Metacognitive beliefs and eating disorders

Metacognition refers to how people think about their own thoughts [1]. The self-regulatory executive function (S-REF) model posits that metacognitive beliefs impact an individual’s response to internal events, including their propensity to engage in excessive worrying and self-focus, utilize maladaptive coping strategies (e.g., avoidance, thought suppression), and place heightened attention on threatening stimuli [2]. Two broad areas of metacognitive beliefs are positive metacognitive beliefs (i.e., beliefs that focus on the benefit of engaging in rumination, worry, threat-monitoring, and counterproductive coping strategies) and negative metacognitive beliefs (i.e., beliefs that focus on the danger and uncontrollability of thoughts) [3, 4].

Eating disorder researchers have expressed a need to expand the understanding of cognitive models used to treat psychiatric patients with eating disorders [5,6,7]. As such, there has been a growing interest in understanding patterns of metacognitive beliefs among eating disorder patients. Existing studies have found that compared to healthy controls, individuals with anorexia nervosa and bulimia nervosa manifest elevated levels of metacognitive beliefs about the danger and uncontrollability of thoughts, the need to control thoughts, and cognitive self-consciousness [4, 5, 8]. Anxious thoughts about weight are likely perceived as dangerous and uncontrollable by individuals with eating disorders and attempts to control these anxious thoughts are made through restricted eating and other weight management behaviors [8]. The existing research supports metacognitive beliefs as one mechanism that contributes to the development and maintenance of eating disorder pathology.

Metacognitive beliefs and emotional eating

To the best of our knowledge, no studies to date have investigated what role metacognitive beliefs play in the manifestation of emotional eating, a well-known risk factor for the development of eating disorders. Emotional eating describes the act of eating in response to negative emotions such as anxiety and depression [9]. As emotional eating is often used as a coping mechanism when faced with distressing emotions, over time it can lead to body dissatisfaction and the utilization of compensatory behaviors including restricted eating and purging [10, 11]. Elucidating the associations between metacognitive beliefs and emotional eating may be informative to the development of eating disorder prevention programs, especially those programs aimed at adolescents since emotional eating typically begins to manifest during this developmental period. If metacognitive beliefs are associated with emotional eating, it may be beneficial to integrate metacognitive treatments [12] into eating disorder prevention programs that target emotional eating in adolescents. While we are unaware of any existing metacognitive therapies that have been evaluated with adolescents in this manner, Esbjørn and colleagues assessed an 8-week group-based metacognitive therapy for youth of ages 7–13 years with generalized anxiety disorder. The intervention used a manual that was adapted from a metacognitive treatment for adults with generalized anxiety disorder and consisted of 8-weekly, 2-h group sessions for youth and two workshops for their parents [13]. Participants demonstrated significant reductions in anxiety following treatment and at a 6-month follow-up, with effects sizes in the large range [13], supporting the efficacy of a metacognitive therapy in adolescents experiencing psychological difficulties.

Purpose of the current study

The purpose of the current study was to assess the associations between metacognitive beliefs and emotional eating among a sample of adolescents. Based on the extant literature with patients with eating disorders [4, 5, 8], we predicted that adolescents who endorsed higher levels of emotional eating would manifest significantly greater negative metacognitive beliefs compared to adolescents who endorsed lower levels of emotional eating. We also hypothesized that negative metacognitive beliefs would be positively associated with emotional eating after accounting for covariates in a hierarchical multiple linear regression analysis.

Materials and methods

Participants

The sample was comprised of 135 adolescents enrolled in 8th grade health education classes at a middle school in the Southern part of the United States. Table 1 provides descriptive statistics for the sample. The mean age of the sample was 13.62 years (SD = 0.57). Participants ranged in age from 13 to 15 years. The majority of participants were female (67.4%) and identified as White (62.2%). Most participants had a parent with a college degree or higher (67.5%).

Table 1 Demographic variables for the sample
Full size table

Measures

Emotional eating scale adapted for children and adolescents, short-form [14]

The EES-C Short-Form is a 10-item self-report measure that assesses the urge to cope with negative affect by eating in children and adolescents ages 8 to 17 years. It was developed from the original 25-item Emotional Eating Scale Adapted for Children and Adolescents [15]. Participants are asked to rate the extent to which they feel the urge to eat in response to negative emotions (e.g., sad, lonely, jealous, nervous, angry). Responses are scored on a 5-point Likert scale from 1 (No desire to eat) to 5 (Very strong desire to eat). Items are summed together to produce a Total score. Higher scores reflect a stronger desire to eat in response to negative mood. In the present study, Cronbach’s Alpha for the EES-C Short-Form Total Score was 0.86. In another study [14], the EES-C Short-Form demonstrated good internal consistency reliability (α = 0.87) and a high degree of overlapping variance with the original EES-C Total Score and Subscale Scores (r = 0.71–0.96). The EES-C Short-Form also exhibited a unidimensional factor structure in an Exploratory Factor Analysis, supporting construct validity [14].

