emotional and physical sensations (Nemiah & Sifneos, 1970). Furthermore, Heatherton and Baumeister (1991) argued that disordered eating might emerge as a means of distraction from negative thoughts and emotions associated with stressful experiences. Thus, they concluded that alexithymia could be a psychological mechanism by which individuals manage the vulnerability associated with experiencing verbal abuse and potentially humiliating statements from family members (Mazzeo & Espelage, 2002).
Mazzeo and Espelage (2002) and Kent et al. (1999) also considered the role of psychological distress in their models. Kent et al. found that anxiety mediated the association between CEA and DE among 236 undergraduate females; depression was not a significant mediator. These authors argued that the link between CEA and anxiety emerges from the insidious nature of emotional abuse in which the uncertainty of when it might happen could lead to perceptions of personal vulnerability. Individuals might then manage this anxiety through disordered eating behaviors. The influence of depression in the association between emotional abuse and eating disordered behavior is less clear (Kent et al., 1999; Mazzeo & Espelage, 2002). For instance, Mazzeo and Espelage (2002) found that when data were analyzed using structural equation modeling, rather than regression analyses (i.e., Kent et al., 1999), depression did mediate the relation in a sample of 406 undergraduates, as well as in a cross-validation sample of 406 additional participants.
These inconsistent findings might relate to the way in which depression and anxiety manifest in nonclinical samples and the way in which they are assessed. This has led many researchers to propose alternate perspectives to understand the relationship between anxiety and depression. Clark and Watson (1991) argued for a tripartite model for both anxiety and depression. Further studies (Watson et al., 1995) have provided support for a model in which anxiety and depression share a general distress component. Therefore, anxiety and depression were conceptualized as GD throughout this manuscript and in the models tested.
Unlike previous studies on CEA and eating disorders, this paper is the first to consider how psychological mediating variables predict both restrictive eating behaviors and bulimic symptoms. The decision to evaluate restrictive eating and bulimic symptoms as distinct latent variables in the model was based in the literature.
First, considerable debate has emerged about whether eating disorders exist along a continuum versus a discontinuum. Proponents of the continuum argument posit that eating habits range from completely normal eating behavior to a pattern of behaviors meeting DSM criteria for eating disorders (e.g., Tylka & Subich, 1999). Conversely, proponents of the discontinuum hypothesis argue that there is a qualitative difference between individuals with clinical eating disorders and those manifesting subclinical disordered eating behaviors (e.g., Gleaves, Lowe, Snow, Green, & Murphy-Eberenz, 2000). Recent taxometric analyses suggest that binge eating behaviors of those with bulimia nervosa and binge eating disorder are qualitatively different from the eating behaviors of nonclinical individuals (Williamson et al., 2002).
Second, longitudinal research clearly demonstrates that regardless of the continuous or discontinuous nature of these symptoms, there are identifiable factors that reflect subclinical levels of eating disorders that place an individual at risk of developing a clinical eating disorder (Killen et al., 1996; Stice, Killen, Hayward, & Taylor, 1998). These studies found that weight concerns, negative body image, and dieting are strong risk factors in the etiology of clinical eating disorders. Given the undergraduate sample included in this paper, it was hypothesized that there would be more participants with elevated levels of body dissatisfaction and dieting, and fewer with binge eating and bulimic symptoms. Thus, a latent variable consisting of restrictive eating and body dissatisfaction was evaluated as a mediator between GD and bulimic symptoms.
Because there is no comprehensive, empirically validated model of the association between CEA and DE, a structural equation model was tested in a nonclinical sample of college women in this study (Figure 1). This study improved upon previous investigations by using multiple indicators to assess CEA, employing structural equation modeling, and examining restrictive eating and body dissatisfaction as a mediator between psychological constructs and bulimic symptoms. In the conceptual model (as illustrated in Figure 1), we proposed that CEA would be related to greater levels of alexithymia (path a). Alexithymia was expected to be further associated with GD, including anxious and depressed mood represented by path b. In turn, GD was expected to be related to both restrictive and bulimic behaviors (paths c and d). Furthermore, using SEM instead of regression allowed us to hypothesize specific mediating associations. We also hypothesized that alexithymia would mediate the relation between CEA and GD; GD would mediate the relation between alexithymia and disordered eating; alexithymia and GD would mediate the relation between CEA and DE. Finally, the distinct constructs of restrictive eating and bulimic symptoms were expected to be significantly associated. In a nested model, the pathway between restrictive and bulimic behaviors (path e) was removed. Based on longitudinal findings, it was hypothesized that the model without path e would be a significantly poorer fit to the data than the original model.
