# paper.pdf ## Metadata - PDFFormatVersion=1.4 - Language=en-US - IsLinearized=false - IsAcroFormPresent=false - IsXFAPresent=false - IsCollectionPresent=false - IsSignaturesPresent=false - CreationDate=D:20260508164938+02'00' - Creator=Adobe InDesign 18.3 (Windows) - Custom.CrossMarkDomains[1]=springer.com - Custom.CrossMarkDomains[2]=springerlink.com - Custom.CrossmarkDomainExclusive=true - Custom.CrossmarkMajorVersionDate=2010-04-23 - Custom.doi=10.1186/s40337-026-01589-8 - ModDate=D:20260508182329+02'00' - Producer=Adobe PDF Library 17.0; modified using iText® 5.5.13.4 ©2000-2024 iText Group NV (SPRINGER SBM; licensed version) - Trapped=False - crossmark:crossmarkdomains=springer.comspringerlink.com - pdfx:crossmarkdomains=springer.comspringerlink.com ## Contents ### Page 1 © The Author(s) 2026. Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use,sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) andthe source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in thisarticle are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not includedin the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you willneed to obtain permission directly from the copyright holder. To view a copy of this licence, visit. * Co rre s p o n d e n c e : Julia Ulbrich -leipzig.de Full list of author information is available at the end of the article Abstract Background The Child Eating Behaviour Questionnaire (CEBQ) is an internationally applied, parent-reportquestionnaire on children’s eating behaviors, but has mainly been validated in population-based samples or childrenwith obesity. This study presents a first comprehensive validation of the German version of the CEBQ in a treatment-seeking and community-based sample including anorexia nervosa (AN), avoidant-restrictive food intake disorder(ARFID), loss of control (LOC) eating, attention-deficit/hyperactivity disorder (ADHD), and a healthy control group. Methods The German version of the CEBQ was completed by 226 parents of children and adolescents (9 monthsto 17 years) in Germany and Switzerland. Factorial, convergent, and discriminant validity, internal consistency, as wellas sociodemographic correlates were assessed, using objectively measured anthropometrics and well-establishedclinical interviews and questionnaires on eating disorders and associated psychopathology in parent- and self-report. Results The original 8-factor structure showed acceptable model fit and acceptable to excellent internal consistency.Convergent validity was mostly supported by weight status and interview- and questionnaire-measured eatingbehaviors. The CEBQ subscales differentiated between groups associated with overeating (ADHD, LOC eating) versusrestrictive eating (ARFID and AN). Conclusion The results support the German version of the CEBQ as a valid and reliable tool for assessing eatingbehavior in youth with eating disorders and ADHD. Future research using larger and age-specific samples shouldexamine psychometric comparability across developmental stages and may provide norms for the CEBQ to enhanceits utility as a screening tool. Plain English summary The Child Eating Behaviour Questionnaire (CEBQ) is an internationally used questionnaire completed by parents todescribe different eating behaviors in children. This study tested whether the German version of the questionnaireis suitable not only for healthy children or children with obesity, as shown previously, but also for children andadolescents with eating disorders or eating difficulties, such as anorexia nervosa (AN), avoidant-restrictive food Validation of the German Child EatingBehaviour Questionnaire (CEBQ) in childrenand adolescents with eating disordersand ADHD Julia Ulbrich1*[ ![http://orcid.org/0009-0008-2949-6675]() ](http://orcid.org/0009-0008-2949-6675) , Anja Hilbert1,2[ ![http://orcid.org/0000-0003-2775-1296]() ](http://orcid.org/0000-0003-2775-1296) , Simone Munsch3,4[ ![http://orcid.org/0000-0002-6187-8912]() ](http://orcid.org/0000-0002-6187-8912) and Ricarda Schmidt1,2[ ![http://orcid.org/0000-0001-7794-2897]() ](http://orcid.org/0000-0001-7794-2897) Ulbrich et al. Journal of Eating Disorders (2026) 14:108 [ ![http://crossmark.crossref.org/dialog/?doi=10.1186/s40337-026-01589-8&domain=pdf&date_stamp=2026-5-8]() ](http://crossmark.crossref.org/dialog/?doi=10.1186/s40337-026-01589-8&domain=pdf&date_stamp=2026-5-8) ![BioMed Central Logo]() ### Page 2 Background Eating disorders (EDs) are characterized by significantand persistent disturbances in eating behavior associatedwith adverse effects on physical or psychosocial health[]. A distinction can be made between restrictive EDs,such as anorexia nervosa (AN) and avoidant/restrictivefood intake disorder (ARFID), and EDs with binge eat-ing, such as bulimia nervosa (BN) and binge-eating dis-order (BED) []. AN and ARFID are characterized byrestrictive food intake, motivated by weight and shapeconcerns in patients with AN and sensory sensitivities,fear of aversive consequences, or lack of interest in eat-ing in those with ARFID []. In contrast, following the5th Edition of the Diagnostic and Statistical Manual ofMental Disorders (DSM-5), BED and BN are character-ized by objective binge-eating episodes (OBEs), definedby eating objectively large amounts of food accompaniedby a feeling of loss of control (LOC) over eating, eitherwith (for BN) or without (for BED) regular compensa-tory behaviors []. According to the 11th revision of theInternational Classification of Diseases (ICD-11), BEDmay additionally include binge-eating episodes over eat-ing a subjectively large amount of food [], which can bereferred to as subjective binge-eating episodes (SBEs).LOC eating describes the subjective perception of beingunable to control eating, irrespective of the amountconsumed, thus includes both OBEs and SBEs, whichis especially relevant in children and adolescents [,].Additionally, other mental disorders may be associatedwith an increased risk of disordered eating behaviors orcomorbid EDs. Specifically, children with attention-def-icit/hyperactivity disorder (ADHD) may show increasedrisk for overeating behavior, BED, and BN [,]. Reliableand valid measures to assess eating behavior across theseconditions, both in self-report and parent-report, arecrucial []. The Child Eating Behaviour Questionnaire (CEBQ),developed by Wardle et al. [], is a well-established par-ent-reported questionnaire assessing the eating behav-ior of children in eight dimensions. Based on 35 items,four subscales depict food avoidance behaviors (Sati-ety Responsiveness [SR], Slowness in Eating [SE], FoodFussiness [FF], Emotional Undereating [EUE]) and foursubscales cover food approach behaviors (Food Respon-siveness [FR], Enjoyment of Food [EF], Desire to Drink[DD], Emotional Overeating [EOE]). The CEBQ has beentranslated, validated, and used in many studies world-wide (see Table for an overview of CEBQ validationstudies). The psychometric properties of the CEBQ havemainly been examined in community or school-basedsamples of children aged 1–16 years [e.g.,,] and insamples enriched with treatment-seeking children withoverweight or obesity aged 3–13 years [,,]. Mostvalidation studies focused on factorial validity, internalconsistency, and associations of CEBQ subscales withchild weight status. Although the CEBQ has recentlybeen used in studies on ARFID [e.g.,,] and ADHD[,], it has never been validated in samples of indi-viduals with interview-assessed EDs or ADHD. Theonly available study by Ko et al. [] reported acceptableinternal consistency, test-retest reliability, and sufficientitem-subscale correlations for the Korean version of theCEBQ in a sample of 2–9-year-old children with parent-reported symptoms of infantile anorexia assessed via anonline survey. However, factorial and convergent validitywere not examined. The factor structure of the CEBQ has not been repli-cated consistently. Studies using Confirmatory FactorAnalysis (CFA), which is the recommended method forthe factorial validation of questionnaires with a hypoth-esized factor structure [], predominately showed poorto acceptable model fit for the theory-based original8-factor structure. A modified 8-factor structure basedon minor adjustments, such as cross loadings or removalof items, showed improved model fit [e.g.,,]. A recentstudy in a large German sample of primary school chil-dren aged 6–11 years showed acceptable to good modelfit for a shortened 24-item version []. Overall, inter-nal consistency for the original and modified 8-factorstructure was acceptable to good. Using Exploratory Fac-tor Analysis (EFA), various factor models with 3–9 fac-tors have been proposed, but most factor models havenot been replicated [e.g.,,]. However, two different7-factor structures, combining SR and SE [e.g.,,] or intake disorder (ARFID), loss of control (LOC) eating, and attention-deficit/hyperactivity disorder (ADHD). A total of226 parents of children and adolescents in Germany and Switzerland completed the questionnaire. The results, asexpected, showed that the questionnaire measured eight aspects of eating behavior in a reliable