Skip to main content
NCBI home page
NIHPA Author Manuscripts logoLink to NIHPA Author Manuscripts
. Author manuscript; available in PMC: 2012 Jul 11.
Published in final edited form as: Int J Obes (Lond). 2011 Nov 29;36(7):938–943. doi: 10.1038/ijo.2011.225

Construct Validity of the Emotional Eating Scale Adapted for Children and Adolescents

Anna Vannucci 1,2, Marian Tanofsky-Kraff 1,2, Lauren B Shomaker 1,2, Lisa M Ranzenhofer 1,2, Brittany E Matheson 1,2, Omni L Cassidy 1,2, Jaclyn M Zocca 1, Merel Kozlosky 3, Susan Z Yanovski 4, Jack A Yanovski 1
PMCID: PMC3367077  NIHMSID: NIHMS359514  PMID: 22124451

Abstract

Background

Emotional eating, defined as eating in response to a range of negative emotions, is common in youth. Yet, there are few easily administered and well-validated methods to assess emotional eating in pediatric populations.

Objective

The current study tested the construct validity of the Emotional Eating Scale Adapted for Children and Adolescents (EES-C) by examining its relationship to observed emotional eating at laboratory test meals.

Method

One hundred fifty-one youth (8-18 years) participated in two multi-item lunch buffet meals on separate days. They ate ad libitum after being instructed to “eat as much as you would at a normal meal” or to “let yourself go and eat as much as you want.” State negative affect was assessed immediately prior to each meal. The EES-C was completed three months, on average, prior to the first test meal.

Results

Among youth with high EES-C total scores, but not low EES-C scores, higher pre-meal state negative affect was related to greater total energy intake at both meals, with and without the inclusion of age, race, sex, and BMI-z as covariates (ps < 0.03).

Discussion

The EES-C demonstrates good construct validity for children and adolescents’ observed energy intake across laboratory test meals designed to capture both normal and disinhibited eating. Future research is required to evaluate the construct validity of the EES-C in the natural environment and the predictive validity of the EES-C longitudinally.

Keywords: emotional eating, obesity, pediatrics, negative affect, energy intake

Emotional eating has been defined as an enduring behavioral pattern characterized by eating in response to a range of negative emotions [1]. Approximately 10 to 60% of children and adolescents report emotional eating, with higher estimates in adolescent samples [2-4], weight-loss treatment-seeking populations [5], and in youth who report other forms of disinhibited eating [6, 7]. Longitudinal data suggest that pediatric emotional eating is a stable appetitive trait that persists over time [8-10]. Among children and adolescents, emotional eating has been linked to the onset of binge eating episodes [11] and associated with greater eating disorder psychopathology, elevated depressive symptoms, and more internalizing and externalizing problem behaviors [2, 12]. Some [2, 12], but not all [7, 13, 14], studies have found that overweight youth endorse greater emotional eating than non-overweight youth. Despite the adverse psychosocial and physical correlates of emotional eating, few easily administered and validated methods are available to assess emotional eating in youth.

As previously published, we adapted the Emotional Eating Scale (EES) for adults to be used with children and adolescents (EES-C) [7]. An examination of its psychometric properties in a community sample of overweight and non-overweight youth suggested that the EES-C has good internal consistency, temporal stability, and convergent and discriminant validity [7]. However, the construct validity of the EES-C remains unknown. To date, there are no known studies that have examined the construct validity of any self-report measure of pediatric emotional eating.

Laboratory studies of emotional eating in the adult literature typically have involved the employment of negative mood induction paradigms, during which individuals are randomized to either a negative or neutral mood induction (e.g., using a brief film clip) and subsequently given access to food. Most adult studies have demonstrated null findings, such that there was no relationship between self-reports of emotional eating and energy intake in the laboratory following negative (vs. neutral) mood inductions [15, 16]. In the few studies that have reported a positive relationship, self-reported emotional eating was associated with energy intake following a negative mood induction, but only among a subset of participants who actually reported experiencing the intended negative affective state [17, 18]. Indeed, one of the drawbacks of mood induction paradigms is the high degree of inter-individual variability in response to laboratory stressors (e.g., individuals have different levels of emotional reactivity to a standardized negative mood paradigm) [19-21]. Therefore, an alternative approach to capturing “emotional eaters” is to examine the relationship of state negative affect immediately prior to eating with objective assessments of energy intake at a laboratory test meal, in the absence of an artificial mood induction.

In a secondary analysis of a study that investigated normal and disinhibited eating behaviors in youth [22], we aimed to test the construct validity of the EES-C by examining the associations among EES-C scores, pre-meal state negative affect, and measured energy consumption. Specifically, we hypothesized an interaction between the EES-C and pre-meal state negative affect, such that the association between pre-meal state negative affect and energy intake would be more positive among youth with high EES-C scores than among those with low EES-C scores. We expected to observe these associations independent of BMI-z and other relevant demographic covariates.

