Negative Affect as a Mediator of the Relationship between Weight-Based Teasing and Binge Eating in Adolescent Girls

Abstract

Previous research has established a link between weight-based teasing and binge eating, though the precise mechanisms that drive this relationship remain unknown. This study examined negative affect as a mediator of the relationship between weight-based teasing and binge eating. Participants included 265 adolescent female twins (aged 10–15 years). Self-report measures assessed binge eating, weight-based teasing, and negative affect. Mediation was tested within hierarchical linear models to control for the non-independence of the twin data. Significant positive associations were observed between binge eating, teasing, and negative affect. In the regression analyses, negative affect partially mediated associations between weight-based teasing and binge eating. Results suggest that increases in negative affect are one way in which weight-based teasing leads to binge eating in girls. Future studies should examine additional mediators and assess possible clinical applications of these findings.

1. Introduction

Weight-based teasing has frequently been found to be a precursor to binge eating (e.g. Thompson, Cattarin, Fowler, & Fisher, 1995). Nonetheless, it remains unknown whether there are direct relationships between weight-based teasing and binge eating, or whether these relationships are mediated by other variables. The aim of the present study was to examine the nature of weight-based teasing and binge eating relationships by investigating negative affect as one possible mediator of this association. Negative affect is one of the strongest longitudinal predictors of binge eating (e.g. Stice, 1998) and has been classified as a nonspecific “risk factor” for eating pathology in two separate meta-analyses (Jacobi, Hayward, de Zwaan, Kraemer, & Agras, 2004; Stice, 2002). Negative affect is also associated with weight-based teasing (e.g. Eisenberg, Neumark-Sztainer, & Story, 2003). Thus, negative affect may serve as a mediator of the association between weight-based teasing and binge eating, such that weight-based teasing increases negative affect which, in turn, increases binge eating (see Figure 1). Although researchers have previously called for the need to examine these mediating relationships (Haines & Neumark-Sztainer, 2006), no previous studies have examined this mediation model.

Figure 1.

Negative Affect as a Mediator of Weight-Teasing and Binge Eating

2. Methods

2.1 Participants

Data come from the Michigan State University Twin Registry (MSUTR; see Klump & Burt, 2006 for study recruitment details). The current study examined a total of 265 MSUTR female twins aged 10–15 (M = 12.5; SD 1.4). One of the primary aims of the projects from which data were drawn was to examine disordered eating/steroid hormone relationships. Therefore, exclusion criteria included medication use or medical conditions that could affect steroid hormone levels. The majority of participants in the current study were Caucasian (87%), while the remainder identified themselves as African American (8%) or other ethnicity (5%).

2.2 Measures

2.21 Binge Eating

The binge eating subscale of the Minnesota Eating Behavior Survey (MEBS; Klump, McGue, & Iacono, 2000)¹.

2.22 Teasing

The weight-based teasing subscale of Perception of Teasing Scale (POTS; Thompson, Cattarin, Fowler, & Fisher, 1995). Participants were asked to consider teasing that occurred while growing up from age 5 through the present.

2.23 Negative Affect

The Early Adolescent Temperament Questionnaire-Revised (EATQ-R; Ellis & Rothbart, 2001), was used to examine negative affect.

2.24 Pubertal Development

The Pubertal Development Scale (PDS; Petersen, Crockett, Richards, & Boxer, 1988) was used to assess overall pubertal development.

2.25 Body Mass Index

Calculated (Weight in kg / Height in m²) from height and weight measurements obtained using a wall-mounted ruler and digital scale, respectively.

2.3 Statistical Analyses

Binge eating and BMI were log transformed to account for positive skew. Arctan transformations were used to account for positive skew and kurtosis for weight-based teasing. Pearson product correlations were used to examine initial relationships between all variables. The mediation model was then tested using methods developed by Baron and Kenny (1986) and regression analyses implemented within hierarchical linear models (HLM) to account for the non-independence of the twin data. BMI, age, and pubertal development, were included as covariates in all analyses.