and consistentway. The parents’ answers corresponded well with the children’s weight and information from clinical interviewsand conceptually related questionnaires. The CEBQ also revealed differences between groups of children oradolescents who tend to overeat (ADHD, LOC eating) compared to those who restrict their eating (ARFID and AN).Overall, the findings provide support for the German CEBQ as a useful tool for assessing eating behaviors in youthacross clinical groups, including eating disorders and ADHD. Keywords Child eating behavior questionnaire (CEBQ), Validation, Eating disorders, ADHD, Psychometric properties,Factor analysis, Assessment ### Page 3 | Study | Country | Method | N | Age,years | Recruitment setting | Factors, model fit | Note on factors | | - | - | - | - | - | - | - | - | | Wardle et al., 2001 [] | UK | PCA | 208 | 2–9 | School-based | 7a | SE and SR combined | | Viana et al., 2008 [] | Portugal | PCA | 240 | 3–13 | School-based, health-care setting (obesity) | 6 | EF and FR combined,SE and SR combined | | Sleddens et al., 2008 [] | Netherlands | PCA | 135 | 6–7 | School-based | 7 | EOE and FR combined | | Santos et al., 2011 [] | Chile | PCA | 294 | 6–12 | School-based, health-care setting (obesity) | 7 | SE and SR combined | | Svensson et al., 2011 [] | Sweden | PCA | 174 | 1–6 | School-based | 7 | EOE and FR combined | | Sparks and Radnitz, 2012[] | USA | CFA | 229 | 2–5 | School-based | 7 poor-acceptableb | SE and SR combined | | PCA | 3 (15 items) | Disinhibition, FoodInterest, Undereating | | Cao et al., 2012 [] | China | PCA | 219 | 1–1.5 | Population-based | 7 | FR split into two fac-tors, no EF, no SR | | Loh et al., 2013 [] | Malaysia | PCA | 362 | 13 | School-based | 9 | FF split into 2 factors | | CFA | 8 poor-acceptableb | FF split into 2 factors,no SR | | Mallan et al., 2013 [] | Australia | CFA | 663 | 1–5 | School-based,population-based | 8o, m poor-goodb | - | | Domoff et al., 2015 [] | USA | CFA | 1002 | 3–4 | School-based | 8o, macceptable-goodb | - | | 3 poorb | Based on Sparks andRadnitz, 2012 | | Ek et al., 2016 [] | Sweden | CFA | 478 | 3–8 | School-based, health-care setting (obesity) | 8o poorc | - | | 7 poorc | EOE and FR combined | | 8m poor-goodb | - | | Sirirassamee and Hun-changsith, 2016 [] | Thailand | PCA | 680 | 6–11 | School-based | 8 | - | | Njardvik et al., 2018 [] | Iceland | Compara-tive CFA | 560 | 5–12 | School-based | 8o poor-acceptableb | - | | 7 poor-acceptableb | SE and SR combined | | 7 poor-acceptableb | EOE and FR combined | | 6 poorb | EF and FR combined,SE and SR combined | | Quah et al., 2017 [] | Singapore | CFA | 636 | 3 | Population-based | 8o, mpoor-acceptableb | - | | EFA | 7 | No FR | | Quah et al., 2019 [] | Singapore | CFA | 1102 | 5–6 | Population-based | 8o, m poorb | - | | EFA | 6 | No FR, no SR | | Spahić and Pranjić, 2019[] | Bosnia | PCA | 2500 | 3–10 | School-based,population-based | 8 | - | | Purwaningrum et al., 2020[] | Indonesia | CFA | 238 | 2–5 | Population-based | 7 poor-goodb | EOE and FR combined | | 8macceptable-goodb | - | | Gebru et al., 2021 [] | Ethiopia | CFA | 542 | 3–6 | School-based | 8m poor-goodb | - | | Oyama et al., 2021 [] | Samoa | CFA | 113 | 1 (21months) | Healthcare setting(birth cohort) | 8o, mpoor-acceptableb | - | | PCA | 3 (15 items) | Dysregulated Eating,Joy of Eating, Attrac-tion to Food | | Al-Hamad et al., 2021 [] | Saudi-Arabia | PCA | 200 | 2–6 | Healthcare setting(healthy, acute care) | 8 | - | | Ayre et al., 2022 [] | Vietnam | EFA | 102 | 2–5 | Healthcare setting | 7 | No SR | | 6 | EF and FR combined,no SR | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | Table 1 Published factor structures of the CEBQ ### Page 4 EOE and FR (overeating) [e.g.,,] have been repeat-edly replicated in previous studies and showed acceptableor good model fit (see Table). CFAs in samples of chil-dren aged 2–5 years in Indonesia or 5–12 years in Icelandshowed a significant advantage of model fit for the 8-fac-tor over both hypothesized 7-factor structures [,]. Findings regarding age and sex differences in CEBQsubscales were not consistent, but there was a trend forboys to score higher on food approach subscales andgirls to score higher on food avoidance subscales [,].Associations between children`s standardized body massindex (BMI, kg/m²) and CEBQ subscales have been dem-onstrated in most studies, with food approach subscalesbeing positively and food avoidance subscales being neg-atively correlated with standardized BMI scores []. The aim of this study was to evaluate the psychomet-ric properties of the German version of the CEBQ inchildren and adolescents with AN, ARFID, LOC eat-ing, ADHD as well as a healthy control group. To ourknowledge, the present study is the first to examine thepsychometric properties of the CEBQ in a sample of chil-dren and adolescents with EDs and ADHD determinedvia clinical interview. Validating the CEBQ in childrenand adolescents with ADHD addresses an important gapby extending its applicability to a clinically relevant popu-lation at increased risk for dysregulated eating behaviorsthat may not meet full-threshold diagnostic criteria forEDs. Finally, a validation of the complete questionnaire inan exclusively German-speaking sample is lacking so far. Specifically, this study sought to (1) examine item char-acteristics, (2) investigate the factor structure by compar-ing repeatedly proposed factor models, specifically, theoriginal 8-factor structure [] and two 7-factor structuresthat showed a combined SR and SE factor [e.g.,,] orcombined EOE and FR factor [e.g.,,], (3) examineinternal consistency, (4) evaluate the convergent validityof the CEBQ by analyzing the association between CEBQsubscales and standardized BMI as well as well-estab-lished questionnaires on eating behaviors, (5) assess thediscriminant validity by analyzing differences in CEBQ | Study | Country | Method | N | Age,years | Recruitment setting | Factors, model fit | Note on factors | | - | - | - | - | - | - | - | - | | Manzano et al., 2021 [] | USA | EFA | 148 | 8–12 | Healthcare setting(overweight/ obesity) | 3 (27 items) | Picky Eating, Emo-tional Eating, Reward-based Eating | | Zhou & Sun, 2021 [] | China | CFA | 372 | School-based | 8m poor-good | - | | 8o poor-acceptable | - | | Hunot-Alexander et al.,2022 [] | Mexico | CFA | 842 | 3–13 | Healthcare setting(healthy, acute care) | 8o poor-acceptableb | - | | Jimeno-Martinez et al.,2022 [] | Spain | CFA | 197 | 3–6 | School-based, health-care setting (at risk forobesity) | 8m goodc | - | | Malczyk et al., 2022 [] | Poland | PCA | 151 | 3–10 | Details not reported | 8 | - | | Mou et al., 2024 [] | China | PCA | 11780 | 2–5 | School-based,population-based | 8 | - | | Peuckert et al., 2025 [] | Brazil | EFA | 205 | 3–13 | School-based | 5 | - | | García et al., 2025 [] | Spain | CFA | 283 | 6–16 | School-based, popula-tion based | 8o poor-goodb | - | | Nigam et al., 2025 [] | USA | EFA | 814 | 2–12 | Healthcare setting(obesity) | 6 | EF and FR combined,SE and SR combined | | Ruden & Warschburger,2025 [] | Germany | EFA | 1337 | 6–11 | School-based,population-based | 7 (24 items) | SE and SR combined | | EFA | 8 (24 items) | - | | CFA | 8 (24 items)acceptable-goodb | - | | PCA Principal Components Analysis, CFA Confirmatory Factor Analysis, EFA Exploratory Factor Analysis, SR Satiety Responsiveness, SE Slowness in Eating, FF FoodFussiness, EUE Emotional Undereating, FR Food Responsiveness, EF Enjoyment of Food, DD Desire to Drink, EOE Emotional OvereatingDefinition of recruitment settings: School-based: Recruited through educational institutions, including preschools/ childcare centers. Population-based: Recruitedto represent a defined geographic population. Healthcare setting: Recruited through hospitals, outpatient clinics, or health services. Included diagnostic groupsand further specifications of sample characteristics are indicated in parenthesesa result of factor analysis deviating from hypothesized factor structure based on theoretical considerations, items retained in hypothesized scalesb based on following fit indices: Model-fit is acceptable if: 2.00 < χ2/df ≤ 3.00, 0.90 ≤ NFI < 0.95, 0.95 ≤ CFI < 0.97, 0.90 ≤ TLI < 0.95, 0.08 < SRMR ≤ 0.10. and 0.06 < RMSEA≤ 0.10. Model fit is good if: χ2/df ≤ 2.00, NFI ≥ 0.95, CFI ≥ 0.97, TLI ≥ 0.95, SRMR ≤ 0.08 and RMSEA ≤ 0.06 [,]c no values specified, evaluation based on own statement of the authorso original 8-factor model without modificationsm modified version of the original 8-factor model, e.g., cross loadings or removal of items, use of modification indices | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | Table 1 (continued) ### Page 5 subscales between children and adolescents with ARFID,AN, LOC eating, ADHD, and a healthy control group,and (6) examine differences in sex and age. Methods Participants and study design Data for the present study (N = 230) were derived from astudy on the validation of the ARFID module 2.0 for theEating Disorder Examination (EDE) [] and the SwissUniversity Study of Nutrition (SUN) []. The studieswere approved by the local Ethics