Methods

Participants

The present investigation is a secondary analysis of a previously published study [22]. Children and adolescents (8-18 years) were healthy volunteers participating in a non-intervention study at the National Institutes of Health (NIH) that investigated eating behaviors in overweight and non-overweight children and adolescents. Detailed recruitment methods are described elsewhere [22]. Briefly, boys and girls of any race or ethnicity who had a body mass index (BMI, kg/m2) at or above the 5th percentile for age and sex [23] were eligible for participation. Exclusion criteria included: 1) the presence of a significant medical condition; 2) abnormal hepatic, renal, or thyroid function; 3) use of medication known to impact body weight; 4) greater than 5 pounds (2.3 kilograms) of weight loss in the previous three months; 5) enrollment in a weight loss treatment program; or 6) the presence of a psychiatric disorder that might impede protocol compliance. Additionally, individuals were excluded if they reported disliking more than 50% of the foods to be offered at the test meals or if they were unable to acclimate to laboratory conditions (i.e., inability to consume at least 50% of a high-calorie shake during a laboratory screening visit that took place prior to the test meals). The latter exclusion criterion was used to confirm that youth could eat comfortably in the laboratory setting. Youth provided written assent and parents gave written consent for participation in the study. This study was approved by the Eunice Kennedy Shriver National Institute of Child Health and Human Development institutional review board and registered at ClinicalTrials.gov (NCT00320177). Data were collected between February 2004 and June 2009.

Procedures

Data were collected from each child during three outpatient visits, on different days, to the NIH Hatfield Clinical Research Center. All visits took place following an overnight fast.

Body Measurements

At the screening appointment, participants’ weight and height were measured using calibrated electronic instruments, as previously described [22]. BMI was calculated as weight in kilograms divided by the square of height in meters. BMI standard deviation (BMI-z) scores were calculated according to the Centers for Disease Control and Prevention 2000 growth charts [23]. Body composition was measured using air displacement plethysmography (Life Measurement Inc., Concord, CA) to determine fat-free mass (kilograms) and percent fat mass.

Assessment of Reported Emotional Eating

The Emotional Eating Scale Adapted for Children and Adolescents (EES-C) was administered during the screening visit, which occurred three months, on average, prior to the first test meal. The EES-C, designed for use with 8-18-year-old children [7], was adapted from the Emotional Eating Scale for adults [24]. The EES-C is a 25-item self-report measure used to assess the propensity to cope with negative affect by eating. Respondents rate their desire to eat in response to each emotion on a 5-point scale from “I have no desire to eat” through “I have a very strong desire to eat.” The EES-C generates three subscales reflecting the urge to eat in response to: 1) Anger, anxiety, and frustration, 2) Depressive symptoms, and 3) Feeling unsettled. Based on parameters indicated by Rosner [25], the EES-C has demonstrated very good internal consistency (αs = 0.83 – 0.95) and adequate temporal stability (rs = 0.59 – 0.74, ps < 0.001) [7]. The EES-C also had good convergent validity in that children reporting recent episodes of loss of control eating had higher EES-C scores (ps < 0.05). Finally, the EES-C demonstrated good discriminant validity, such that its subscales were unrelated to general measures of psychopathology [7].

Summing of the individual EES-C items generates an EES-C total score. Psychometric properties of the EES-C total score were examined for the current sample. Cronbach's alpha for the EES-C total score was 0.94. In addition to administration at the screening visit, the EES-C was also administered to participants up to one year prior to the screening visit; this allowed us to assess temporal stability for the current study. The EES-C total score had adequate temporal stability (r = 0.64, p < 0.001) [25]. The EES-C total score showed good convergent validity with loss of control eating (p < 0.01). The EES-C total score demonstrated good discriminant validity with measures of depressive symptoms, trait anxiety, and internalizing and externalizing problem behaviors (data not shown, but available upon request).

Assessment of State Negative Affect Prior to the Laboratory Test Meals

Immediately prior to each laboratory test meal, participants completed the well-validated Brunel Mood Scale (BRUMS) [26], which measures present mood state and generates six subscales pertaining to anger, confusion, depression, fatigue, tension, and vigor. As an overall measure of pre-meal state negative affect, a total BRUMS score was computed by summing individual items with the exception of items from the vigor subscale. The vigor subscale was not included because it reflects positive affect and does not correspond to the constructs assessed in the EES-C. Cronbach's alpha for the BRUMS total score in the current study was 0.83, indicating very good internal consistency.