Section snippets
Participants
Participants were 608 undergraduate and graduate women recruited from sororities and a departmental subject pool at a large Midwestern university. In addition to the measures described below, participants were asked to complete a demographic questionnaire. Participants represented the following racial/ethnic groups: 69.4% (n = 408) White, 15.0% (n = 88) African American, 8.2% (n = 48) Asian American/Asian, 7.0% (n = 41) Latina, and .3% (n = 2) Native American. Regarding year in school, 22.1% (n = 130) were
Descriptive statistics
Descriptive statistics included subscale means, standard deviations, and bivariate correlations among scales as reported in Table 1. All subscale means and standard deviations were comparable to norms for undergraduate nonclinical samples. All of the above subscales were deemed suitable for use in the modeling analyses.
Notably, anxiety and depression subscales correlated between .55 and .78, which is consistent with previous findings suggesting that these subscales measure overlapping
Discussion
Very little attention has been given to the long-term effects of CEA. Only a few studies have examined the link between emotional abuse and disordered eating. The current study was an attempt to work toward recommendations for creating a deeper understanding of the impact of emotional abuse. Support was found for an association between CEA and DE through a series of mediating relations. The present study adds to only one other rigorous study (Mazzeo & Espelage, 2002) in shedding light on the
Acknowledgements
We thank Drs. Helen Neville and James Rounds for their assistance with this project. We appreciate Drs. Howard Berenbaum's and Suzanne Mazzeo's willingness to share their knowledge. We also thank Ross Wantland for his assistance with data entry.
References (38)
- et al.
The twenty-item Toronto Alexithymia Scale-I: Item selection and cross-validation of the factor structure
Journal of Psychosomatic Research
(1994) Childhood abuse, alexithymia and personality disorder
Journal of Psychosomatic Research
(1996)- et al.
Childhood emotional abuse and eating psychopathology
Clinical Psychology Review
(2000) - et al.
The measurement of psychological maltreatment: Early data on the Child Abuse and Trauma Scale
Child Abuse & Neglect
(1995) - et al.
Relationships between alexithymia and psychological characteristics associated with eating disorders
Journal of Psychosomatic Research
(1996) - et al.
The moderator-mediator variable distinction in social psychological research: Conceptual, strategic, and statistical considerations
Journal of Personality and Social Psychology
(1986) - et al.
Childhood trauma questionnaire: A retrospective self-report: Manual
(1998) Structural equations with latent variable
(1989)Eating disorders: Obesity, anorexia nervosa, and the person within
(1973)- et al.
Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications
Journal of Abnormal Psychology
(1991)
The national women's study: Relationship of victimization and posttraumatic stress disorder to bulimia nervosa
International Journal of Eating Disorders
BSI: Administration, scoring, and procedures manual for the Brief Symptom Inventory
Examining the construct validity of the Eating Disorder Inventory
Psychological Assessment
A contextual analysis of trauma: Theoretical considerations
Eating Disorder Inventory – 2: Professional Manual
The eating attitudes test: Psychometric features and clinical correlates
Psychological Medicine
Continuity and discontinuity models of bulimia nervosa: A taxometric investigation
Journal of Abnormal Psychology
Binge eating as an escape from self-awareness
Psychological Bulletin
Fit indices in covariance structure modeling: Sensitivity to underparameterized model misspecification
Psychological Methods
Cited by (111)
Specific correlation between childhood trauma and social cognition in Chinese Han first-episode, drug-naïve major depressive disorder
2023, Journal of Affective DisordersThe mediating role of impulsivity in the associations between child maltreatment types and past month substance use
2022, Child Abuse and NeglectThe role of emotion dysregulation and alexithymia in the link between types of child abuse and neglect and psychopathology: A moderated mediation model
2021, European Journal of Trauma and DissociationCitation Excerpt :Research findings show that emotional abuse and neglect may play a role in the lacking ability to identify feelings (Brown, Fite, Stone, Richey, & Bortolato, 2018), meaning that childhood emotional abuse (Aust, Härtwig, Heuser, Bajbouj, & Gold, 2013), physical and emotional neglect (Minnich, Gordon, Kwan, Troop-Gordon, & Liu, 2017) have been linked to alexithymia. However, physical and sexual abuse have not been associated with alexithymic features (Hund & Espelage, 2006). Moreover, evidence show that child abuse and neglect have an effect on psychopathology through emotion dysregulation, with higher levels of abuse/neglect being associated with emotion dysregulation than in turn is being associated with higher levels of psychopathology (Jennissen, Holl, Mai, Wolff, & Barnow, 2016; Mclaughlin, Hatzenbuehler, & Hilt, 2009; Moulton, Newman, Power, Swanson, & Day, 2015; Titelius et al., 2018).
- ☆
This research was supported in part by a University of Illinois Graduate College Master's Thesis Grant. A version of this article was presented at the 112th Annual Convention of the American Psychological Association, Honolulu, Hawaii, August 2004.