What Facets of Picky Eating Relate to Eating Concerns in Undergraduate Students?: The Moderating Effects of Negative Psychological Correlates

Purpose: Picky eating (PE) can occur in adulthood and is associated with mental health concerns. PE is often conceptualized as distinct from disordered eating (DE), but recent research maps positive relationships between these maladaptive eating phenotypes. Precisely what PE facets relate to DE remain unknown, as do factors such as negative psychological correlates that might explain relationships between PE facets and DE. Methods: A large, undergraduate sample (N=509) completed an online survey assessing PE facets (Adult Picky Eating Questionnaire; meal presentation, food variety, meal disengagement, and taste aversion), disordered eating (Eating Disorder Examination Questionnaire), specically eating concerns, and negative psychological correlates such as mental health concerns (Depression, Anxiety and Stress Scale - 21 Items) and inexible eating (Inexible Eating Questionnaire). Results: Positive relationships emerged between PE facets, eating concerns, and negative psychological correlates. Negative psychological correlates moderated relationships between PE facets and eating concerns. Meal presentation, meal disengagement, and taste aversion were more strongly associated with eating concerns when mental health concerns and inexible eating were higher. Food variety did not signicantly explain variance in eating concerns. Conclusions: Considering PE multidimensionally may yield important insights beyond the broader construct. Mental health concerns and inexible eating may be treatment and research targets in addressing the overlap between PE facets such as meal presentation, meal disengagement, and taste aversion and eating concerns.


Introduction
Picky eating (PE) is often a vexing behavior in children. Not surprisingly, existing knowledge of PE correlates focuses largely on children [1][2][3][4], a striking limitation considering that PE persists over time [5-6, but see 7], that childhood PE predicts adult PE [3,[8][9], and that approximately 35% of adults engage in PE [10]. Adult PE is associated with poor dietary quality [11][12] and signi cant psychological impairment including a range of mental health concerns and disordered eating [13][14][15] (e.g., dietary restraint, eating concerns, binge eating, and overall eating pathology) [13]. Furthermore, relationships between adult PE and disordered eating, speci cally eating concerns, are strengthened by negative psychological correlates such as mental health concerns and in exible eating in undergraduate students [13]. A critical piece missing from this research is precisely which facets of PE explain these relationships.
Picky Eating: De nitions, Multidimensional Structure, And Negative Psychological Correlates PE includes di culty trying new foods, eating a limited range of foods, and marked rigidity and sensory sensitivity around food presentation and preparation [1]. PE also includes the rejection of familiar foods which distinguishes it from related constructs like food neophobia [16]. Because PE in adults is less understood than PE in children, a multidimensional tool was created to assess four distinct facets of adult PE [17]: (1) meal presentation, or strong preferences regarding food preparation and presentation; (2) food variety, or restricted dietary intake across food groups; (3) meal disengagement, or avoidant behavior around mealtimes; and (4) taste aversion, or rejection of bitter or sour foods [17][18]. This multidimensional approach to assessing PE has been identi ed as a key area of future research to de ne and elaborate on the nuances of adult PE [19], including the overlap between PE, disordered eating, and negative psychological correlates.
Some research has examined PE in relation to disordered eating, with mixed results. One study found that a comorbid class (high PE and disordered eating symptoms) was the largest class via latent class analysis, and was associated with psychosocial impairment and poor mental health outcomes [14]. Other research found that all PE facets were related to social anxiety around eating as well as anxiety sensitivity when controlling for disordered eating symptoms [17]. In another study [19] picky eaters endorsed less disordered eating (including shape and weight concern behaviors) compared to "approaching" eaters, or eaters who had high enjoyment of food. However, picky eaters also endorsed more eating-related impairment and depressive symptoms than "moderate" eaters, or eaters with a moderate food approach and avoidance [19]. Related to this is research suggesting that childhood PE did not predict disordered eating in adulthood [20]. These mixed ndings and others [21][22][23][24] suggest unclear relationships between PE and disordered eating, warranting further research mapping these relationships.
Relationships between PE and negative psychological correlates are clearer. Adult picky eaters endorse more obsessive-compulsive disorder and depressive symptoms than non-PE adults [14,22]. Relatedly, positive relationships have been identi ed between PE and anxiety symptoms [17,25], and preliminary evidence suggests that adult PE is positively associated with stress symptoms [13]. Additionally, picky eaters with Avoidant Restrictive Food Intake Disorder (ARFID) reported elevated eating in exibility (rigidity around eating rules) [24]. Using a multidimensional approach to PE, one study found that meal presentation and meal disengagement predicted psychological in exibility and psychosocial impairment [17]. Furthermore, meal disengagement predicted depressive symptoms [17], suggesting that meal disengagement may be a more proximal predictor of poor mental health than other facets of PE. Taken together, while PE is consistently associated with psychological impairment, there is mixed evidence regarding relationships with disordered eating, and little is known about which speci c facets of PE are associated with negative psychological correlates and disordered eating.