Observed Intake During Laboratory Test Meals

Participants were asked to consume their lunch ad libitum from a multiple-item buffet test meal on two separate days, scheduled at least two weeks apart. On the day of each test meal, children and adolescents arrived in the morning after an overnight fast and were provided with a standard 280 kcal breakfast (7% protein, 19% fat, 74% carbohydrate). Participants remained at the NIH Clinical Center for the next six hours, during which they were observed to ensure that they consumed no calorie-containing foods or beverages. Children were allowed to participate only in sedentary activities. The laboratory test meals began at 2:30pm. Each participant was then presented with a multiple-item, 9,835-kcal food array varied in macronutrient composition (51% carbohydrate, 12% protein, 37% fat, across all foods) and containing a wide assortment of foods, including meal-type foods and snack- and dessert-type foods [28]. In random order, a normal eating episode or a disinhibited eating episode was modeled by providing participants with the pre-recorded instructions. Prior to the “normal meal”, children were instructed to “eat as much as you would at a normal meal.” Prior to the “disinhibited meal”, children were instructed to “let yourself go and eat as much as you want.” Other than the different instruction, all other aspects of the conditions were identical. The experimental design of the test meals was based upon the adult literature that has examined normal and binge eating behaviors in the laboratory setting [27]. Youth were then given unrestricted time to consume their meal. All food items presented were weighed to the nearest 0.1g before and after the test session. Energy content and macronutrient composition were calculated using data from the USDA National Nutrient Database for Standard Reference (USDA, Agricultural Research Service, Beltsville, MD) and food manufacturer information, when available.

Data Analytic Plan

All analyses were performed with SPSS 16.0. Data were screened for normality. Skew and kurtosis were satisfactory on all variables, and outliers were adjusted to fall 1.5 times the interquartile range below or above the 25th or 75th percentile [29]. This strategy was used because it minimizes outliers’ influence on the characteristics of the distribution, minimally changes the distribution overall, and avoids potential bias associated with eliminating outliers altogether [29].

Linear mixed models with repeated measures were used to examine the associations among emotional eating, pre-meal state negative affect, and observed eating. A mixed models approach was selected to account for within-subject correlation, since each subject was observed on up to two occasions. Unlike repeated measures analysis of variance, the mixed models approach can accommodate data from participants who did not provide complete outcome data for both occasions. The repeated measure was meal instruction (normal versus disinhibited). The dependent variables were total energy intake (kcal; logarithm transformed) and percent macronutrient content consumed (carbohydrate, fat, and protein; arcsine transformed). The independent variables were the main and interactional effects of emotional eating status and pre-meal state negative affect. Subject was treated as a random effect to allow for within-subject correlation; the remaining independent variables were treated as fixed effects in the model. Consistent with prior adult studies examining the construct validity of emotional eating [15], EES-C scores were categorized into High-Emotional Eating (High-EE) and Low-Emotional Eating (Low-EE) using a median split. In a second analysis, meal instruction (normal versus disinhibited) also was included as an independent variable in order to test the three-way interaction of meal instruction, emotional eating status (coded as High-EE or Low-EE), and pre-meal state negative affect. Finally, this same model was conducted accounting for the contribution of anthropometric and demographic information. Specifically, we adjusted for age (years), race (coded as non-Hispanic Caucasian or other), sex, and BMI-z score. Total energy consumed (kcal; logarithm transformed) was included as a covariate in the models examining percent macronutrient content intake (arcsine transformed). Meal type randomization order was considered as a covariate, but was removed because it did not significantly contribute to any model. Models were fit using the restricted maximum likelihood estimation procedure—a standard method for fitting linear mixed models. The restricted maximum likelihood estimation is appropriate because it provides significance tests and confidence intervals for variables of interest, minimizes the effect of nuisance parameters, and produces unbiased estimates of variance and covariance parameters. Associations were considered significant when p values were ≤ 0.05. All tests were two-tailed.

Results

One-hundred fifty-one participants (8 to 18 years) completed the EES-C, BRUMS, and the two laboratory test meals. The average time between the administration of the EES-C and the first test meal was 11.31 ± 8.33 weeks, ranging from the same day (prior to the test meal) to 65 weeks. The average time between the administration of the EES-C and the second test meal was 15.30 ± 9.39 weeks, ranging from two days to 66 weeks. The racial/ethnic background of the sample was 56.0% non-Hispanic Caucasian, 32.9% Black or African American, 6.0% Hispanic, 3.3% Asian, and 1.8% self-identified “Other.” Children represented a wide range of weight strata (BMI-z score: -1.50 to 3.20; 0.91 ± 1.10) [Table 1].

Table 1.