Metacognitions questionnaire for children (MCQ-C) [16]

The MCQ-C is a 24-item self-report measure that was adapted from the 30-item adolescent version of the Metacognitions Questionnaire. The MCQ-C assesses four domains of metacognitive beliefs in youth ages 7 to 17 years including positive metacognitive beliefs, negative metacognitive beliefs, cognitive monitoring, and superstitious/punishment/responsibility beliefs. For the purposes of the current study, we focused on the positive metacognitive beliefs (e.g., “If I worry about things now, I will have fewer problems in the future”) and negative metacognitive beliefs (e.g., “It is not a good idea to worry because worrying is bad for me”) subscales. A 4-point Likert response scale (“do not agree,” “agree slightly,” “agree moderately,” “agree very much”) is used, with higher scores indicative of greater endorsement of the statements in the scale. Cronbach’s Alpha for the MCQ-C positive metacognitive beliefs and negative metacognitive beliefs scales were 0.80 and 0.72, respectively. Other studies have demonstrated satisfactory internal consistency reliability for the positive and negative metacognitive beliefs scales (alphas range from 0.75 to 0.86) [16] and construct validity of the MCQ-C [17], as well as criterion and convergent validity [18, 19].

Positive and negative affect schedule for children (PANAS-C) [20]

The PANAS-C is a 30-item self-report measure that assesses positive and negative affect in youth in the 4th through 8th grades. For the present study, we utilized the Negative Affect subscale which includes 15 descriptors of feelings and emotions related to negative affect (e.g., sad, frightened). A 5-point Likert response scale is used (“very slightly or not at all,” “a little,” “moderately,” “quite a bit,” “extremely”). Youth are asked to report on their feelings and emotions over the past few weeks and higher scores are indicative of greater negative affect. In the present study, Cronbach’s Alpha for the PANAS-C Negative Affect Scale was 0.91. The PANAS-C has demonstrated good internal consistency reliability (alphas range from 0.87 to 0.92) and convergent validity in existing research [21].

Demographic information

A demographic questionnaire was completed by participants to obtain the following information: age, sex, race/ethnicity, and parents’ highest level of education.

Procedures

The present study was cross-sectional in nature. Participants were recruited from a middle school in the Southern part of the United States. A graduate research assistant visited 8th grade health classes in Spring 2018 and Winter 2018 to provide students with informational packets about the study. The packets included detailed information about the study, a parental consent form, and participant assent form. Students were eligible for the study if they had a signed parental consent and participant assent form and were able to complete study questionnaires in English. Students in special education classes were excluded from participation in the current study since their different academic needs potentially could impact their ability to provide valid responses on self-report questionnaires [22]. Participants completed written questionnaires, including the demographic questionnaire, EES-C Short-Form, MCQ-C, and PANAS-C, in their health classroom or a nearby school lounge. Participants were not compensated for their participation in the current study. All study procedures were approved by the Baylor University Institutional Review Board.

Statistical analysis

Missing data were examined using Little’s test. If data were missing at random, they were imputed utilizing expectation–maximization (EM). Low and high emotional eating classification was determined by a median split as recommended by Vannucci and colleagues (2012). In the current study, the median score for Total EES-C Short-Form was 20, which is similar to the median score of 22 found in the original validation study of the EES-C Short-Form [14]. In the present study, participants with a Total EES-C Short-Form score less than or equal to 20 were classified as low emotional eaters (N = 65, 48.1%) and participants with a Total EES-C Short-Form score greater than 20 were classified as high emotional eaters (N = 70, 51.9%). Independent samples t tests were used to investigate whether there were statistically significant differences (p < 0.05) in positive and negative metacognitive beliefs between adolescents who engaged in low levels of emotional eating versus high levels of emotional eating. Effect sizes were calculated by taking the difference between the low emotional eating sample mean and the high emotional eating sample mean, divided by the pooled standard deviation. Effect sizes for differences in means are categorized as small (0.20), medium (0.50), and large (0.80) in magnitude.

Hierarchical multiple linear regression analysis was used to examine multivariate associations between study variables. The Total EES-C Short-Form score was the criterion variable in the regression model. Socio-demographic variables (age, gender, race/ethnicity, parental highest education) and negative affect were entered into Block 1 as control variables. Negative affect was included as a control variable as it serves as a non-specific risk for experiencing emotional distress [23]. In Block 2 positive and negative metacognitive beliefs were simultaneously entered to determine if they incremented the variance explained in the final adjusted model for which emotional eating was the outcome.