Negative Psychological Correlates as Moderators of Relationships Between Picky Eating and Disordered Eating
Recently, Barnhart and colleagues [13] explored a range of negative psychological correlates as moderators of relationships between broad PE behavior and binge eating, dietary restraint, eating concerns, and overall eating pathology in undergraduate students. Mental health concerns (e.g., anxiety, depression, and stress symptoms) and in exible eating moderated relationships between PE and eating concerns such that PE was more strongly associated with eating concerns when mental health concerns and in exible eating were higher [13]. As it is measured on the Eating Disorder Examination-Questionnaire, eating concerns captures a range of disordered eating attitudes and cognitions including the fear of loss of control during eating, social aspects of eating such as eating in secrecy and guilt about eating, and a preoccupation with eating, food, calories [26]. The PE literature suggests some overlap with eating concerns, especially in terms of the social aspects of eating [14,17,27]. However, it remains unclear precisely what it is about PE that is moderated by these negative psychological correlates to explain meaningful variance in eating concerns. To this end, Barnhart and colleagues [13] outlined the need to examine these research questions with a focus speci c PE facets.

The Present Study
The present study builds on previous research demonstrating positive relationships between PE and disordered eating [14][15], speci cally eating concerns [13], by examining relationships between PE facets and eating concerns. Given recent research pointing to overlap between broad PE, disordered eating, and negative psychological correlates including mental health concerns and in exible eating [13], this study had two exploratory aims. First, we examined bivariate relationships between PE facets, eating concerns, and negative psychological correlates. Second, we examined negative psychological correlates as moderators of relationships between PE facets and eating concerns.

Participants
This secondary analysis [13] utilized undergraduate participants from a large, Midwestern university who completed an online survey on eating behaviors. Initially, 579 participants were recruited; of those, 70 participants were excluded due to completing 75% or less of the survey (n = 42), completing the survey multiple times (n = 17), or failing to meet quality standards (e.g., incorrect responses to attention checks; n = 11). The nal sample (N = 509) ranged in age from 18 to 25 years (M = 19.96, SD = 2.93) and was primarily women (n = 390), White (n = 438), heterosexual (n = 436), and middle class (n = 270). Additionally, 6.3% and 2% of participants reported past and current eating disorder diagnosis, respectively. See Barnhart and colleagues [13] for additional participant information.

Anthropometry
Body mass index (BMI; kg/m 2 ) was calculated using self-reported height in feet and weight in pounds.

Demographics
Participants self-reported demographic characteristics including age, gender, socioeconomic status, year in school, race, and sexual orientation. Data were also collected on current and past eating disorder diagnosis, current and past PE, and adherence to a particular eating style (e.g., vegetarian, vegan, etc.).

Adult Picky Eating Questionnaire (APEQ)
Adult PE was examined using the APEQ [17]. The questionnaire yields four subscales that are typically averaged to create a total score: meal presentation, food variety, meal disengagement, and taste aversion.
This study focused on the subscales, not the total score. Example items are "I have a strong preference toward speci c food presentation (meal presentation)," "I eat a limited number of items from each food group (food variety)," "I usually feel that I have something better to do than eating (meal disengagement)," and "I reject bitter foods, even if they are only slightly bitter (taste aversion)." Participants completed 16 items, on a ve-point Likert scale (1 = Never, 5 = Always). The APEQ has shown strong psychometric properties [17]. Higher scores indicated greater PE.

Eating Disorder Examination Questionnaire (EDE-Q)
Disordered eating was self-reported using the EDE-Q [26]. The 28-item scale assesses attitudes, cognitions, and behaviors relating to restraint, weight, shape, and eating concerns on a seven-point Likert scale (1 = No days; 7 = Everyday). This study focused on the eating concerns subscale. Each of the 28 items is preceded with "on how many of the past 28 days" and followed by the statements such as "Have you gone for long periods of time (8 waking hours or more) without eating anything at all in order to in uence your shape or weight?" The EDE-Q has strong psychometric properties [26]. The mean is calculated across all items in each scale, and higher scores indicate greater disordered eating.

