Introduction

Coping has been defined by Skinner and Wellborn [1] as the way, “people regulate their behavior, emotion, and orientation under conditions of psychological stress” (p. 112). While there are numerous types of coping strategies [2], two dimensions that have been supported in the empirical literature are problem- versus emotion-focused coping and engagement (approach) versus disengagement (avoidance) coping [3]. The first dimension focuses on acting on the cause of the stress in the situation (problem coping) compared to acting on the negative feelings that emerge from a stressful situation (emotion-focused coping); engagement coping strategies (e.g., seeking social support) involve reactions to stressors that are focused towards the cause of the stressor or an individual’s feelings or thoughts whereas disengagement coping (e.g., withdrawal, denial) involves strategies that shift an individual away from the stressor or their feelings or thoughts [3].

Emotional eating is defined as overeating in response to negative emotional states [4, 5]. According to psychosomatic theory [6], emotional eating serves as a non-adaptive coping strategy to temporarily alleviate distress [7] and over time can lead to increased weight gain, problems with weight loss, and obesity [8]. Consistent with this theory, a study of both eating-disordered and healthy women from the community found reliance on emotion-focused and avoidance-distraction coping styles was related to higher emotional eating [9]. In another study of adult females, experiential avoidance mediated the association between negative emotions and emotional eating [10]. Emotion-focused coping mediated the relationship between unsupportive social interactions and emotional eating in a sample of female university students [11]. Taken together, these studies indicate avoidant and emotion-focused coping strategies are associated with emotional eating in adults.

Despite data that suggest childhood stressors are correlated with emotional eating in childhood and adulthood [12, 13], little is known about coping strategies used by youth that may place them at risk for developing emotional eating. Studies with adolescents indicate deficient emotion regulation [14] and avoidant coping strategies are associated with eating disorder risk and binge eating [15,16,17]. However, these studies did not specifically examine emotional eating. It is important to understand whether certain coping strategies are associated with emotional eating in adolescents because emotional eating has been found in non-clinical adolescent populations [18] and may intensify to more problematic disordered eating if left untreated in this population [19]. Thus, elucidating coping strategies associated with emotional eating in adolescents may facilitate in the identification of youth at risk for developing more severe disordered eating and inform interventions aimed at the prevention of eating disorders.

In a subset of the present sample, we investigated factorial invariance of the Emotional Eating Scale (EES) Adapted for Children and Adolescents across adolescent males and females [20]. Based on our finding that the EES largely measured the same construct across adolescent males and females [20], in the present study we sought to assess whether coping styles moderated the relationship between perceived stress and emotional eating (as measured by the EES) in a combined sample of male and female adolescents. Adolescence has been described as a susceptible developmental period for stress given the numerous physical, physiological, and emotional changes [21]. Based on psychosomatic theory [6], and previous research [9, 22,23,24], we hypothesized (1) the association between perceived stress and emotional eating would be stronger for adolescents who endorse avoidance coping styles, (2) the association between perceived stress and emotional eating would be stronger for adolescents who endorse distraction coping styles, (3) the association between perceived stress and emotional eating would be weaker for adolescents who endorse more active coping styles, and (4) the association between perceived stress and emotional eating would be weaker for adolescents who endorse support-seeking coping styles.

Methods

Participants

The study sample comprised 277 eighth-grade students attending a middle school in the Southern United States. Demographic characteristics of the sample are summarized in Table 1.

Table 1 Demographic characteristics of study participants
Full size table

Measures

Demographic questionnaire

Students completed a demographic questionnaire constructed to gather information about their gender, age, race/ethnicity, parent relationship status, and parents’ highest educational attainment.

Body mass index (BMI)

Research assistants individually measured weight and height for each student. Weight was measured with a Tanita scale (model BF680 W) and height was taken with a measuring tape mounted to the wall. The measurements were performed after the removal of shoes, hats, and heavy clothing (e.g., jackets). Both weight and height were recorded on the student’s demographic questionnaire.