Results

Participants classified as high emotional eaters reported statistically significant higher negative metacognitive beliefs (Mean = 15.56; SD = 4.22) compared to participants classified as low emotional eaters (Mean = 12.85; SD = 4.31; p < 0.001; t =  − 3.69). The effect size (d = 0.64) was in the medium range. Participants classified as high emotional eaters reported similar levels of positive metacognitive beliefs (Mean = 10.02; SD = 3.80) compared to participants classified as low emotional eaters (Mean = 9.25; SD = 3.56; p = 0.230; t =  − 1.20).

Among the potential covariates, negative affect (r = 0.22, p = 0.010), but none of the sociodemographic variables (age, gender, race, parental education) shared bivariate associations with emotional eating (rs ranging from − 0.13 to 0.13, ps > 0.074). As such, only negative affect was retained as a covariate for the multivariate analysis. Table 2 presents standardized beta coefficients for the hierarchical multiple linear regression analysis from the total model, which included all predictors. As shown in Table 2, the variance explained by the total adjusted model, for which emotional eating was the outcome, was significant and the amount of total variance accounted for in emotional eating specific to the metacognitive beliefs was significant. In the final model, MCQ-C Negative Metacognitive Beliefs Subscale was significantly correlated with the EES-C Short-Form Total Score (standardized beta coefficient = 0.29; p = 0.006).

Table 2 Multiple linear regression analysis examining predictors of emotional eating
Full size table

Discussion

The purpose of the current study was to assess the associations between positive and negative metacognitive beliefs and emotional eating among a sample of adolescents. As predicted, we found adolescents classified as high emotional eaters reported statistically significant higher negative metacognitive beliefs compared to adolescents classified as low emotional eaters. The effect size was in the medium range. Higher levels of emotional eating were associated with greater endorsement of negative metacognitive beliefs in adolescents after controlling for negative affect. The finding that the significant positive association between negative metacognitive beliefs and emotional eating remained after controlling for negative affect is noteworthy since a strong association between negative affective states like depression and emotional eating has been documented in adolescents [9]. Although negative affect shared a bivariate association with emotional eating in this study, negative affect was no longer associated with emotional eating once statistically controlling for the effects of metacognitive beliefs in the multivariate analysis. Our finding suggests that negative metacognitive beliefs potentially play a unique role in the manifestation of emotional eating in this population.

While the cross-sectional nature of this study precludes us from inferring causation, our findings indicate a need for future prospective studies to elucidate the temporal associations between emotional eating and negative metacognitive beliefs. Such information could provide support for assessing the efficacy of metacognitive therapies for the prevention of emotional eating in adolescents. One metacognitive therapy that may warrant particular attention in the treatment and prevention of eating disorder pathology is the attention training technique (ATT) [12, 24]. ATT, an auditory listening technique developed to foster disengagement from worry through strengthening attentional control, is one component of metacognitive therapy that alters metacognitive beliefs, particularly negative metacognitive beliefs about the uncontrollability and danger of worry [25]. Gaining experience disengaging from the worry process leads to disconfirming evidence of negative metacognitive beliefs [26]. As a result, worry becomes less threatening and individuals can more easily disengage from the worry process. There are data supporting ATT as a standalone intervention for depression and anxiety [27, 28]. If future prospective studies find negative metacognitive beliefs contribute to emotional eating, it may be valuable for future research to evaluate the efficacy of metacognitive therapy or individual treatment components, such as ATT, for reducing eating disorder pathology in adolescents.

The current study had a number of limitations. In addition to the cross-sectional nature of the study, participants completed self-report measures in a school setting, which may have influenced their propensity to respond in socially desirable ways. Our sample of adolescents was predominantly White, female, and from homes who had at least one parent with a college degree or higher. As such, our findings may not generalize to males and more diverse adolescent populations. It will be important for future studies to replicate our findings with males and diverse adolescent samples, including non-white and older adolescents. We did not collect data on height and weight in the current study. Consequently, we were not able to compute Body Mass Index (BMI) for participants. Given previous findings that support different patterns of emotional eating between youth in varying BMI categories, future studies should assess if the pattern of correlations between emotional eating and metacognitive beliefs are influenced by BMI classification.

In conclusion, higher levels of emotional eating were associated with greater endorsement of negative metacognitive beliefs in our sample of adolescents and those adolescents classified as high emotional eaters reported statistically significant higher negative metacognitive beliefs compared to adolescents classified as low emotional eaters. These findings provide preliminary evidence that negative metacognitive beliefs may play a role in the manifestation of emotional eating in adolescents.

What is already known on this subject?

Existing research supports metacognitive beliefs as one mechanism that contributes to the development and maintenance of eating disorder pathology. No studies to date have investigated what role metacognitive beliefs play in the manifestation of emotional eating, a well-known risk factor for the development of eating disorders.

What does this study add?

The current findings provide preliminary evidence that negative metacognitive beliefs may play a role in the manifestation of emotional eating in adolescents and highlight the need for future prospective studies to elucidate the temporal associations between emotional eating and negative metacognitive beliefs.