Depression, Anxiety and Stress Scale − 21 Items (DASS-21)
Mental health concerns were assessed using the DASS-21, a short version of the 42-item instrument [28].
The DASS-21 measures depression, anxiety, and stress symptoms on a four-point Likert scale (0 = Did not apply to me at all; 3 = Applied to me very much, or most of the time). This study averaged scores from the three subscales for a total score, titled mental health concerns. Example items are "I experienced trembling (anxiety)," "I felt that I had nothing to look forward to (depression)," and "I found it hard to wind down (stress)." Higher scores indicate greater mental health concerns.
In exible Eating Questionnaire (IEQ) In exible eating was assessed using the IEQ [29]. The 11-item scale assesses engagement in in exible and rigid eating rules on a ve-point Likert scale (1 = Fully disagree; 5 = Fully agree). This study used the scale's total score with items such as, "To manage my eating through rules gives me a sense of control" and "not following my eating rules makes me feel inferior." The IEQ has high internal consistency, construct and temporal stability [29], and higher scores indicate greater in exible eating. Procedure Procedures were approved by the Institutional Review Board prior to data collection (IRB protocol #1530232). University instructors shared a recruitment script about the survey with their students.
Interested students then provided informed consent and accessed the survey on Qualtrics.
Participants provided demographic information followed by previously described self-report measures; median survey completion time was approximately 27 minutes. Participants received course credit or extra credit following survey conclusion and were asked to recruit their parents to complete a similar survey. However, only student responses were used in the present study.
Model 1 moderation analyses [30] were calculated in PROCESS macro in SPSS 27. Moderation analyses examined if PE facets (i.e., meal presentation, food variety, meal disengagement, taste aversion) interacted with negative psychological correlates (i.e., mental health concerns and in exible eating) in relation to eating concerns. Moderation models in PROCESS macro calculate an overall effect of all independent variables, including covariates, on the dependent variable; main effects of the independent variables and covariates on the dependent variable; the interaction of the independent variables on the dependent variable; and conditional moderation effects, which plot the interaction at various levels of the moderator (e.g., -1 SD below average, average, and + 1 SD above average). Eight moderation analyses were calculated, four examining the interaction of PE facets and mental health concerns on eating concerns, and four examining the interaction of PE facets and in exible eating on eating concerns. BMI (mean centered) and gender (0 = men, 1 = women) were included in moderation analyses given they are related to the study variables [31][32] and thus could confound ndings if unaddressed.

Preliminary Results
See Table 1 for descriptive statistics and bivariate correlations. There were signi cant positive bivariate correlations between each facet of PE and eating concerns and negative psychological correlates ( Table 1). Assumptions of multiple regression were con rmed via visual inspection of histograms, Q-Q plots, and scatterplots. Diagnostics con rmed there were no issues of multicollinearity (e.g., tolerance and variance in ation factors within acceptable ranges; [33]).

Negative Psychological Correlates as Moderators of Relationships Between Picky Eating Facets and Eating Concerns
The overall models assessing mental health concerns and meal presentation (R 2 = .33, p < .001), food variety (R 2 = .31, p < .001), meal disengagement (R 2 = .36, p < .001), and taste aversion (R 2 = .31, p < .001) contributed signi cant variance in eating concerns. Furthermore, mental health concerns signi cantly interacted with meal presentation in relation to eating concerns (b = .31, p = .001, ΔR 2 = .02, ΔF(1,475) = 11.19) such that meal presentation was more strongly associated with eating concerns when mental health concerns were higher (Fig. 1). While interactions between meal disengagement (b = .15, p = .05) and taste aversion (b = .14, p = .06) and mental health concerns were not signi cant in relation to eating concerns, conditional moderation effects were observed such that relationships between meal disengagement and taste aversion were more strongly associated with eating concerns when mental health concerns were higher. Interactions between mental health concerns and food variety were not signi cant, nor were conditional moderation effects present in relation to eating concerns (Table 2). Signi cant effects were retained with inclusion of BMI and gender covariates ( Table 2).