Participant Characteristics Based Upon High and Low Emotional Eating (EE)

Low EE EES-C Total Score ≤ 13 High EE EES-C Total Score > 13 t
Age at Screening (y) 12.48 ± 2.68 13.82 ± 2.71 -3.03**
BMI (kg/m2) 24.00 ± 8.52 24.37 ± 8.98 -0.26
BMI-z score 0.94 ± 1.15 0.86 ± 1.06 0.47
Fat Mass (%) 26.64 ± 12.96 26.55 ± 12.84 0.04
Fat-Free Mass (%) 73.36 ± 12.96 73.44 ± 12.83 -0.04
Disinhibited Meal (n = 150)
    Energy Intake (kcal) 1444.42 ± 586.85 1569.43 ± 585.82 -1.30
    Protein Intake (%) 14.70 ± 3.91 14.76 ± 3.68 -0.09
    Fat Intake (%) 37.92 ± 6.18 36.65 ± 5.68 1.31
    Carbohydrate Intake (%) 48.57 ± 6.18 49.76 ± 7.86 -0.88
    Pre-Meal State Negative Affect (Brunel Mood Scale total score) 4.31 ± 4.51 7.42 ± 6.12 -3.50**
Normal Meal (n = 151)
    Energy Intake (kcal) 1379.06 ± 576.80 1430.62 ± 549.24 -0.56
    Protein Intake (%) 14.69 ± 3.67 14.79 ± 3.24 -1.80
    Fat Intake (%) 38.33 ± 5.59 36.10 ± 6.58 2.26*
    Carbohydrate Intake (%) 48.22 ± 7.62 50.35 ± 8.39 -1.64
    Pre-Meal State Negative Affect (Brunel Mood Scale total score) 3.66 ± 4.30 6.90 ± 5.14 -4.22**
Open in a new tab

Note:

*

p < 0.05

**

p < 0.01

EES-C = Emotional Eating Scale adapted for Children and Adolescents.

EES-C Scores

All three EES-C subscales were highly inter-correlated with each other (rs = 0.67 – 0.80, ps < 0.001) and with the EES-C total score (rs = 0.86 – 0.96, ps < 0.001). Only results using the EES-C total score are presented because, for all analyses, the pattern and significance of results for the EES-C subscales were consistent with findings using the EES-C total score.

The median for the EES-C total score was 13. Therefore, youth with EES-C scores less than or equal to 13 were categorized as Low-EE, and those with total scores that exceeded 13 were categorized as High-EE. Youth in the Low- and High-EE groups did not differ on demographic or anthropometric variables, with the exception that youth with High-EE were marginally, but significantly, older (mean ± SD = 13.82 ± 2.71 years) than Low-EE youth (12.48 ± 2.68 years; Table 1). Per design, Low-EE youth's EES-C total score was significantly lower (4.93 ± 4.25) than High-EE youth (27.59 ± 10.52; p < 0.01). Median scores for Low- and High-EES-C youth were 4 and 26, respectively.

Model Fit

Assumptions of normality and linearity were examined using histograms, scatter plots (for continuous variables) and box and whisker plots (for dichotomous variables), of model residuals, and no violations of either normality or linearity were observed. Because the within-subjects factor has two possible values (normal versus disinhibited), only a single parameter is required to estimate the within-subjects correlation. Comparison of possible covariance structures is therefore not necessary. All -2 Restricted Log Likelihood values ranged between -277.8 through -216.5.

Relationships Among Emotional Eating, Pre-Meal State Negative Affect, and Energy Intake

In an unadjusted model, there was no main effect of either emotional eating status (Low- or High-EE; p = 0.46) or pre-meal state negative affect (BRUMS total score; p = 0.60) on youths’ total energy intake (kcal). However, there was a significant interaction between emotional eating status and pre-meal state negative affect (F (1, 284.33) = 5.31, p = 0.02). Among youth with High-EE, higher reports of pre-meal state negative affect were positively associated with total energy intake (Figure 1), with each 1-unit increase in reported pre-meal state negative affect predicting, on average, a 1.38% (95% CI: 0.08, 2.7%) increase in energy intake. In contrast, for participants with Low-EE, each 1-unit increase in reported pre-meal state negative affect was associated with an average 0.90% (95% CI: -2.29, 0.51) decrease in energy intake. The difference in energy intake associated with a 1-unit increase in pre-meal state negative affect was 2.30% (95% CI: 0.35, 4.29) greater among youth with High-EE compared to Low-EE. Based on the average total intake approaching 1,500 kcal across both meals, a 2.30% change in energy intake corresponds to the consumption of approximately an additional 35 calories more in the High-EE compared to Low-EE for every unit difference in pre-meal negative affect.

Figure 1.

Figure 1

Open in a new tab

Pre-meal state negative affect, measured by the Brunel Mood Scale total score, in relation to total energy intake across laboratory test meals among youth with High-Emotional Eating (EE) and Low-EE, measured by the Emotional Eating Scale adapted for Children and Adolescents total score, p = 0.02. Adjusted, back-transformed geometric means are presented.