BMI was computed by entering the student’s height, weight, age, and sex into CDC pediatric growth charts for children between the ages of 2 and 20. After BMI was computed, the BMI percentile and z-score (standard deviation) were calculated. Each BMI number was plotted on the CDC BMI-for-age growth charts (for either girls or boys) to obtain a percentile ranking. The BMI percentile and z-score indicate the student’s BMI placement compared to same age and sex peers.

Stress

The Perceived Stress Scale (PSS-10) was used to assess the degree of global perceived stress [25]. The PSS-10 is a 10-item self-report questionnaire (e.g., how often have you been upset because of something that happened unexpectedly, how often have you felt nervous and stressed, how often have you felt difficulties were piling up so high that you could not overcome them) congruent with a transactional model of stress to assess feelings of being overwhelmed, ability to control events, or predictability of events in one’s life. Respondents are asked to report their feelings and thoughts during the last month by rating each item in one of five ways from never, almost never, sometimes, fairly often, to very often with scores of 0, 1, 2, 3, and 4, respectively. To score the PSS-10, positively stated items (items 4, 5, 7, and 8) are reversed scored. All 10 items are summed together to yield a total perceived stress score. Total perceived stress scores range from 0 to 40, with higher scores indicating higher perceived stress. Researchers have used this measure for older children and adolescents [26, 27]. Studies have reported a coefficient alpha ranging from .87 to .89 for early adolescents [18, 26, 28]. In this study, the PSS exhibited an acceptable internal consistency reliability (α = 0.84).

Coping

The Children’s Coping Strategies Checklist-Revision 1 (CCSC-R1) is a 54-item self-report inventory for children to rate how they typically cope with problems in their daily lives (e.g., you tried to stay away from the problem, you talked to someone who could help you figure out what to do, you did not think about it). Items are classified into 14 subscales that form four dimensions of children’s coping strategies: active coping, distraction, avoidance, and support seeking [29, 30].

Active coping consists of using problem focused (cognitive decision making, direct problem solving, and seeking understanding subscales) and positive reframing coping (positive thinking, optimistic thinking, control, and minimization subscales). Distraction coping consists of physical release of emotions and use of distracting actions. Avoidance coping consists of avoidant actions, repression, and wishful thinking. Support seeking coping includes support for actions and feelings. Children are asked to describe how often they usually use each behavior when they have a problem (never = 1, sometimes = 2, often = 3, and most of the time = 4). Scoring for the four major factors of coping is reached by taking the mean of the subscale scores for the subscales/dimensions that comprise that factor. Adequate reliability and validity have been previously documented in the literature [31, 32]. The CCSC-R1 was theoretically derived and the factor structure has been supported by earlier confirmatory factor analytic studies [29, 33]. In the present study, Cronbach’s alphas for the dimensions were as follows: 0.90 for active coping, 0.60 for distraction strategies, 0.70 for avoidance strategies, and 0.90 for support seeking.

Emotional eating

The Emotional Eating Scale for Children and Adolescents (EES-C) is a self-report questionnaire that has been adapted from the EES for adults [34]. The EES-C has been utilized with 8–17-year-olds to assess the urge to cope with negative affect by consuming food [34, 35]. The EES-C consists of 25 items across three subscales reflecting the desire to eat in response to: (1) anger, anxiety, and frustration (EES-C-AAF, 12 items), (2) depressive symptoms (EES-C-DEP, 7 items), and (3) feeling unsettled (EES-C-UNS, 4 items). Questions are presented on a five-point Likert scale for each emotion, with responses ranging from “I have no desire to eat”, “I have a small desire to eat”, “I have a moderate desire to eat”, “I have a strong desire to eat”, and “I have a very strong desire to eat” and scores ranging from 0, 1, 2, 3, and 4, respectively. The three subscales on the EES-C are calculated by totaling the items on each respective scale, while the total EES-C score is calculated by adding all of the items together. Greater total EES-C scores indicate a stronger desire to eat in response to a negative mood.