Discussion
First, we explored relationships between PE facets, eating concerns, and negative psychological correlates. Eating concerns and PE facets such as meal presentation and meal disengagement were positively related with medium strength, whereas eating concerns and food variety and taste aversion had small positive relations. Thus, some PE facets were more strongly related to eating concerns than others, a nding that is complemented by our bivariate data that suggest that meal presentation and meal disengagement were also more strongly related to mental health concerns and in exible eating. These ndings map onto previous research suggesting that PE is positively associated with signi cant psychological impairment [13-14, 17, 22, 24-25], as well as previous research that meal presentation and meal disengagement speci cally were related to a range of negative psychosocial correlates [17].
Building on these bivariate data, in the second aim, moderation analyses examined negative psychological correlates as moderators of relationships between PE facets and eating concerns. Meal presentation was more strongly associated with eating concerns when mental health concerns and in exible eating were higher. Furthermore, meal disengagement and taste aversion were more strongly associated with eating concerns when in exible eating was higher. Put another way, PE attitudes and behaviors such as strong preferences for speci c food presentation, negative affect when food is not prepared/cooked the "right" way (meal presentation), avoiding mealtime, disengagement while sitting down for mealtime (meal disengagement), and rejection of bitter/sour food (taste aversion) were more strongly associated with eating concerns in undergraduates who self-reported higher mental health concerns and in exible eating. This expands on previous research [13] by demonstrating that speci c facets of PE such as meal presentation, meal disengagement, and taste aversion may explain these relationships. Though models examining mental health concerns as a moderator of relationships between meal disengagement and taste aversion and eating concerns were non-signi cant, conditional moderation effects revealed that at higher levels of mental health concerns, meal disengagement and taste aversion were more strongly associated with eating concerns. Although these preliminary data should be interpreted with caution, they suggest that future research could add to our understanding of their generality (e.g., clinical versus non-clinical PE and disordered eating presentations).
Perhaps surprisingly, for the PE facet of food variety, no signi cant interaction nor conditional moderation effects were evidenced in relation to eating concerns. This pattern of null ndings is notable as food variety (i.e., eating a limited number of foods from known food groups (e.g., fruits, vegetables), not trying new types of foods, and eating from a narrow range of foods [17]) often sits at the forefront of PE conceptualizations [e.g., 1,16], both in public and academic domains. Interestingly, this effect may map onto food variety re ecting food exposure whereas meal presentation, meal disengagement, and taste aversion may re ect more anxiety and/or control around food. Findings from the present work add to existing research mapping PE and disordered eating [13][14][15] such that it generates questions about precisely which facets of PE explain these relationships. Importantly, conclusions drawn from this research may bene t both researchers and clinicians through a more nuanced understanding of the complex links between PE, disordered eating, and negative psychological correlates.

Limitations And Future Directions
Of course, this study is not without limitations and future directions. First, cross-sectional data do not allow us to make causal attributions across study variables. While the statistical analyses, speci cally moderation analyses, allow us to discern the role of both moderators on the relationship between PE facets and eating concerns, it is important for future research to con rm these exploratory ndings with both experimental and experience sampling (e.g., ecological momentary assessment) methods to better map the temporal order and mechanistic links between PE, negative psychological correlates, and eating concerns. Second, though large and well-powered, the sample is convenient and primarily WEIRD (i.e., Western, Educated, Industrialized, Rich, and Democratic [34]); thus, ndings are bound to undergraduate participants. Still, these effects suggest several important research avenues to build on constraints on generality [35], including how these effects unfold in more severe PE, disordered eating, and mental health populations, as well as how these effects unfold across development (e.g., from childhood to adulthood). Finally, data collected in the present study were self-reported; thus, concerns common to descriptive, survey-based research such as social desirability and recall biases may compromise the validity of these ndings. Future research using diverse methods to probe these constructs (e.g., behavioral) may build and substantiate observed effects.

Conclusions
Understanding precisely which PE facets interact with negative psychological correlates in relation to eating concerns provides useful information to researchers interested in targeting mutual mechanisms of PE and disordered eating, speci cally eating concerns. The present study suggests that PE facets such as meal presentation, meal disengagement, and taste aversion (but not limited food variety) were more strongly associated with eating concerns when mental health concerns and in exible eating were higher. These data inform clinical practice through knowledge of mental health concerns and in exible eating as correlates that may worsen relationships between PE facets and disordered eating, which in turn may provide information about potential therapeutic targets in concurrent PE and disordered eating presentations.

What Is Already Known On This Subject?
Adult picky eating (PE) is positively associated with signi cant psychological impairment and disordered eating. Some negative psychological correlates such as mental health concerns and in exible eating may strengthen the relationship between PE and disordered eating, speci cally eating concerns, but which facets of PE describe these relationships is unknown.

What Does This Study Add?
Maps relationships between PE facets, eating concerns, mental health concerns, and in exible eating. Identi ed speci c PE facets (e.g., meal presentation, meal disengagement, taste aversion) that are more strongly related to eating concerns in adult participants with higher mental health concerns and in exible.

Declarations
Con icts of Interest: None of the authors have potential con icts of interest to disclose. Ethics Approval: This study was performed in line with the principles of the Declaration of Helsinki. The present study received approved IRB review prior to data collection by Bowling Green State University's Institutional Review Board prior to data collection (#1530232) Consent to Participate: All participants gave online informed consent prior to participating in the present study.
Consent for Publication: All participants have consented to having de-identi ed, group data published in a journal article.
[33] Belsley DA, Kuh E, Welsch RE (1980) Regression diagnostics: Identifying in uential data and sources of collinearity. New York, NY: John Wiley.      Conditional moderation of taste aversion and eating concerns by in exible eating.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download. Tables.pdf