Committees (Leipzig,Germany/Fribourg, Switzerland), and informed consentand assent was provided by parents and children ≥ 8 yearsold. Participants of the validation study on the ARFIDmodule 2.0 for the EDE (N = 170) [] were recruited atLeipzig University Medical Center, a support group forindividuals with ARFID, advertisements on the Internet,and from the population-based Leipzig Research Centerfor Civilization Diseases (LIFE) Child study [] betweenFebruary 2018 and May 2021. The sample comprised atreatment-seeking group with avoidant/restrictive eat-ing behaviors, including current and lifetime diagnosesof ARFID (n = 39) or AN (n = 24), determined using theEDE interview [,], or its child version [,] andthe ARFID module 2.0 []. A control group recruitedfrom the population following a telephone screening toexclude symptoms of EDs was matched for age and sex.Inclusion criteria involved age between 0 and 17 yearsand sufficient German language skills, and no specificexclusion criteria were applied. German-speaking participants of the SUN study(n = 60) were recruited from primary school screeningsin Switzerland between November 2011 and July 2014.The presence of LOC eating and ADHD was determinedthrough the EDE adapted for children [,] and theSchedule of Affective Disorders and Schizophrenia forSchool-age Children, Present and Lifetime Version [,]. Inclusion criteria were an age between 8 and 13years, regular school attendance, and sufficient Germanlanguage skills. Exclusion criteria included > 1 episodeof compensatory behavior during the past 3 months,AN, BN, psychotic disorders, current overweight treat-ment, medication affecting eating, and serious medicalproblems. Participants with LOC eating (n = 14) wererequired to have ≥ 3 LOC eating episodes over the past3 months with distress and/or ≥ 2 out of 5 behavioralsymptoms. Children with ADHD (n = 24) were requiredto fulfill Diagnostic and Statistical Manual of Mental Dis-orders, Fourth Edition, Text Revision (DSM-IV-TR) orDSM-5 criteria for ADHD, and medical treatment wasallowed if intake was stable over the previous 3 months[,]. Children who met the inclusion criteria for bothLOC eating and ADHD were included in the LOC eatinggroup. A control group (n = 22) without lifetime or cur-rent LOC eating, compensatory behavior, ED or ADHDsymptoms was matched for age, sex, and BMI percentileto the LOC group. Because a completion of at least 75% of the CEBQitems was required for inclusion in this study, n = 4 par-ticipants were excluded, leaving a total sample of N = 226participants (see Table for descriptive statistics). Childage in the total sample ranged from 9 months to 17years, with a mean age of 9.81 (SD = 4.79) years. Mostchildren had German (n = 163, 72.1%) or Swiss (n = 57,25.2%) nationality. The mean BMI standard deviationscore (BMI-SDS) was − 0.47 (SD = 1.31), ranging from− 4.39 ≤ BMI-SDS ≤ 3.66. Most of the children were ofnormal weight (n = 150, 66.4%). Measures Child eating behaviour questionnaire The CEBQ [] is a 35-item parent-report questionnaireassessing child food avoidance and food approach behav-iors with 8 subscales (see Background). All items areanswered on a 5-point Likert Scale ranging from 1 = neverto 5 = always and subscale mean scores are calculated,with higher scores indicating a stronger representation ofthe corresponding eating behavior. Reverse-coded items(3, 4, 10, 16, and 32) were recoded prior to all analysesby inverting the 5-point Likert scale (1 to 5, 2 to 4, 3 to 3,4 to 2, 5 to 1). The authorized German translation of theCEBQ was performed by Anja Hilbert and controlled in abacktranslation procedure by a licensed translator, bothof whom had language expertise and cultural knowledge.A congruence check confirmed consistency between theoriginal and backtranslated versions. Measures for convergent validity Body mass index standard deviation score Objectivelymeasured body height and weight were used to calculatethe BMI (kg/m2). Based on age- and sex-specific Germannormative data [], BMI-SDS was computed. Positivecorrelations with food approach subscales and negativecorrelations with food avoidance subscales were assumed. Eating disorder examination The EDE [,] and itschild version [,] were used to assess the number ofOBEs and SBEs during the past month. Positive correla-tions with food approach subscales and negative correla-tions with food avoidance subscales were assumed. EDE ARFID module 2.0 The ARFID module 2.0 for theEDE [] in its child and parent version was used to assessthe number of days with avoidant/restrictive food intakeduring the past month. Positive correlations with food ### Page 6 avoidance subscales and negative correlations with foodapproach subscales were assumed. Eating disturbances in youth-questionnaire (EDY-Q) The EDY-Q [,] is a 14-item self-report ques-tionnaire evaluating restrictive eating disturbances on a7-point Likert scale (0 = never to 6 = always). A total meanscore can be obtained from items 1–5 and 8–12, withhigher scores indicating greater restrictive eating distur-bances. Cronbach`s Alpha (α) and McDonalds Omega (ω)were 0.75 and 0.73 for this sample. Positive correlationswith food avoidance subscales and negative correlationswith food approach subscales were assumed. Data analytic plan All statistical analyses were performed using IBM SPSSStatistics version 29.0.0.0 and IBM SPSS Amos version 26for Windows. The data analytic plan and hypotheses werepre-specified. Item analyses Item distributions were tested for normality using theShapiro-Wilk and Kolmogorov normality tests and byitem skewness and kurtosis, interpreted as normally(absolute value = 0.00), slightly non-normally (absolutevalue < 1.00), moderately non-normally (1.00 ≤ abso-lute value ≤ 2.30), and severely non-normally distributed(absolute value > 2.30) []. Item difficulty was calculatedas pm = sum of item scores/ (N * maximal item score), | Table 2 Descriptive statistics for the total sample and diagnostic subgroups | | - | | Total sampleN = 226 (100%) | ARFIDn = 39 (17.3%) | ANn = 24 (10.6%) | LOC eatingn = 14 (6.2%) | ADHDn = 24 (10.6%) | Controln = 124 (54.9%) | | Child sociodemographics and anthropometrics | | Sex (n, %) | | Girls | 126 (55.8%) | 16 (41.0%) | 23 (95.8%) | 7 (50.0%) | 10 (41.7%) | 69 (55.6%) | | Boys | 100 (44.2%) | 23 (59.0%) | 1 (4.2%) | 7 (50.0%) | 14 (58.3%) | 55 (44.4%) | | Nationality (n, %) | | German | 163 (72.1%) | 39 (100%) | 24 (100%) | 0 | 0 | 99 (79.8%) | | Swiss | 57 (25.2%) | 0 | 0 | 14 (100%) | 22 (91.7%) | 21 (16.9%) | | Age, years (M, SD) | 9.81 (4.79) | 7.85 (5.24) | 14.83 (1.63) | 10.97 (1.02) | 10.37 (1.24) | 9.15 (5.03) | | Age group (n, %) | | 0–7 years | 63 (27.9%) | 20 (51.3%) | 0 | 0 | 0 | 43 (34.7%) | | 8–13 years | 113 (50.0%) | 12 (30.8%) | 6 (25.0%) | 14 (100%) | 24 (100%) | 57 (46.0%) | | 14–17 years | 50 (22.1%) | 7 (17.9%) | 18 (75.0%) | 0 | 0 | 24 (19.4%) | | Weight status (n, %) | | Severe underweight | 34 (15.0%) | 18 (46.2%) | 8 (33.3%) | 0 | 0 | 8 (6.5%) | | Underweight | 23 (10.2%) | 7 (17.9%) | 2 (8.3%) | 1 (7.1%) | 2 (8.3%) | 11 (8.9%) | | Normal weight | 150 (66.4%) | 14 (35.9%) | 14 (58.3%) | 8 (57.1%) | 18 (75.0%) | 95 (76.6%) | | Overweight | 10 (4.4%) | 0 | 0 | 3 (21.4%) | 2 (8.3%) | 5 (4.0%) | | Obesity | 9 (4.0%) | 0 | 0 | 2 (14.3%) | 2 (8.3%) | 5 (4.0%) | | BMI, kg/m2 (M, SD) | 17.54 (4.20) | 14.8 (1.58) | 17.29 (2.05) | 20.17 (3.53) | 18.11 (3.55) | 18.03 (4.86) | | BMI-SDS (M, SD) | -0.47 (1.31) | -1.54 (0.93) | -1.35 (-0.94) | 0.72 (1.12) | 0.12 (1.19) | -0.21 (1.20) | | BMI percentile (M, SD) | 39.01 (30.56) | 14.02 (17.57) | 18.83 (17.23) | 69.32 (30.50) | 52.56 (29.70) | 45.34 (28.78) | | Parent sociodemographics | | Age, years (M, SD) | 41.13 (7.64) | 37.58 (6.81) | 47.71 (4.91) | 43.18 (3.67) | 41.07 (4.52) | 40.65 (8.31) | | BMI, kg/m2 (M, SD) | 25.0 (5.69) | 24.90 (6.59) | 24.37 (4.51) | 24.84 (5.63) | 23.97 (5.14) | 25.38 (5.75) | | Family status (n, %) | | Single | 42 (18.6%) | 9 (23.1%) | 4 (16.7%) | 0 | 1 (4.2%) | 27 (21.8%) | | Married | 146 (64.6%) | 27 (69.2%) | 11 (45.8%) | 10 (71.4%) | 19 (79.2%) | 79 (63.7%) | | Separated | 9 (4.0%) | 0 | 3 (12.5%) | 0 | 4 (16.7%) | 6 (4.8%) | | Divorced | 27 (11.9%) | 2 (5.1%) | 6 (25.0%) | 4 (28.6%) | 0 | 11 (8.9%) | | Widowed | 1 (0.4%) | 0 | 0 | 0 | 0 | 1 (0.8%) | | University degree (n, %) | 60 (26.5%) | 10 (25.6%) | 7 (29.2%) | 3 (21.4%) | 1 (4.2%) | 39 (31.5%) | | German reference data [] for the BMI-SDS were used to classify child weight status into severe underweight (BMI-SDS ≤ -1.88), underweight (-1.88 < BMI-SDS ≤-1.28), normal weight (-1.28 < BMI-SDS < 1.28), overweight (1.28 ≤ BMI-SDS < 1.88), and obesity (BMI-SDS ≥ 1.88)ARFID avoidant/restrictive food intake disorder, AN anorexia nervosa, LOC loss of control, ADHD attention-deficit/hyperactivity disorder, BMI body mass index (kg/m2), BMI-SDS body mass index-standard deviation scoreDue to missing data, values may not sum up to N = 226 (100%)A detailed age-by-diagnosis distribution in single-year increments is provided in Supplementary Table S4 | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | ### Page 7 with scores closer