In a second model examining the main and interactional effects of emotional eating status, pre-meal state negative affect, and meal instruction on energy intake, there was no main effect of emotional eating status (p = 0.53), pre-meal state affect (p = 0.98), or meal instruction (p = 0.75). The interaction between emotional eating status and pre-meal state negative affect remained significant (F (1, 291.12) = 5.27, p = 0.02), such that the change in total energy intake associated with a 1-unit increase in pre-meal state negative affect was 1.60% (95% CI: -0.54, 3.77) more among youth with High-EE compared to Low-EE. Based on the average total intake approaching 1,500 kcal across both meals, a 1.60% difference in energy intake corresponds to the consumption of approximately an additional 25 calories more among High-EE compared to Low-EE for every unit difference in pre-meal negative affect.

There also was a significant two-way interaction between meal instruction and pre-meal state negative affect; there was a stronger positive association between pre-meal state negative affect and energy intake during the disinhibited, compared to the normal meal, among all youth regardless of emotional eating status (F (1, 155.73) = 5.26, p = 0.02). The three-way interaction of emotional eating status, pre-meal state negative affect, and meal instruction was not significant (p = 0.27), indicating that youth with High-EE ate more in response to negative affect relative to Low-EE youth, regardless of meal instruction condition.

In a third model controlling for sex, race, age, and BMI-z, neither emotional eating status (p = 0.25) nor pre-meal state affect (p = 0.76) were significantly associated with total energy intake. However, the significant interaction between emotional eating status and pre-meal state negative affect remained significant for total adjusted intake (p = 0.03). Consistent with the previous models, the relationship between pre-meal state negative affect and subsequent energy intake was more positive among High-EE compared to Low-EE youth (Table 2). For High-EE youth, each 1-unit increase in pre-meal state negative affect was associated with a 1.09% (95% CI: -0.05, 2.25%) increase in energy intake, where as for youth with Low-EE, each 1-unit increase in pre-meal state negative affect was associated with a 0.81% (95% CI: -2.07, 0.28) decrease in energy intake. The difference in energy intake associated with a 1-unit increase in pre-meal state negative affect was 1.92% (95% CI: 0.19, 3.68) higher among youth with High- compared to Low-EE. Based on the average total intake approaching 1,500 kcal across both meals, a 1.92% difference in energy intake corresponds to the consumption of approximately an additional 30 kcal more among High-EE compared to Low-EE for every unit difference in pre-meal negative affect.

Table 2.

Linear Mixed Model of Reported Emotional Eating Predicting Energy Intake

Predictor variables Adjusted Estimates in Model Standard Error DF t p
Intercept 2.77 .07 153.47 42.23 .000
Age .02 .01 146.29 5.13 .000
Race .06 .02 146.29 2.49 .014
Sex .13 .02 146.31 5.86 .000
BMI-z Score .04 .01 145.00 3.50 .001
Disinhibited Meal .03 .01 150.53 2.85 .005
EE Status .04 .03 205.45 1.15 .249
Pre-Meal State Negative Affect .01 .001 249.22 1.88 .061
EE Status X Pre-Meal State Negative Affect -.01 .004 279.71 -2.17 .031
Open in a new tab

Note: Energy intake (kcal) was log transformed; N = 151; Race coded: 0 = other, 1 = non-Hispanic Caucasian; Sex coded: 0 = male, 1 = female; Meal type coded: 0 = normal, 1 = disinhibited; Emotional eating (EE) status coded: 0 = High EE, 1 = Low EE.

For all models, although the association between negative affect and total energy intake did not differ significantly from zero in either the High- or Low-EE groups, the association between mood and intake among High- versus Low-EE differed significantly from each other.

No significant main or interactional effect was observed for macronutrient content intake, meal-type foods, or snack- and dessert-type foods for any model.

Discussion

The current investigation examined the construct validity of the Emotional Eating Scale Adapted for Children and Adolescents (EES-C) [7] by comparing self-reported emotional eating with children's actual emotional eating, measured via the relationship of pre-meal state negative affect and subsequent energy intake. We found that among those reporting high emotional eating, based upon the EES-C total score, pre-meal state negative affect was positively associated with total observed energy consumed, to a greater extent than among those reporting low emotional eating.