3. Results

Expected positive associations were found between binge eating, weight-based teasing and negative affect (see Table 1). In steps one and two of the mediation analyses, weight-based teasing significantly predicted both binge eating and negative affect. Negative affect predicted binge eating in step three, even when controlling for weight-based teasing. In the final step, we evaluated negative affect as a mediator of the association between weight-based teasing and binge eating. Results revealed that the relationship between weight-based teasing and binge eating decreased but remained statistically significant (β=1.6, p=.006), suggesting that negative affect partially mediates associations between weight-teasing and binge eating. Sobel tests (Baron & Kenny, 1986) confirm that negative affect significantly carries the effect of weight-based teasing on binge eating (Z=4.68; p<.0001). Importantly, the final step of the model accounted for one-fifth of the variance in binge eating (pseudo R²=.20; Kenny, Kashy, & Cook, 2006).

Table 1.

Pearson Correlations

Variable	Binge Eating	Weight-Based Teasing	Negative Affect	BMI	Age	Pubertal Development
Binge Eating	1.0	--	--	--	--	--
Weight-Based Teasing	.30**	1.0	--	--	--	--
Negative Affect	.42**	.37**	1.0	--	--	--
BMI	.07	.48**	.11	1.0	--	--
Age	−.02	−.07	.08	.22**	1.0	--
Pubertal Development	−.03	.07	.15*	.39**	.67**	1.0

Note. BMI = body mass index.

^*p<.05,

^**p<.001.

4. Conclusions

The current study extended prior research by evaluating whether negative affect mediated the well-documented association between weight-based teasing and binge eating. Formal tests of mediation suggested that being teased about one’s weight may increase negative affect, which in turn, leads to binge eating. These findings could not be explained by differences in BMI, pubertal status, or age. Importantly, the direct effect of weight-based teasing, and the mediating effects of negative affect, were moderate-to-large (Table 2), suggesting that weight-based teasing may be an important stressor that increases negative affect and begins the chain of events that lead to binge eating.

Table 2.

Hierarchical Linear Model Examining Negative Affect as a Mediator of the Relationship between Weight-Teasing and Binge Eating

Model	β (SE)	t	df	p
Step 1: WT → BE (pseudo R2 = .09)	3.04 (.57)	5.32	250	< .001
Covariates
BMI	−.27 (.25)	−1.09	224	.28
Age	.01 (.02)	.720	191	.47
Pubertal Development	−.005 (.01)	−.85	259	.40
Step 2: WT → NA (pseudo R2 = .17)	139 (19)	19.5	233	< .001
Covariates
BMI	−19.4 (9)	−2.15	241	.03
Age	−.52 (.58)	.886	193	.38
Pubertal Development	.34 (.19)	1.78	259	<.08
Step 3: NA → BE (pseudo R2 = .20)a	.01 (.002)	6.2	254	< .001
Covariatesa
BMI	−.16 (.23)	−.24	229	.81
Age	.01 (.01)	.411	190	.68
Pubertal Development	−.01 (.01)	−1.6	258	.11
Step 4: WT and NA → BE (pseudo R2 = .20)a	1.6 (.58)	2.76	245	< .01
Covariatesa
BMI	−.16 (.23)	−.24	229	.81
Age	.01 (.01)	.411	190	.68
Pubertal Development	−.01 (.01)	−1.6	258	.11

Note. WT=Weight Teasing Frequency Subscale of POTS; BE=Binge Eating Subscale of MEBS; BMI=Body Mass Index; NA=Negative Affect Subscale of EATQ. Calculations of pseudo R² were made with the following equation taken from (Kenny et al., 2006): $R^2=1-\frac{S_{dd}+{S_e}^2}{S_{dd}\prime +{S_e}^2\prime }.$

^aPseudo R² for steps three and four are identical because these parameters are calculated within the same equation.

Our study had limitations that must be acknowledged. First, findings may not generalize to clinical populations, as participants were not clinically diagnosed with BN or BED. Secondly, our cross sectional data did not allow us to infer a temporal relationship among these variables. For example, it could be hypothesized that teasing mediates the relationship between negative affect and binge eating. However, post-hoc analyses demonstrated that this is not the case, as associations between negative affect and binge eating remained unchanged when controlling for levels of teasing (data not shown). Nonetheless, future prospective and experimental studies should examine both models to determine direction of effects. Finally, we used self-report data to assess binge eating, teasing, and negative affect. Additional research using other assessment methods (e.g., interviews) and informants is needed to corroborate our results.

Nevertheless, this study improves understanding of possible mechanisms that drive the relationship between weight-based teasing and binge eating. Our findings suggest that weight-based teasing and negative affect may be important factors to target in intervention and prevention programs aimed at reducing binge eating, particularly in adolescent girls.