The EES-C-AAF, EES-C-DEP, and EES-C-UNS subscales and the EES-C total score have demonstrated strong internal consistent reliability throughout previous studies [34, 35]. The EES-C exhibited good convergent validity, evidenced by children reporting loss of control during eating episodes scoring higher on all EES-C subscales in a study conducted by Vannucci et al. [35]. The EES-C-AAF and EES-C-UNS subscales have demonstrated good discriminant validity, while the EES-C-DEP displayed adequate discriminant validity with general measures of psychopathology in the original study of the EES-C [34]. In the present study, the EES-C-AAF, EES-C-DEP, and EES-C-UNS subscales and the EES-C total score demonstrated good internal consistency reliability (αs = 0.90, 0.81, 0.71 and 0.93, respectively).

Procedures

Students were recruited through a health and lifestyle class at their middle school. They were considered eligible for the study if they were enrolled in the middle school and able to provide self-report responses on questionnaires in English. Exclusion criteria included students enrolled in special education programs or services.

Recruitment took place over 3 days. On the first day, the research assistant introduced the study, distributed an information packet, and provided time to answer any questions or concerns. Five hundred and forty-seven students were sent home with an information packet, including information about the study’s purpose and procedures, written informed consent, and written informed assent forms. Parents were instructed to review the consent and assent forms, which were approved by the Baylor University Institutional Review Board and the School District. Two days following the introductory visit, the research assistant returned to the middle school to collect the returned and signed consent and assent forms. On a subsequent day, the surveys were administered to the 277 students that returned their consent and assent forms and were willing to participate in the study (52 students who originally received packets were absent that day). Participating adolescents were asked to complete paper-and-pencil versions of the demographic questionnaire, the Perceived Stress Scale (PSS), the Children’s Coping Strategies Checklist-Revision 1 (CCSC-R1), and the Emotional Eating Scale for Children and Adolescents (EES-C) in a classroom. After the youth completed the self-report measures, a research assistant measured height and weight in a private room to calculate BMI. For all participants, study materials were de-identified. Each participating adolescent received an entry into a drawing for one of five $10 gift cards to Walmart upon completion of data collection procedures.

Statistical analysis

Multiple linear regression analyses were conducted to determine if coping styles moderate the relationship between stress and emotional eating. For each analysis, the assumptions of linearity, independence of errors, homoscedasticity, unusual points and normality of residuals were met. Correlational analyses between the variables were conducted to assess for multicollinearity, which indicated many of the variables were collinear (Table 2). To reduce potential problems with multicollinearity, stress and coping style variables were centered prior to conducting the analyses [36]. Centering entails subtracting the sample mean value from each adolescent’s score on the variable to form deviation scores with a sample mean of zero. Centering scores has no adverse effect on the correlation among variables and yields meaningful interpretations of the relation of predictors to dependent variables [36]. Following being centered, four interaction terms were created by multiplying the centered total stress score by each centered coping style (avoidance coping, distraction coping, active coping, support seeking coping).

Table 2 Intercorrelations among stress, coping, and Emotional Eating Scores
Full size table

Demographic variables including gender, race/ethnicity, highest level of education attained by parent, parental relationship status, and BMI were entered into the first block as control variables. The second block comprised the centered stress scores. The third block comprised the four-centered coping style scores (avoidance coping, distraction coping, active coping, and support seeking coping). Finally, the fourth block consisted of the interaction terms for centered stress with the four-centered coping styles (avoidance coping, distraction coping, active coping, and support seeking coping). A total of four regression analyses were conducted, including one with the total emotional eating score (EES-C) as the dependent variable and three with the emotional eating subscales (anxiety, anger, frustration subscale (EES-C-AAF); depression subscale (EES-C-DEP); unsettled subscale (EES-C-UNS) as the dependent variables. Table 3 presents the standardized beta coefficients for each variable comprising the models, with the presence of a statistically significant interaction determined by observing R 2 and results of the hierarchical regression F test of the step containing the interaction term. All statistical analyses were conducted in SPSS, Version 23.