to 1 indicating easier and closer to 0indicating more difficult items. For corrected item-sub-scale correlations Pearson`s r was assessed and inter-preted as sufficient if r ≥ .30 []. Missing values of CEBQitem responses were analyzed for frequency and patterns.Little’s MCAR test was significant (p < .05), indicatingthat the data were not missing completely at random(MCAR). However, the overall proportion of missingdata was very low (maximum 1.8% per item, 1–5 missingitems per participant, M = 0.19, SD = 0.73; see Table),with no systematic pattern across items or participantsand no association with observed variables, such asdiagnostic group, age, or sex. Based on these findings,missingness was considered consistent with a missing atrandom (MAR) assumption, and missing item responseswere imputed using the expectation-maximization algo-rithm for factor and reliability analyses. Factor analyses A CFA was conducted to compare the original 8-fac-tor structure [] and two previously replicated 7-factorstructures that showed a combined SR and SE factor [e.g.,,]) or combined EOE and FR factor [e.g., 23, 24]. Mul-tivariate normality was assessed through Mardia’s testand rejected if the absolute value of the critical ratio (c.r.)was > 1.96 []. In this case, Bollen-Stine bootstrapping | Table 3 Item characteristics of the CEBQ original 8-factor model, N = 226 | | - | | Item | Originalsubscale | M | SD | Skewness | Kurtosis | Item difficulty | Correcteditem-subscalecorrelation | Missing itemresponsesn (%) | | 1 | EF | 3.35 | 1.11 | -0.40 | -0.56 | 0.59 | 0.84 | 0 (0%) | | 2 | EOE | 1.38 | 0.71 | 1.96 | 3.29 | 0.10 | 0.69 | 3 (1.3%) | | 3* | SR | 3.05 | 1.03 | 0.17 | -0.38 | 0.51 | 0.58 | 2 (0.9%) | | 4* | SE | 3.14 | 1.15 | 0.04 | -0.98 | 0.54 | 0.36 | 4 (1.8%) | | 5 | EF | 3.47 | 1.09 | -0.45 | -0.50 | 0.62 | 0.58 | 0 (0%) | | 6 | DD | 2.87 | 1.16 | 0.22 | -0.81 | 0.47 | 0.55 | 1 (0.4%) | | 7 | FF | 3.02 | 1.27 | 0.02 | -1.04 | 0.51 | 0.83 | 1 (0.4%) | | 8 | SE | 3.06 | 1.23 | 0.20 | -1.08 | 0.52 | 0.76 | 0 (0%) | | 9 | EUE | 2.65 | 1.33 | 0.24 | -1.12 | 0.41 | 0.72 | 3 (1.3%) | | 10* | FF | 3.19 | 1.21 | -0.22 | -0.80 | 0.55 | 0.83 | 1 (0.4%) | | 11 | EUE | 3.15 | 1.11 | -0.22 | -0.66 | 0.54 | 0.60 | 0 (0%) | | 12 | FR | 2.43 | 1.03 | 0.21 | -0.89 | 0.36 | 0.57 | 0 (0%) | | 13 | EOE | 1.44 | 0.76 | 2.01 | 4.52 | 0.11 | 0.65 | 3 (1.3%) | | 14 | FR | 1.69 | 1.04 | 1.37 | 0.90 | 0.17 | 0.72 | 0 (0%) | | 15 | EOE | 1.24 | 0.52 | 2.08 | 3.46 | 0.06 | 0.59 | 1 (0.4%) | | 16* | FF | 3.05 | 1.23 | 0.11 | -0.98 | 0.51 | 0.79 | 0 (0%) | | 17 | SR | 2.83 | 1.20 | 0.20 | -0.86 | 0.46 | 0.55 | 0 (0%) | | 18 | SE | 2.12 | 1.17 | 0.85 | -0.21 | 0.28 | 0.61 | 0 (0%) | | 19 | FR | 1.61 | 0.93 | 1.57 | 1.92 | 0.15 | 0.72 | 0 (0%) | | 20 | EF | 3.24 | 1.07 | -0.30 | -0.51 | 0.56 | 0.81 | 0 (0%) | | 21 | SR | 3.01 | 0.96 | -0.17 | -0.56 | 0.50 | 0.67 | 2 (0.9%) | | 22 | EF | 3.32 | 1.17 | -0.46 | -0.58 | 0.58 | 0.85 | 1 (0.4%) | | 23 | EUE | 2.59 | 1.14 | 0.13 | -0.85 | 0.40 | 0.53 | 0 (0%) | | 24 | FF | 2.53 | 1.12 | 0.42 | -0.65 | 0.38 | 0.64 | 1 (0.4%) | | 25 | EUE | 2.57 | 1.29 | 0.21 | -1.19 | 0.39 | 0.75 | 2 (0.9%) | | 26 | SR | 3.10 | 1.03 | 0.09 | -0.57 | 0.53 | 0.68 | 3 (1.3%) | | 27 | EOE | 1.91 | 1.03 | 0.80 | -0.49 | 0.23 | 0.55 | 1 (0.4%) | | 28 | FR | 2.37 | 1.21 | 0.53 | -0.75 | 0.34 | 0.61 | 0 (0%) | | 29 | DD | 1.86 | 1.03 | 1.07 | 0.28 | 0.22 | 0.69 | 1 (0.4%) | | 30 | SR | 2.40 | 1.19 | 0.48 | -0.75 | 0.35 | 0.59 | 3 (1.3%) | | 31 | DD | 2.31 | 1.22 | 0.55 | -0.80 | 0.33 | 0.66 | 2 (0.9%) | | 32* | FF | 3.38 | 1.16 | -0.21 | -0.81 | 0.60 | 0.87 | 1 (0.4%) | | 33 | FF | 2.86 | 1.27 | 0.10 | -1.05 | 0.47 | 0.83 | 1 (0.4%) | | 34 | FR | 1.52 | 0.87 | 1.92 | 3.52 | 0.13 | 0.73 | 3 (1.3%) | | 35 | SE | 2.54 | 1.12 | 0.31 | -0.73 | 0.39 | 0.46 | 2 (0.9%) | | CEBQ Child Eating Behaviour Questionnaire, SR Satiety Responsiveness, SE Slowness in Eating, FF Food Fussiness, EUE Emotional Undereating, FR Food Responsiveness,EF Enjoyment of Food, DD Desire to Drink, EOE Emotional Overeating* Reversed items were previously inversed | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | ### Page 8 with n = 1000 bootstraps was applied. The factor variancewas set to 1 and correlations between each of the latentvariables were allowed. Initial analyses included no cross-loadings, and error terms were uncorrelated. Model fit was evaluated using various fit indices: χ2/df, Root Mean Square Error of Approximation (RMSEA),Comparative Fit Index (CFI), Normed Fit Index (NFI),Tucker-Lewis-Index (TLI), Standardized Root MeanSquared Residual (SRMR) and interpreted as acceptableif 2.00 < χ2/df ≤ 3.00, 0.90 ≤ NFI < 0.95, 0.95 ≤ CFI < 0.97,0.90 ≤ TLI < 0.95, 0.08 < SRMR ≤ 0.10. and 0.06 < RMSEA≤ 0.10 or as good if χ2/df ≤ 2.00, NFI ≥ 0.95, CFI ≥ 0.97,TLI ≥ 0.95, SRMR ≤ 0.08 and RMSEA ≤ 0.06 [,,].The Akaike Information Criterion (AIC) was assessedfor model comparison, with lower scores indicating bet-ter model fit []. Χ2 difference tests were performed formodel comparison of nested models. To identify the best-fitting model, modification indices were used to allow forcrossloadings and correlated error terms if plausible incontent. Factor loadings and squared multiple correla-tions were presented for the unmodified original 8-factormodel, with factor loadings < 0.40 considered as random[] and squared multiple correlations (variance in itemresponses explained by factor) interpreted as good if ≥0.40 []. Spearman rank correlation coefficients wereused to compute intercorrelations between the subscalesof the CEBQ, and interpreted as small (ρ ≤ 0.29), medium(0.30 ≤ ρ ≤ 0.49), and strong (ρ ≥ 0.50) []. All the following analyses were conducted for theunmodified original 8-factor model. Reliability Internal consistency was evaluated by Cronbach`s Alpha(α) and McDonalds Omega (ω) and considered as accept-able if α, ω ≥ 0.70, good if α, ω ≥ 0.80, and excellent if α,ω ≥ 0.90 []. Convergent validity Spearman rank correlation coefficients were used tocompute correlations between CEBQ subscales andmeasures for convergent validity (see 2.2.2) and inter-preted according to Cohen []. In the case of nonsig-nificant Spearman correlations between CEBQ subscalemean scores and the BMI-SDS, a Multivariate Analysis ofCovariance (MANCOVA) was performed on the effect ofweight group (severe underweight, underweight, normalweight, overweight and obesity) on CEBQ subscale meanscores to identify non-monotonic relationships, control-ling for age, sex, and diagnostic group. Discriminant validity To examine differences in CEBQ subscales among diag-nostic groups (ARFID, AN, LOC eating, ADHD, control),a MANCOVA was conducted, controlled for age, sex,and BMI-SDS. In case of significant multivariate effects,the MANCOVA was followed by one-way analyses ofcovariance (ANCOVAs) for each CEBQ subscale andfurther by univariate post-hoc tests. Effect sizes (partialη²) were calculated for the ANCOVAs and interpreted assmall (0.01), medium (0.06), or large (0.14) []. For post-hoc tests, significance was set at p < .05 (two-tailed) withadjustments according to Bonferroni correction, andeffect sizes (Cohen`s d) for significant univariate asso-ciations were calculated and interpreted as small (0.2),medium (0.5), or large (0.8) []. For LOC eating andADHD groups, higher mean scores were assumed in thefood approach subscales and lower mean scores in thefood avoidance subscales, and vice versa for ARFID andAN groups. Differences in CEBQ subscale mean scores by sex and agegroup Effects of sex (boys, girls) and age (0–7 years, 8–13 years,and 14–17 years) on CEBQ subscale mean scores wereexamined by separate MANCOVAs. The analyses werecontrolled for sex (except the MANCOVA on sex), age(except the MANCOVA on age), BMI-SDS, and diag-nostic group. As age groups were highly confounded bydiagnostic groups, only the control group (n = 124) wasincluded in this analysis. Results Item analyses Table shows item characteristics and subscale meanscores of the CEBQ. Shapiro-Wilk and Kolmogorov-Smirnov tests indicated non-normality for all items (p <.05). Most items showed slight to moderate non-normal-ity according to skewness and kurtosis, with exception ofitems 2, 13, 15 (part of the EOE subscale), and 34 (part ofthe FR subscale), which showed a severe positive devia-tion in kurtosis (-1.19 ≤ kurtosis ≤ 4.52). The item diffi-culty varied between 0.06 ≤ pm ≤ 0.60 and fell within anacceptable range, although there was a tendency towardsgreater difficulty. Notably, the EOE (0.06 ≤ pm ≤ 0.23) andthe FR subscales (0.13 ≤ pm ≤ 0.36) contained particularlydifficult items. Corrected item-subscale correlations weresufficient for all items (0.36 ≤ r ≤ .87). Factor analyses Mardia’s test indicated multivariate non-normality (c.r.