These findings suggest that the EES-C adequately discriminates between youth who consume more energy when they experience negative affect compared to those who do not consume more following reports of negative affect. Youth reporting high EES-C scores consumed approximately 1-2% more total energy (in kcal) for every 1-unit increased in reported pre-meal state negative affect. With the average total intake approaching 1,500 kcal across both meals, a 1-2% change in energy intake corresponds to a difference of approximately 15-30 kcal for every unit difference in pre-meal negative affect. In the current laboratory paradigm, pre-meal state negative affect among youth with high emotional eating was approximately 3 units higher than youth with low emotional eating, suggesting about a 50-100 kcal difference in total intake, at each of the test meals. Although relatively innocuous during a single meal, such a pattern of eating, if sustained over time in the natural environment, might contribute to excess intake and lead to inappropriate weight gain and obesity [30]. Prospective data are needed to investigate whether youth with high EES-C scores gain excessive weight or have an increased risk for obesity onset as compared to youth with low EES-C scores.

There was no significant three-way interaction among emotional eating status, pre-meal negative affect, and meal instruction, which suggests that the EES-C differentiated High- from Low-emotional eating youth independent of the meal instruction. Whether children were instructed to eat “as you would at a normal meal” or “let yourself go and eat as much as you want,” the EES-C still identified children who demonstrated a pattern of eating wherein negative affect was associated with increased intake. Notably, there were no main or interactional associations of the EES-C or pre-meal state negative affect with percentage of macronutrients. These data are consistent with prior studies in youth that found no association between emotional eating and snacking behavior or fat intake [4, 31]. Taken together, these findings suggest the EES-C, which is a brief and easily administered questionnaire, may be a good measure to assess risk for overconsumption of energy in response to negative mood states. Our findings suggest that youth reporting an EES-C total score equal to or greater than 13 may consume greater overall energy than those with a score of less than 13; it is necessary to determine whether these findings generalize from laboratory to naturalistic settings. To enhance the utility of the EES-C as a screening tool, future studies would benefit from empirically examining optimal EES-C cut-points that identify youth who are likely to consume excess energy in response to negative emotions and who are at heightened risk for obesity.

In contrast to the pattern of eating demonstrated by youth with reported high emotional eating, pre-meal negative affect was associated with somewhat, but not significantly, lower energy intake during the meal for children with reported low emotional eating. This pattern of eating is consistent with a hypothesized “normative” stress response during which appetite is suppressed [32]. Blissett and colleagues (2010) found that pre-school age children whose mothers often used food to regulate their children's emotions (i.e., inducing emotional eating) consumed more chocolate following a negative mood induction than children in the neutral mood condition. Conversely, children with mothers who did not report often using food to alleviate emotions consumed less in the negative mood condition and more in the neutral condition. Given these findings and our pattern of results, it seems possible that children's normative response to emotional stressors may be to decrease their food intake. However, for some children, the development of recurrent, persistent emotional overeating may be an alternative, possibly learned, response to coping with negative mood states during early childhood. Further exploration of this hypothesis in prospective studies is warranted.

Strengths of the current study include the large sample size, the use of objective methods to examine eating behaviors, the representation of both non-Hispanic Caucasian and African American youth, and the inclusion of non-treatment-seeking boys and girls of a broad age and weight range. Another primary strength of the current study is the assessment of pre-meal state negative affect to assist in capturing emotional eating behaviors in laboratory, which eliminates the need for a potentially ineffective or variably effective negative mood induction experimental paradigm [19-21]. Moreover, the inclusion of two laboratory test meals, a disinhibited and normal meal, enhances the power of the study and increases the generalizability of the findings. That is, findings suggest that youth reporting a high propensity for emotional eating are likely to consume more energy in response to negative affect across at least two distinct eating situations (e.g., during a normal meal and contexts during which youth eat in a disinhibited style or ‘let themselves go’). It is also notable that among youth reporting high emotional eating, the relationship between pre-meal state negative affect and total energy intake was significant regardless of meal type, suggesting that mood may be driving youth's overconsumption rather than the meal instruction or type of eating.

Limitations include that all participants were provided the same standardized breakfast prior to the test meal, regardless of body weight. It is possible that heavier youth required considerably more energy than what was provided for breakfast to have equal satiety as compared to their leaner or non-overweight peers. However, the relationship between reported emotional eating and actual eating in response to pre-meal state negative affect was identified even after accounting for BMI-z score. A large buffet, although effectively used in prior studies [22, 28], may have impacted participants’ consumption; the amount of food eaten generally increases when greater amounts of food are presented [33, 34]. Therefore, studies examining emotional eating and energy intake in naturalistic settings are an important next research step.

In conclusion, the EES-C appears to demonstrate good construct validity with observed emotional eating in the laboratory, and may be an easily administered screening tool for children and adolescent's propensity to eat in response to negative affect. Should future studies determine that high EES-C scores prospectively predict the development of excess weight gain over time, emotional eating should be considered as a potential target for obesity prevention in youth..