Table 3 Multiple linear regression analyses
Full size table

Results

Emotional Eating Scale for Children and Adolescents: Total Score (EES-C)

The first multiple linear regression analysis had the EES-C total score as the dependent variable. In the first step, demographic variables did not account for a significant amount of variance in the EES-C total score, R 2 = .020, F(5, 242) = 1.013, p = .411. Analyses revealed that the second step with demographics and the addition of the centered perceived stress scores significantly predicted EES-C total score, R 2 = .057, F(6, 241) = 2.424, p = .027. The third step with demographic variables, centered perceived stress scores, and the addition of centered coping style scores did significantly predict EES-C total scores, R 2 = .076, F(10, 237) = 1.938, p = .041, respectively. The fourth step with interaction term of centered perceived stress scores and centered coping style scores did not account for a significant amount of variance in the EES-C total score, with all variables accounting for only 8.7% of the variance (R 2 = .087, F(14, 233) = 1.586, p = .084).

Emotional Eating Scale for Children and Adolescents: Depression subscale (EES-C-DEP)

The second multiple linear regression analysis had the EES-C-DEP subscale score as the dependent variable. In the first step, demographic variables did not account for a significant amount of variance in the EES-C-DEP subscale score, R 2 = .029, F(5, 242) = 1.463, p = .203. Analyses revealed that the second step with demographics and the addition of the centered perceived stress scores did significantly predict EES-C-DEP subscale score, R 2 = .083, F(6, 241) = 3.646, p = .002. The third step with demographic variables, centered perceived stress scores, and the addition of centered coping style scores did significantly predict EES-C-DEP subscale scores, R 2 = .114, F(10, 237) = 3.062, p = .001, respectively. The fourth step with interaction term of centered perceived stress scores and centered coping style scores did account for a significant amount of variance in the EES-C-DEP subscale score, with all variables accounting for 13.5% of the variance (R 2 = .135, F(14, 233) = 2.595, p = .002). In the fourth step, the standardized beta coefficient for the interaction term between avoidant coping style and perceived stress was significantly associated with depressive emotional eating (β = −.109, p = .014).

Emotional Eating Scale for Children and Adolescents: Anxiety, Anger Frustration Subscale (EES-C-AAF)

The third multiple linear regression analysis had the EES-C-AAF subscale score as the dependent variable. In the first step, demographic variables did not account for a significant amount of variance in the EES-C-AAF subscale score, R 2 = .021, F(5, 242) = 1.005, p = .386. Analyses revealed that the second step with demographics and the addition of the centered perceived stress scores did not significantly predict EES-C-AAF subscale score, R 2 = .041, F(6, 241) = 1.710, p = .119. The third step with demographic variables, centered perceived stress scores, and the addition of centered coping style scores did not significantly predict EES-C-AAF subscale scores, R 2 = .060, F(10, 237) = 1.518, p = .133, respectively. The fourth step with interaction term of centered perceived stress scores and centered coping style scores did not account for a significant amount of variance in the EES-C-AAF subscale score, with all variables accounting for only 6.5% of the variance (R 2 = .065, F(14, 233) = 1.156, p = .311).

Emotional Eating Scale for Children and Adolescents: Unsettled Subscale (EES-C-UNS)

The fourth multiple linear regression analysis had the EES-C-UNS subscale score as the dependent variable. In the first step, demographic variables did not account for a significant amount of variance in the EES-C-UNS subscale score, R 2 = .029, F(5, 242) = 1.460, p = .204. Analyses revealed that the second step with demographics and the addition of the centered perceived stress scores did significantly predict EES-C-UNS subscale score, R 2 = .055, F(6, 241) = 2.354, p = .031. The third step with demographic variables, centered perceived stress scores, and the addition of centered coping style scores did not significantly predict EES-C-UNS subscale scores, R 2 = .071, F(10, 237) = 1.818, p = .058, respectively. The fourth step with interaction term of centered perceived stress scores and centered coping style scores did not account for a significant amount of variance in the EES-C-UNS subscale score, with all variables accounting for only 7.8% of the variance (R 2 = .078, F(14, 233) = 1.417 p = .146).