= 20.89). Consequently, the Bollen-Stine bootstrappingmethod was applied. In comparison to two previouslyhypothesized 7-factor models, the original 8-factor modelshowed the best model fit, as indicated by the minimumAIC value and significant χ2 difference tests (p < .001),with fit indices indicating poor (NFI = 0.79, CFI = 0.88) toacceptable (χ2/df = 2.13, RMSEA = 0.07 and SRMR = 0.08)model fit (Table). Inspection of modification indices ### Page 9 revealed values up to 32.87. All items showed accept-able to high factor loadings (0.43–0.94; see Table). Thevariance in item responses explained by the latent fac-tor ranged between 18% and 88% and was interpreted asgood (> 40%) for most items, except for items 4, 5, 6, 12,17, 30, and 33. Based on modification indices, reasonable crossload-ings and correlated error terms were allowed (see Tableand suppl. Table S1 for details), which further signifi-cantly (p < .001) improved the model fit of the 8-factormodel, with some fit indices achieving a good model fit(χ2/df = 1.71, RMSEA = 0.06 and SRMR = 0.06). However,all items that were modified showed higher or equal fac-tor loadings on their original subscales, while factor load-ings on their new subscales were low (0.18 ≤ λ ≤ 0.44),indicating limited deviation from the hypothesized factorstructure. Given the confirmatory nature of the analysis,the lack of a strong theoretical rationale for the proposedmodifications, the risk of overfitting and capitalization onchance [,], and to support comparability with theexisting literature, the unmodified 8-factor model wasretained for subsequent analyses. Intercorrelations between CEBQ subscales are shownin Table. Significantly positive intercorrelations werefound between most food approach subscales, withmedium to large effect, and between food avoidance sub-scales, with small to large effect. Significant correlationsbetween food approach and food avoidance subscalesexhibited negative associations of up to large effect, withexception of small positive correlations between EOE andEUE and as well between DD and EUE. DD displayednonsignificant correlations with most subscales, exceptfor small to medium positive associations with EUE,EOE, and FR. Reliability All subscales of the CEBQ showed acceptable to excellentinternal consistency, with Cronbach`s α ranging between0.75 ≤ α ≤ 0.91 and McDonald`s ω ranging between0.78 ≤ ω ≤ 0.92 (Table). Convergent validity All significant correlations corresponded to the hypoth-esized directions (Table). The number of days withavoidant/ restrictive food intake as assessed by parent- orself-report significantly showed small to medium-sizedcorrelations with food avoidance subscales, except forthe SE subscale, while correlations with most of the foodapproach subscales were nonsignificant, except for the EFsubscale showing small (for self-report) or medium-sized(for parent-report) negative correlations. OBEs were sig-nificantly positively correlated with the subscales FR andEF, and negatively correlated with the subscales SR andSE, with small effects. Correlations between SBEs andCEBQ subscales were nonsignificant. The EDY-Q totalmean score significantly showed small to large correla-tions with CEBQ subscales, with positive associationswith food avoidance and negative associations with foodapproach subscales, except for EUE and DD. CEBQ sub-scales showed significantly small to medium-sized corre-lations with the BMI-SDS, with positive associations forfood approach and negative associations for food avoid-ance subscales, with exception of the DD subscale. Over-all, the DD subscale showed no significant correlationswith measures of convergent validity. To further explore possible non-monotonic relation-ships between weight status and CEBQ subscale meanscores, a MANCOVA was conducted. Weight groups dif-fered significantly in the CEBQ subscales, p < .001, F(32,852) = 3.44. The higher the weight status, the higher werethe mean values of the food approach subscales and thelower the mean values of the food avoidance subscales,except for the DD subscale. Specifically, significant groupdifferences were found for all subscales except for EOEand DD, with medium to large effects (Suppl. Table S2).Post-hoc tests showed small to large effects (Suppl. Fig-ure S1). Discriminant validity The diagnostic groups differed significantly in CEBQsubscales, F(32, 852) = 3.09, p < .001. Significant group | Table 4 CFA of the CEBQ: Model fit for the original 8-factor model and revised factor models, N = 226 | | - | | Model | Modifications | χ2/df | NFI | CFI | RMSEA | SRMR | TLI | AIC | | 8 factors | - | 2.13Acceptable | 0.79Poor | 0.88Poor | 0.07 Acceptable | 0.08Acceptable | 0.86Poor | 1329.94 | | 8 factors | Crossloadings, correlated error terms | 1.71Good | 0.83Poor | 0.92Poor | 0.06Good | 0.06Good | 0.91Acceptable | 1106.67 | | 7 factors (SE and SR combined) | - | 2.33Acceptable | 0.77Poor | 0.85Poor | 0.08Acceptable | 0.09Acceptable | 0.84Poor | 1436.36 | | 7 factors (FR and EOE combined) | - | 2.71Acceptable | 0.72Poor | 0.81Poor | 0.09Acceptable | 0.17Poor | 0.79Poor | 1650.49 | | CFA Confirmatory Factor Analysis, CEBQ Child Eating Behaviour Questionnaire, CFI Comparative Fit Index, NFI Normed Fit Index, RMSEA Root Mean Square Error ofApproximation, SRMR Standardized Root Mean Squared Residual, TLI Tucker-Lewis-Index, AIC Akaike Information Criterion, SR Satiety Responsiveness, SE Slownessin Eating, FR Food Responsiveness, EOE Emotional OvereatingModel-fit is acceptable if: 2.00 < χ2/df ≤ 3.00, 0.90 ≤ NFI < 0.95, 0.95 ≤ CFI < 0.97, 0.90 ≤ TLI < 0.95, 0.08 < SRMR ≤ 0.10. and 0.06 < RMSEA ≤ 0.10. Model fit is good if: χ2/df ≤ 2.00, NFI ≥ 0.95, CFI ≥ 0.97, TLI ≥ 0.95, SRMR ≤ 0.08 and RMSEA ≤ 0.06 [,,] | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | ### Page 10 differences were seen for most subscales, except for DDand EUE (p > .05) (p < .05; see Table), with medium tolarge effects. Post-hoc tests revealed medium to large significantunivariate effects (see Fig.). As expected, significantlymore food approach behaviors were seen for LOC andADHD groups relative to ARFID and AN, while the lattershowed greater food avoidance behaviors. Thus, signifi-cant differences emerged between groups associated withovereating versus restrictive eating patterns. In contrast,none of the subscales significantly differentiated withinthese subgroups, that is, between LOC and ADHD or | Table 5 Reliability and factorial validity of the original 8-factor model | | - | | Latent factor | Observed factor(Item) | Standardized regres-sion weight | Varianceexplainedby factor | | Subscale | M (SD) | Cronbach`s Alpha | McDonald`sOmega | | FR | 1.92 (0.81) | 0.85 | 0.85 | 12 | 0.61 | 0.37 | | 14 | 0.82 | 0.67 | | 19 | 0.81 | 0.66 | | 28 | 0.65 | 0.43 | | 34 | 0.81 | 0.65 | | EF | 3.35 (0.97) | 0.89 | 0.90 | 1 | 0.90 | 0.80 | | 5 | 0.60 | 0.36 | | 20 | 0.87 | 0.76 | | 22 | 0.94 | 0.88 | | EOE | 1.49 (0.61) | 0.79 | 0.80 | 2 | 0.79 | 0.63 | | 13 | 0.79 | 0.63 | | 15 | 0.68 | 0.46 | | 27 | 0.68 | 0.47 | | DD | 2.34 (0.96) | 0.79 | 0.80 | 6 | 0.61 | 0.37 | | 29 | 0.88 | 0.78 | | 31 | 0.76 | 0.58 | | SR | 2.86 (0.81) | 0.82 | 0.82 | 3* | 0.82 | 0.68 | | 17 | 0.56 | 0.32 | | 21 | 0.72 | 0.51 | | 26 | 0.75 | 0.56 | | 30 | 0.56 | 0.32 | | SE | 2.71 (0.88) | 0.75 | 0.78 | 4* | 0.43 | 0.18 | | 8 | 0.90 | 0.80 | | 18 | 0.72 | 0.52 | | 35 | 0.64 | 0.40 | | EUE | 2.74 (0.99) | 0.82 | 0.83 | 9 | 0.84 | 0.70 | | 11 | 0.63 | 0.40 | | 23 | 0.55 | 0.30 | | 25 | 0.89 | 0.79 | | FF | 3.00 (1.04) | 0.91 | 0.92 | 7 | 0.87 | 0.75 | | 10* | 0.89 | 0.79 | | 16* | 0.81 | 0.65 | | 24 | 0.67 | 0.44 | | 32* | 0.92 | 0.84 | | 33 | 0.84 | 0.71 | | CEBQ Child Eating Behaviour Questionnaire, SR Satiety Responsiveness, SE Slowness in Eating, FF Food Fussiness, EUE Emotional Undereating, FR Food Responsiveness,EF Enjoyment of Food, DD Desire to Drink, EOE Emotional OvereatingAll values are significant at p < .001* Reversed items were previously inversed | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | | Table 6 Crossloadings of the modified 8-factor model | | - | | Item | Original subscale(Factor loading) | New subscale(Factor loading) | | 24 | FF (0.44) | SR (0.44) | | 23 | EUE (0.63) | EF (0.27) | | 35 | SE (0.52) | EUE (0.27) | | 5 | EF (0.47) | FF (-0.26) | | 3* | SR (1.05) | EUE (-0.41) | | 33 | FF (0.75) | SR (0.18) | | SR Satiety Responsiveness, SE Slowness in Eating, FF Food Fussiness, EUEEmotional Undereating* Reversed Items were previously inversed | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | ### Page 11 between ARFID and AN. The only exception was the FFsubscale, which significantly distinguished ARFID fromall other groups, including AN, with the highest FF valuesobserved in the ARFID group. Differences in CEBQ subscale mean scores by sex and agegroup No significant differences were found in the CEBQ sub-scales by sex, F(16, 226) = 2.44, p = .18. Effect sizes for allsubscales were small (ηp² ≤ 0.01), indicating negligibledifferences between boys and girls. Age groups differedsignificantly in CEBQ subscales, F(16, 226) = 2.44, p =.002. Univariate analyses showed significant group differ-ences for all food avoidance subscales (p < .05) with smallto medium effect (Suppl. Table S3), but no significantdifferences for food approach subscales. Food avoidancescores were significantly higher in the 0–7 than 8–13years age group for SR, SE, and EUE (medium effects),and compared to the 14–17 years age group for FF andSR (large effects). Discussion