Acknowledgments

We thank the families who participated in these studies and the staff of the metabolic kitchen at the NIH Clinical Center. We also thank Cara Olsen for her invaluable consultation regarding the statistical analyses for the current study. J Yanovski and M Kozlosky are Commissioned Officers in the United States Public Health Service, Department of Health and Human Services. MTK, SZY and JAY designed the study. AV and MTK conceived the hypothesis for this article and drafted the manuscript. MTK and JAY supervised the data collection. BEM, OLM, JMZ, and MK collected the data. AV, MTK, LBS, and LMR conducted the data analysis. All authors participated in the interpretation of the results and approved the final version of the manuscript.

The funding organization played no role in design and conduct of the study; collection, management, analysis, and interpretation of the data; nor preparation or review of the manuscript.

Research support: Intramural Research Program, NIH, grant 1ZIAHD000641 from the NICHD with supplemental funding from NIMHD (to JAY).

Footnotes

ClinicalTrials.gov ID: NCT00320177

None of the authors had any conflict of interest.

References

  • 1.Faith MS, Allison DB, Geliebter A. Emotional eating and obesity: Theorectical considerations and practical recommendations. In: Dalton S, editor. Obesity and Weight Control: The Health Professional's Guide to Understanding and Treatment. Gaithersburg, MD; Aspen: 1997. pp. 439–65. [Google Scholar]
  • 2.Braet C, et al. Differences in eating style between overweight and normal-weight youngsters. J Health Psychol. 2008;13:733–43. doi: 10.1177/1359105308093850. [DOI] [PubMed] [Google Scholar]
  • 3.Nguyen-Rodriguez ST, et al. BMI as a moderator of perceived stress and emotional eating in adolescents. Eat Behav. 2008;9:238–46. doi: 10.1016/j.eatbeh.2007.09.001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4.Wardle J, et al. Eating style and eating behaviour in adolescents. Appetite. 1992;18:167–83. doi: 10.1016/0195-6663(92)90195-c. [DOI] [PubMed] [Google Scholar]
  • 5.Shapiro JR, et al. Evaluating binge-eating disorder in children: Development of the Children's Binge Eating Disorder Scale (C-BEDS). Int J Eat Disord. 2007;80(1):82–9. doi: 10.1002/eat.20318. [DOI] [PubMed] [Google Scholar]
  • 6.Marcus MD, Kalarchian MA. Binge eating in children and adolescents. Int J Eat Disord. 2003;34(Supple):S47–57. doi: 10.1002/eat.10205. [DOI] [PubMed] [Google Scholar]
  • 7.Tanofsky-Kraff M, et al. Validation of emotional eating scale adapted for use in children and adolescents (EES-C). Int J Eat Disord. 2007;40(3):232–40. doi: 10.1002/eat.20362. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 8.Ashcroft J, et al. Continuity and stability of eating behaviour traits in children. European Journal of Clinical Nutrition. 2007;62(8):985–90. doi: 10.1038/sj.ejcn.1602855. [DOI] [PubMed] [Google Scholar]
  • 9.van Strien T, et al. Parental control and the dopamine D2 receptor gene (DRD2) interaction on emotional eating in adolescence. Appetite. 2010;54(2):255–61. doi: 10.1016/j.appet.2009.11.006. [DOI] [PubMed] [Google Scholar]
  • 10.van Strien T, van der Zwaluw CS, Engels RC. Emotional eating in adolescents: A gene (SLC6A4/5-HTT) - depressive feelings interaction analysis. Journal of Psychiatry Research. 2010;44(15):1035–42. doi: 10.1016/j.jpsychires.2010.03.012. [DOI] [PubMed] [Google Scholar]
  • 11.Stice E, Presnell K, Spangler D. Risk factors for binge eating onset in adolescent girls: A 2-year prospective study. Health Psychol. 2002;21(2):131–38. [PubMed] [Google Scholar]
  • 12.Braet C, van Strien T. Assessment of emotional, externally induced and restrained eating behaviour in non to twelve-year-old obese and non-obese children. Behav Res Ther. 1997;35(9):863–73. doi: 10.1016/s0005-7967(97)00045-4. [DOI] [PubMed] [Google Scholar]
  • 13.van Strien T, Bazelier FG. Perceived parental control of food intake is related to external, restrained, and emotional eating in 7-12-year-old boys and girls. Appetite. 2007;49:618–25. doi: 10.1016/j.appet.2007.03.227. [DOI] [PubMed] [Google Scholar]