Discussion

The objective of the present study was to assess coping styles as a moderator of the relationship between perceived stress and emotional eating. To the best of our knowledge, this is the first empirical study to assess these relationships in a community sample of adolescents. Our hypotheses were partially supported in that at higher levels of perceived stress, an avoidant coping style increased an adolescent’s propensity for depressive emotional eating (EES-C-DEP). This result corresponds to past research findings, which have indicated avoidant coping styles were associated with increased binge eating and emotional eating in adults across various weight categories [9, 22, 24]. Similar findings have been observed for avoidant coping with other disturbed eating patterns and eating disorders in youth. Avoidant coping in adolescents has been related to unhealthier eating behaviors such as skipping meals or watching television while eating [23], higher loss of control eating [37], and increased binge eating [17]. Since the interaction term between stress and avoidant coping was only significant for the depressive emotional eating model, it is possible that avoidant coping styles were predominantly utilized in response to depressive negative emotions captured on the EES-C (e.g., down, not doing enough, sad, discouraged). One study that assessed binge eating in youth found similar results, with individuals endorsing depressive emotions having the expectation that eating reduces their negative emotions [38]. While the present study assessed avoidant coping, one study found decentering (an emotion regulation process that allows one to distance the self from aspects of conscious thoughts and emotions that are potentially threatening to one’s identity) attenuated the effect of weight/body shape and shame on disordered eating in adult women [39]. It will be important for future studies with youth to examine what role decentering plays in moderating the relationship between perceived stress and emotional eating.

The interaction between perceived stress and active coping style did not account for a significant amount of variance in the prediction of emotional eating. Results of the present study contrasted with some previous findings in the active coping literature. In a female college sample, it was found that positive reframing of thoughts prevented body comparison and displeasure [40]. Previous studies have found that the adolescents reporting binge eating used positive strategies less frequently [17, 41], with similar results for emotional eating [42]. Disparities with the present findings may be attributed to other studies surveying Dutch and Spanish adolescents [41, 42], using the Dutch Eating Behavior Questionnaire opposed to the EES-C for emotional eating, or measuring binge eating instead of emotional eating.

The interaction between perceived stress and social support coping style did not account for a significant amount of variance in the prediction of emotional eating in the present study. Contrary to these findings, some adult studies have reported unsupportive social interactions were associated with emotional eating [11] and less social support was associated with binge eating [22]. An adult study with multiple mediation analyses indicated that emotion-focused coping mediated the relation between unsupportive social interactions and emotional eating [11]. Social support has been shown to have a direct impact on health behavior in youth [43], including dietary practices [44]. One study that investigated social coping in adolescent girls found low scores on assertiveness and social coping were associated with increased dieting [45]. A few important distinctions may have accounted for the different findings. Some studies used the variable of social support or interactions [11, 43, 44], not social coping as utilized in the present study. Additionally, mediation models were tested, such as one study testing if emotion-focused mediated the relationship between unsupportive social interactions and emotional eating [11]. Another possible explanation suggested by researchers is that social support may be more greatly connected with positive experiences over stressful ones, such that adolescents equate social support with feeling happy and being entertained [23].

Findings from the present study have clinical implications for healthcare providers who work with youth. Adolescence is associated with an increased risk of disordered eating and represents an ideal time for early intervention [46]. The present study provides evidence that emotional eating patterns are present even in a non-clinical sample of adolescents. At higher levels of perceived stress, an avoidant coping style may increase an adolescent’s propensity for depressive emotional eating. These findings provide preliminary support for targeting an avoidant coping style in preventative interventions, particularly for youth that have the propensity to overeat in response to feelings of depression. Creating interventions and increasing accessibility of interventions that target avoidant coping styles and teach healthy alternatives for emotional dysregulation in adolescence may prevent higher levels of emotional eating and more extreme forms of disordered eating behaviors extending into adulthood [18, 47, 48].

There are several limitations to the present study. Considering the present sample was restricted to 8th grade students at one middle school in the Southern United States, the findings may have limited generalizability. Generalizability is further limited by the fact that adolescents and their parents selected themselves to participate in the study. A parent with more concern about their adolescent’s physical and mental health may have been less likely to sign the consent form and allow their adolescent to participate, which limits external validity. We also did not have any measure of family income in the present study. Another limitation of the present study was results were based solely on self-report measures of stress, coping, and emotional eating. Social desirability and response bias may have influenced youth during the completion of the measures in a health and physical education class. Finally, the cross-sectional nature of this study prevents us from inferring causation. Future longitudinal studies are necessary to examine what predictive role coping styles play in moderating the relationship between stress and emotional eating in adolescents.