This is the first study examining the validity and reliabilityof the CEBQ in children and adolescents with interview-assessed EDs such as AN and ARFID as well as LOCeating and ADHD. Overall, the CEBQ showed accept-able validity and reliability. Compared to two previouslyproposed 7-factor models [e.g.,,], the original 8-fac-tor structure [] exhibited the best fit for this sample,with acceptable to high factor loadings across all items.Internal consistency for all subscales was acceptable toexcellent. As expected, convergent validity, as assessedthrough objectively derived BMI-SDS and well-estab-lished questionnaires, was acceptable across all subscalesexcept for the DD subscale. The CEBQ subscales, exceptfor DD and EUE subscales, significantly differentiatedbetween diagnostic groups associated with overeating(ADHD and LOC) and those characterized by restrictiveeating (ARFID and AN), as hypothesized. No significantdifferences in CEBQ subscale mean scores were observedfor child sex, whereas food avoidance behavior variedacross age groups. Missing item responses occurred rarely suggestinggood comprehensibility and applicability of the CEBQto the diverse sample including eating disturbances andADHD. Correlations between each item and its subscalewere sufficient for all items, as previously reported in asample of Dutch primary school children []. Items ofthe subscales EOE and FR showed increased difficulty,reflected in low mean scores and positive skewness, aspreviously reported for the EOE subscale [e.g.,,].Particularly younger children may tend to undereatrather than overeat in emotional situations [,]. | Table 7 Spearman rank correlation coefficients of CEBQ subscales for the original 8-factor model | | - | | Subscale | FR | EF | EOE | DD | SR | SE | EUE | FF | | FR | 1 | | EF | 0.54** | 1 | | EOE | 0.57** | 0.38** | 1 | | DD | 0.34** | 0.11 | 0.33** | 1 | | SR | − 0.37** | − 0.66** | − 0.25** | 0.01 | 1 | | SE | − 0.31** | − 0.42** | − 0.18** | 0.08 | 0.50** | 1 | | EUE | 0.03 | − 0.10 | 0.26** | 0.15* | 0.32** | 0.26** | 1 | | FF | − 0.11 | − 0.56** | − 0.08 | − 0.02 | 0.45** | 0.23** | 0.24** | 1 | | CEBQ Child Eating Behaviour Questionnaire, SR Satiety Responsiveness, SE Slowness in Eating, FF Food Fussiness, EUE Emotional Undereating, FR Food Responsiveness,EF Enjoyment of Food, DD Desire to Drink, EOE Emotional Overeating* p < .05, ** p < .01 | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | | Table 8 Measures for convergent validity of the CEBQ original 8-factor model: Spearman rank correlation coefficients | | - | | Measure | n | FR | EF | EOE | DD | SR | SE | EUE | FF | | BMI-SDS | 226 | 0.37** | 0.43** | 0.21** | − 0.03 | −0.48** | − 0.36** | − 0.19** | − 0.19** | | Days with avoidant/restrictive food intake, parent-report | 166 | − 0.13 | − 0.42** | − 0.06 | − 0.03 | 0.30** | 0.11 | 0.24** | 0.43** | | Days with avoidant/restrictive food intake, self-report | 127 | 0.10 | − 0.28** | 0.02 | 0.02 | 0.29** | 0.06 | 0.24* | 0.40** | | OBEs | 158 | 0.27** | 0.24** | 0.12 | 0.11 | −0.25** | − 0.21** | − 0.06 | − 0.06 | | SBEs | 158 | 0.04 | 0.07 | 0.13 | 0.07 | 0.04 | 0.05 | 0.08 | 0.04 | | EDY-Q total mean score | 128 | − 0.24** | − 0.60** | − 0.27** | − 0.03 | 0.57** | 0.38** | 0.10 | 0.48** | | BMI-SDS body mass index standard deviation score, CEBQ Child Eating Behaviour Questionnaire, EDY-Q Eating Disturbances in Youth-Questionnaire, SR SatietyResponsiveness, SE Slowness in Eating, FF Food Fussiness, EUE Emotional Undereating, FR Food Responsiveness, EF Enjoyment of Food, DD Desire to Drink, EOEEmotional Overeating, OBEs Objective binge-eating episodes, SBEs Subjective binge-eating episodes* p < .05 (two-tailed)** p < .01 (two-tailed) | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | ### Page 12 Comparative factor analyses revealed the superiority ofthe theory-based original 8-factor model over two pre-viously identified 7-factor structures, combining SE andSR factors [,], or combining EOE and FR factors [,], as previously demonstrated in nonclinical samples[,]. Regarding the CEBQ’s factorial validity, previousstudies in nonclinical samples showed comparable resultswith poor to acceptable model fit of the original 8-factormodel [e.g.,], acceptable to high factor loadings [e.g.,,], and acceptable to excellent internal consistency forall subscales [e.g.,,]. The similarity of the CFA resultsbetween the clinical sample of the current study and non-clinical samples suggests that the CEBQ is universallyapplicable to both healthy and treatment-seeking indi-viduals. As demonstrated previously, the CEBQ’s modelfit was improved using modification indices [e.g.,,].Some of the applied crossloadings in the present study(see Table) have already been reported in other studies,for example for Item 24 [] and Item 5 []. Althoughthe modified 8-factor model showed an improved modelfit to our data with up to good fit indices, analyses forconvergent and discriminant validity were conductedusing the unmodified 8-factor model. Given that all mod-ified items had equal or higher factor loadings on theiroriginal subscales and the factor loadings on the newsubscales were low, the unmodified model was retainedto avoid randomness and possible overfit to the data. Aligning meta-analytic findings [], CEBQ subscalescorrelated significantly positively (for food approach sub-scales) or negatively (for food avoidance subscales) withchildren’s BMI-SDS. Only the DD subscale did not cor-relate with children`s BMI-SDS, as previously indicated[e.g.,,]. Additionally, this study used interview- andquestionnaire-based measures on children’s EDs andassociated psychopathology for evaluating convergentvalidity of the CEBQ. As expected, the self-reportednumber of OBEs assessed by interview was significantlypositively correlated with the subscales FR and EF andnegatively with the subscales SR and SE, indicating anassociation between binge eating and general overeat-ing behavior in children and adolescents, consistent withlaboratory-based test meal studies in children with LOCeating [e.g.,,]. The correlation between the numberof self-reported OBEs and the EOE subscale was nonsig-nificant, although negative affect has been proposed asan antecedent of overeating and LOC eating in childrenand adolescents []; however, test meal and ecologi-cal momentary assessment studies have not consistentlysupported this association [,]. In contrast to OBEs,the number of SBEs did not correlate significantly withany of the CEBQ subscales, as the CEBQ is based onparent-report and parents might have difficulties noticingSBEs in their children. | Table 9 Differences in original CEBQ subscales by diagnostic groups, controlled for sex, age, and BMI-SDS | | - | | CEBQ subscale | Controln = 124 | ARFIDn = 39 | ANn = 24 | LOCn = 14 | ADHDn = 24 | ANCOVAs | | M | SD | M | SD | M | SD | M | SD | M | SD | F(4, 217) | p | ηp² | | FR | 1.94a | 0.76 | 1.61a | 0.57 | 1.30a | 0.33 | 2.90b | 0.98 | 2.45b | 0.81 | 7.63 | < 0.001 | 0.12 | | EF | 3.55a | 0.79 | 2.44b | 0.96 | 2.71b | 0.96 | 4.25a | 0.55 | 3.83a | 0.65 | 12.75 | < 0.001 | 0.19 | | EOE | 1.47a, b | 0.61 | 1.34a, c | 0.38 | 1.21a | 0.34 | 1.96b, c | 0.80 | 1.89c | 0.73 | 4.80 | < 0.001 | 0.08 | | DD | 2.27 | 0.87 | 2.52 | 1.11 | 2.21 | 1.18 | 2.43 | 0.88 | 2.51 | 0.93 | 0.57 | 0.69 | 0.01 | | SR | 2.77a, b | 0.76 | 3.59c | 0.62 | 3.24b, c | 0.81 | 2.00a | 0.43 | 2.58a, b | 0.80 | 7.64 | < 0.001 | 0.12 | | SE | 2.67a, b | 0.82 | 3.17a | 0.88 | 3.17a | 0.89 | 1.95b | 0.54 | 2.21b | 0.83 | 3.44 | < 0.001 | 0.09 | | EUE | 2.61 | 0.97 | 3.28 | 1.01 | 2.48 | 0.96 | 2.54 | 0.72 | 2.90 | 0.94 | 1.65 | 0.16 | 0.03 | | FF | 2.87a | 1.02 | 3.91b | 0.78 | 2.76a | 0.86 | 2.49a | 0.77 | 2.81a | 1.11 | 6.68 | < 0.001 | 0.11 | | CEBQ Child Eating Behaviour Questionnaire, SR Satiety Responsiveness, SE Slowness in Eating, FF Food Fussiness, EUE Emotional Undereating, FR Food Responsiveness, EF Enjoyment of Food, DD Desire to Drink, EOEEmotional Overeatinga, b, c Identical letters indicate nonsignificant differences between the respective groups. p < .05, adjustments according to Bonferroni-correction | | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | - | ### Page 13 The parent-reported number of days with avoidant-restrictive food intake showed significantly medium-sized positive and negative correlations with the SR, FF,and EF subscales, corresponding to avoidant-restrictivefood intake being defined by restrictions in the quantity(measured by SR) or variety (measured by FF) of foodintake [], and reflecting clinical observations that chil-dren with avoidant-restrictive food intake hardly enjoyeating (measured by EF). Overall, all significant correla-tions between the CEBQ subscales and the measures forconvergent validity corresponded to the hypothesizeddirections, which supports convergent validity. Oneexception is the DD subscale, which showed no signifi-cant associations with measures for convergent validityand inconsistent intercorrelations with other CEBQ sub-scales, in line with previous research [e.g.,,]. Regarding discriminant validity, significant differencesby diagnostic group were seen