  • 14.Caccialanza R, et al. Validation of the Dutch Eating Behaviour Questionnaire parent version (DEBQ-P) in the Italian population: A screening tool to detect differences in eating behaviour among obese, overweight, and normal-weight preadolescents. Eur J Clin Nutr. 2004;58:1217–22. doi: 10.1038/sj.ejcn.1601949. [DOI] [PubMed] [Google Scholar]
  • 15.Evers C, de Ridder DT, Adriaanse MA. Assessing yourself as an emotional eater: Mission impossible? Health Psychol. 2009;28(6):717–25. doi: 10.1037/a0016700. [DOI] [PubMed] [Google Scholar]
  • 16.O'Connor DB, O'Connor RC. Perceived changes in food intake in response to stress: The role of conscientiousness. Stress and Health. 2004;20:279–91. [Google Scholar]
  • 17.Oliver G, Wardle J, Gibson EL. Stress and food choice: A laboratory study. Psychosomatic Medicine. 2000;62:853–65. doi: 10.1097/00006842-200011000-00016. [DOI] [PubMed] [Google Scholar]
  • 18.O'Connor DB, et al. Effects of daily hassles and eating style on eating behavior. Health Psychol. 2008;27:s20–31. doi: 10.1037/0278-6133.27.1.S20. [DOI] [PubMed] [Google Scholar]
  • 19.Cohen S, Hamrick N. Stable individual differences in physiological response to stressors: Implicaitons for stress-elicited changes in immune related health. Behavior, Brain, and Immunity. 2003;17:407–414. doi: 10.1016/s0889-1591(03)00110-7. [DOI] [PubMed] [Google Scholar]
  • 20.Goldstein MJ. Individual differences in response to stress. Am J Comm Psychol. 1973;1(2):113–37. doi: 10.1007/BF00880131. [DOI] [PubMed] [Google Scholar]
  • 21.van Eck MM, et al. Individual differences in cortisol responses to a laboratory speech task and their relationshiop to responses to stressful daily events. Bio Psychol. 1996;43(1):69–84. doi: 10.1016/0301-0511(95)05159-7. [DOI] [PubMed] [Google Scholar]
  • 22.Tanofsky-Kraff M, et al. Laboratory assessment of the food intake of children and adolescents with loss of control eating. Am J Clin Nutr. 2009;89:1–8. doi: 10.3945/ajcn.2008.26886. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 23.Kuczmarski RJ, et al. 2000 CDC Growth Charts for the United States: methods and development. Vital Health Stat 11. 2002;246:1–190. [PubMed] [Google Scholar]
  • 24.Arnow B, Kenardy J, Agras WS. The Emotional Eating Scale: The development of a measure to assess coping with negative affect by eating. Int J Eat Disord. 1995;18(1):79–90. doi: 10.1002/1098-108x(199507)18:1<79::aid-eat2260180109>3.0.co;2-v. [DOI] [PubMed] [Google Scholar]
  • 25.Rosner B. Fundamental of Biostatistics. Sixth ed. Duxbury Press; 2005. [Google Scholar]
  • 26.Terry PC, et al. Development and validation of a mood measure for adolescents. J Sports Sci. 1999;17:861–72. doi: 10.1080/026404199365425. [DOI] [PubMed] [Google Scholar]
  • 27.Walsh BT, Boudreau G. Laboratory studies of binge eating disorder. Int J Eat Disord. 2003;34(Suppl):S30–8. doi: 10.1002/eat.10203. [DOI] [PubMed] [Google Scholar]
  • 28.Mirch MC, et al. Effects of binge eating on satiation, satiety, and energy intake of overweight children. Am J Clin Nutr. 2006;84:732–38. doi: 10.1093/ajcn/84.4.732. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 29.Behrens JT. Principles and procedures of exploratory data analysis. Psychol Methods. 1997;2:131–160. [Google Scholar]
  • 30.Rosenbaum M, Leibel RL, Hirsch J. Obesity. N Engl J Med. 1997;337(6):396–407. doi: 10.1056/NEJM199708073370606. [DOI] [PubMed] [Google Scholar]
  • 31.van Strien T, Oosterveld P. The children's DEBQ for assessment of restrained, emotional, and external eating in 7-12-year-old children. Int J Eat Disord. 2008;41:72–81. doi: 10.1002/eat.20424. [DOI] [PubMed] [Google Scholar]
  • 32.Blissett J, Haycraft E, Farrow C. Inducing preschool children's emotional eating: Relations with parental feeding practices. Am J Clin Nutr. 2010;92:359–65. doi: 10.3945/ajcn.2010.29375. [DOI] [PubMed] [Google Scholar]
  • 33.Gosnell BA, et al. Food presentation and energy intake in a feeding laboratory study of subjects with binge eating disorder. Int J Eat Disord. 2001;30:441–6. doi: 10.1002/eat.1105. [DOI] [PubMed] [Google Scholar]
  • 34.Shomaker LB, et al. Eating in the absence of hunger in adolescents: Intake after a large-array meal compared with that after a standardized meal. Am J Clin Nutr. 2010;92:697–703. doi: 10.3945/ajcn.2010.29812. [DOI] [PMC free article] [PubMed] [Google Scholar]

RESOURCES