for all CEBQ subscales,except for the subscales DD and EUE, indicating thatthese latter subscales may not be suitable for discriminat-ing different eating disturbances, ADHD, and controls.Despite adequate internal consistency and factor load-ings for the DD subscale in the present sample, its weaksubscale intercorrelations as well as low convergent anddiscriminant validity across different samples [e.g.,,]suggest a psychometric limitation of the subscale. Mean values of the food approach subscales were par-ticularly low for AN and ARFID groups and high for LOCand ADHD groups, and for food avoidance subscales viceversa, while the control group showed medium valuesfor all subscales. These results were expected given thatrestrictive eating behavior is a diagnostic criterion forARFID and AN, while overeating is diagnostically rele-vant for BED [] and common in children with ADHD [,]. The results thus show that the CEBQ can distinguishbetween diagnostic groups associated with overeating(ADHD and LOC) and those involving restrictive eating(ARFID and AN) but reveals only limited discrimina-tion between groups within each of these categories. Ofnote, the FF subscale significantly differentiated ARFIDfrom all other diagnostic groups, including AN, which isconsistent with the understanding that selective eatingbehavior may be a main characteristic feature of ARFID[]. Given small subgroup sizes, particularly in the over-weight, obesity, and LOC eating groups (n < 20), findingsregarding diagnostic differentiation should be interpretedwith caution. In terms of sociodemographic correlates, the presentfindings align with previous studies [e.g.,] on absentsex differences, while general evidence base is mixed[e.g.,,]. Regarding children’s age, only food avoidancesubscales significantly differed by age group, with highest ### Page 14 scores for children aged 0–7 years, which is consistentwith findings that the peak of picky eating and food neo-phobia is usually seen between ages 2–6 years [], anddecreasing food avoidance and constant food approachbehavior from toddlers to young teenagers in CEBQstudies covering a wide age range [,]. Regarding strengths and limitations, one of thestrengths of this study is the sample of children andadolescents with interview-assessed AN, ARFID, LOCeating, ADHD, and a control group of healthy children.Based on a literature review, a comparative CFA con-taining the most frequently proposed factor models wascarried out. Additionally to the objectively measuredanthropometrics, well-established and validated ques-tionnaires on eating behavior were used to determine theconvergent validity of the CEBQ subscales. The inclusionof both treatment-seeking and non-treatment-seekingparticipants enhances the generalizability of findingsto diverse populations but may also introduce variabil-ity that complicates the interpretation of results. Thewide age range of 9 months to 17 years may also be astrength and limitation at the same time. Although thereis no official age limit for the CEBQ, it was developed ina sample including children aged 2–9 years [] and hasso far mainly been validated for the age range from 12months – 13 years [e.g.,,], with exception of onerecent validation study in Spanish children and adoles-cents aged 6–16 years []. Its application in infants andadolescents remained exploratory, and some items maybe interpreted differently across developmental stages,potentially introducing age-related measurement bias.Another limitation of this study pertains to the samplesize, particularly in relation to the limited number of par-ticipants in the diagnostic, weight, and age groups ana-lyzed by MANCOVAs. Finally, the results of this studyexhibit limited generalizability on populations fromnon-Western countries and populations of lower educa-tional levels, as the parents surveyed in this study showedabove-average levels of education [,]. Conclusion In sum, the findings provide support for the German ver-sion of the CEBQ as a valid and reliable tool for assess-ing eating behavior in children and adolescents acrossvarious clinical groups, including specific EDs, eatingdisturbances and ADHD, and can thus be used in futureresearch and clinical settings. Given the broad devel-opmental range of the sample (0–17 years), the presentpsychometric findings cannot be generalized uniformlyacross ages and should primarily be interpreted atthe overall sample level. Future research should con-duct measurement invariance analyses in large clinicalsamples to ensure that the results are equally transfer-able to all age and diagnostic groups. The provision ofpopulation-based and clinical normative scores for theCEBQ mean scores is crucial for evaluating its potentialutility as a screening tool for identifying atypical eatingbehavior indicative of overeating vs. restrictive EDs. Abbreviations ADHD Attention-deficit/hyperactivity disorder AIC Akaike information criterion AN Anorexia nervosa ANCOVA One-way analysis of covariance ARFID Avoidant-restrictive food intake disorder BED Binge-eating disorder BMI Body mass index BMI-SDS BMI standard deviation score BN Bulimia nervosa CEBQ Child eating behaviour questionnaire CFA Confirmatory factor analysis CFI Comparative fit index DD Desire to drink DSM-IV-TR Diagnostic and statistical manual of mental disorders, fourthedition, text revision DSM-5 Diagnostic and statistical manual of mental disorders, 5thedition EDE Eating disorder examination EDs Eating disorders EDY-Q Eating disturbances in youth-questionnaire EF Enjoyment of food EFA Exploratory factor analysis EOE Emotional overeating EUE Emotional undereating FF Food fussiness FR Food responsiveness ICD-11 International classification of diseases 11th revision LIFE Leipzig research center for civilization diseases LOC Loss of control MANCOVA Multivariate analysis of covariance MCAR Missing completely at random NFI Normed fit index OBEs Objective binge-eating episodes RMSEA Root mean square error of approximation SBEs Subjective binge-eating episodes SE Slowness in eating SR Satiety responsiveness SRMR Standardized root mean squared residual SUN Swiss university study of nutrition TLI Tucker-Lewis-index Supplementary Information The online version contains supplementary material available at. Supplementary Material 1. Acknowledgements The authors are grateful to all parents and their children for participation. Author contributions Conceptualization: JU, AH, RS; Data curation: JU, RS; Formal analysis: JU;Funding acquisition: AH, SM; Investigation: AH, RS; Methodology: JU, AH,RS; Project administration: AH, RS; Resources: AH, SM; Supervision: AH, RS;Visualization: JU; Writing - original draft: JU; Writing – review & editing: allauthors. Funding Open Access funding enabled and organized by Projekt DEAL. This publicationwas supported by the German Federal Ministry of Education and Research(grant 01EO1501, 01EO1001) and the Swiss National Science Foundation(SNSF) (grant 100014132045/1). ### Page 15 Data availability The data are available from the author upon reasonable request. Declarations Ethics approval and consent to participate The studies were approved by the local Ethics Committees (Leipzig, Germany/Fribourg, Switzerland), and informed consent and assent was provided byparents and children ≥ 8 years old. Consent for publication Not applicable. Competing interests The authors declare no competing interests. Author details 1Department of Psychosomatic Medicine and Psychotherapy, BehavioralMedicine Research Unit, Integrated Research and Treatment CenterAdiposityDiseases, Leipzig University Medical Center, Philipp-Rosenthal-Strasse 55, 04103 Leipzig, Germany 2German Center for Child and Adolescent Health (DZKJ), partner siteLeipzig/Dresden, Leipzig, Germany 3Department of Psychology, Clinical Psychology and Psychotherapy,University of Fribourg, Fribourg, Switzerland 4Food Research and Innovation Center, Cluster Food and Mental Health/Psychology, University of Fribourg, Fribourg, Switzerland Received: 29 October 2025 / Accepted: 23 March 2026 References 1. American Psychiatric Association. Diagnostic and statistical manual of mentaldisorders. 5th ed. Washington, DC: American Psychiatric Publishing; 2013. 2. World Health Organization. International classification of diseases for mortal-ity and morbidity statistics (11th revision). 2019. 3. Shomaker LB, Tanofsky-Kraff M, Elliott C, Wolkoff LE, Columbo KM, Ran-zenhofer LM, Roza CA, Yanovski SZ, Yanovski JA. Salience of loss of controlfor pediatric binge episodes: does size really matter? Int J Eat Disord.2010;43(8):707–16.. 4. Villa FM, Crippa A, Rosi E, Nobile M, Brambilla P, Delvecchio G. ADHD and eat-ing disorders in childhood and adolescence: an updated minireview. J AffectDisord. 2023;321:265–71.. 5. Kaisari P, Dourish CT, Higgs S. Attention deficit hyperactivity disorder (ADHD)and disordered eating behaviour: A systematic review and a framework forfuture research. Clin Psychol Rev. 2017;53:109–21.. 6. Braet C, Soetens B, Moens E, Mels S, Goossens L, Van Vlierberghe L. Are twoinformants better than one? Parent–child agreement on the eating styles ofchildren who are overweight. Eur Eat Disord Rev. 2007;15(6):410–7.. 7. Wardle J, Guthrie CA, Sanderson S, Rapoport L. Development of the Children’sEating Behaviour Questionnaire. J Child Psychol Psychiatry. 2001;42(7):963–70.. 8. Domoff SE, Miller AL, Kaciroti N, Lumeng JC. Validation of the Children’s Eat-ing Behaviour Questionnaire in a low-income preschool-aged sample in theUnited States. Appetite. 2015;95:415–20.. 9. Njardvik U, Klar EK, Thorsdottir F. The factor structure of the Children’s EatingBehaviour Questionnaire: a comparison of four models using confirmatoryfactor analysis. Health Sci Rep. 2018;1(3):28.. 10. Santos JL, Ho-Urriola JA, González A, Smalley SV, Domínguez-Vásquez P,Cataldo R, et al. Association between eating behavior scores and obesity inChilean children. Nutr J. 2011;10:108.. 11. Somaraki M, Ek A, Sandvik P, Byrne R, Nowicka P. How do young children eatafter an obesity intervention? Validation of the Child Eating Behaviour Ques-tionnaire using the Rasch Model in diverse samples from Australia and Swe-den. Appetite. 2022;169:105822.. 12. Viana V, Sinde S, Saxton J. Children’s Eating Behaviour Questionnaire: Associa-tions with BMI in Portuguese children. Br J Nutr. 2008;100(2):445–50.. 13. Kozak A, Czepczor-Bernat K, Modrzejewska J, Modrzejewska A, Matusik E,Matusik P. Avoidant/restrictive food disorder (ARFID), food neophobia, othereating-related behaviours and feeding practices among children with autismspectrum disorder and in non-clinical sample: A preliminary study. Int JEnviron Res Public Health. 2023;20(10):5822.. 14. Schmidt R, Hiemisch A, Kiess W, von Klitzing K, Schlensog-Schuster F,Hilbert A. Validation study on the child, adult, and parent version of theARFID module 2.0 for the Eating Disorder Examination. Int J Eat Disord.2022;55(12):1708–20.. 15. Leventakou V, Herle M, Kampouri M, Margetaki K, Vafeiadi M, Kogevinas M, etal. The longitudinal association of eating behaviour and ADHD symptoms inschool age children: A follow-up study in the RHEA cohort. Eur Child AdolescPsychiatry. 2022;31(3):511–7.. 16. Wu D, Chen S, Lin S, Huang R, Li R, Huang Y, Chen M, Li X. Eating behavior-mediated association between attention deficit hyperactivity disorder andbody fat mass. Heliyon. 2023;9(3):e13641.. 17. Ko MM, Lee SH, Chang GT, Lee B. Reliability and validity of a Korean version ofthe Children’s Eating Behavior Questionnaire in anorexia context. Front Nutr.2023;10:1247630.. 18. Moosbrugger H, Kelava A. Testtheorie und Fragebogenkonstruktion. Springer;2007. 19. Ek A, Sorjonen K, Eli K, Lindberg L, Nyman J, Marcus C, et al. Associationsbetween parental concerns about preschoolers’ weight and eating andparental feeding practices: Results from analyses of the Child Eating BehaviorQuestionnaire, the Child Feeding Questionnaire, and the Lifestyle BehaviorChecklist. PLoS ONE. 2016;11(1):e0147257.. 20. Ruden JC, Warschburger P. Validation of a 24-Item Version of the Child EatingBehavior Questionnaire (CEBQ-24G) in a Sample of German Parents withPrimary School Children. Appetite. 2025;214:108223.. 21. Loh DA, Moy FM, Zaharan NL, Mohamed Z. Eating behaviour amongmulti-ethnic adolescents in a middle-income country as measured bythe self-reported Children’s Eating Behaviour Questionnaire. PLoS ONE.2013;8(12):e82885.. 22. Sparks MA, Radnitz CL. Confirmatory factor analysis of the Children’s EatingBehaviour Questionnaire in a low-income sample. Eat Behav. 2012;13(3):267–70.. 23. Sleddens EF, Kremers SP, Thijs C. The Children’s Eating Behaviour Question-naire: Factorial validity and association with Body Mass Index in Dutchchildren aged 6–7. Int J Behav Nutr Phys Act. 2008;5:49.. 24. Svensson V, Lundborg L, Cao Y, Nowicka P, Marcus C, Sobko T. Obesity relatedeating behaviour patterns in Swedish preschool children and associationwith age, gender, relative weight and parental weight - Factorial validationof the Children’s Eating Behaviour Questionnaire. Int J Behav Nutr Phys Act.2011;8:134.. 25. Purwaningrum DN, Arcot J, Hadi H, Hasnawati RA, Rahmita RS, Jayasuriya R.A cultural adaptation and validation of a child eating behaviour measure ina low- and middle-income country. Public Health Nutr. 2020;23(11):1931–8.. 26. Gao M, Xue K, Guo H, Suppl. S82–6. doi:. 27. Kininmonth A, Smith A, Carnell S, Steinsbekk S, Fildes A, Llewellyn C. Theassociation between childhood adiposity and appetite assessed using theChild Eating Behavior Questionnaire and Baby Eating Behavior Questionnaire:A systematic review and meta-analysis. Obes Rev. 2021;22(5):e13169.. 28. Kurz S, van Dyck Z, Dremmel D, Munsch S, Hilbert A. Early-onset restrictiveeating disturbances in primary school boys and girls. Eur Child AdolescPsychiatry. 2015;24(7):779–85.. 29. Poulain T, Baber R, Vogel M, Pietzner D, Kirsten T, Jurkutat A, et al. The LIFEChild study: A population-based perinatal and pediatric cohort in Germany.Eur J Epidemiol. 2017;32(2):145–58.. 30. Fairburn CG, Cooper Z, O’Connor ME. Eating Disorder Examination (Edition17.0D). 2021.. Accessed23 Sept 2021. ### Page 16 31. Hilbert A, Tuschen-Caffier B. Eating Disorder Examination. German translation.dgvt-; 2016. 32. Bryant-Waugh RJ, Cooper PJ, Taylor CL, Lask BD. The use of the eating disorderexamination with children: A pilot study. Int J Eat Disord. 1996;19(4):391–7./ ( S I C I ) 1 0 9 8 - 1 0 8 X ( 1 9 9 6 0 5 ) 1 9 : 4 % 3 C 3 9 1 : : A I D- E AT 6 % 3 E 3 . 0 .CO ; 2 -G . 33. Hilbert A. Eating Disorder Examination for children. German translation. dgvt-;2016. 34. Puig-Antich J, Chambers W. Schedule for Affective Disorders and Schizo-phrenia for school-aged children. Unpublished interview schedule. New YorkState Psychiatric Institute; 1978. 35. Delmo C, Weiffenbach O, Gabriel M, Poustka F. 3. Auflage der deutschenForschungsversion des K-SADS-PL, erweitert um ICE-10-Diagnostik. Bern:Huber; 2000. 36. American Psychiatric Association. Diagnostic and statistical manual of mentaldisorders. 4th ed., text rev. Washington, DC: American Psychiatric Association;2000. 37. Kromeyer-Hauschild K, Wabitsch M, Kunze D, Geller F, Geiß HC, Hesse V, et al.Perzentile für den Body-Mass-Index für das Kindes- und Jugendalter unterHeranziehung verschiedener deutscher Stichproben. Monatsschr Kinderhe-ilkd. 2001;149:807–18.. 38. van Dyck Z, Bellwald L, Kurz S, Dremmel D, Munsch S, Hilbert A. Essproblemeim Kindesalter. Z Gesundheitspsychol. 2013;21(2):91–100.. 39. van Dyck Z, Hilbert A. Eating Disorders in Youth-Questionnaire. DeutscheVersion. Universität Leipzig; 2016. 40. Lei M, Lomax RG. The effect of varying degrees of nonnormality in structuralequation modeling. Struct Equ Model. 2005;12(1):1–27.. 41. Nunnally JC, Bernstein IH. Psychometric theory. McGraw-Hill; 1994. 42. Bentler PM, Wu EJ. EQS for Windows user’s guide [Version 5]. Encino: Multi-variate Software Inc; 1995. 43. Hu LT, Bentler PM. Cutoff criteria for fit indexes in covariance structureanalysis: Conventional criteria versus new alternatives. Struct Equ Model.1999;6(1):1–55.. 44. Schermelleh-Engel K, Moosbrugger H, Müller H. Evaluating the fit of struc-tural equation models: Tests of significance and descriptive goodness-of-fitmeasures. Methods Psychol Res Online. 2003;8(2):23–74. 45. Bozdogan H. Akaike’s information criterion and recent developments ininformation complexity. J Math Psychol. 2000;44(1):62–91.. 46. Stevens JP. Applied multivariate statistics for the social sciences. Routledge;2009. 47. Weiber R, Mühlhaus D, Strukturgleichungsmodellierung. Springer; 2014. 48. Cohen J. Quantitative methods in psychology: A power primer. Psychol Bull.1992;112(1):155–9.. 49. DeVellis RF, Thorpe CT. Scale development: Theory and applications. Sage;2022. 50. Cohen J. Statistical power analysis for the behavioral sciences. 2nd ed.Erlbaum; 1988. 51. Jackson DL, Gillaspy JA, Purc-Stephenson R. Reporting practices in confirma-tory factor analysis: an overview and some recommendations. PsycholMethods. 2009;14(1):6–23.. 52. MacCallum RC, Roznowski M, Necowitz LB. Model modifications in covari-ance structure analysis: the problem of capitalization on chance. Psychol Bull.1992;111(3):490–504.. 53. Gebru N, Gebreyesus S, Hassen H, Habtemariam E, Abebe D. Cultural adapta-tion and validation of the Children’s Eating Behaviour Questionnaire inEthiopia. J Nutr Sci. 2021;10:52.. 54. Mallan KM, Liu WH, Mehta RJ, Daniels LA, Magarey A, Battistutta D. Maternalreport of young children’s eating styles. Validation of the Children’s EatingBehaviour Questionnaire in three ethnically diverse Australian samples. Appe-tite. 2013;64:48–55.. 55. Jimeno-Martínez A, Maneschy I, Moreno L, Bueno-Lozano G, Miguel-EtayoP, de Flores-Rojas K, et al. Reliability and validation of the Child EatingBehavior Questionnaire in 3- to 6-year-old Spanish children. Front Psychol.2022;13:705912.. 56. Sweetman C, Wardle J, Cooke L. Soft drinks and ‘desire to drink’ in preschool-ers. Int J Behav Nutr Phys Act. 2008;5:60.. 57. Hilbert A, Tuschen-Caffier B, Czaja J. Eating behavior and familial interactionsof children with loss of control eating: A laboratory test meal study. Am J ClinNutr. 2010;91(3):510–8.. 58. Tanofsky-Kraff M, McDuffie JR, Yanovski SZ, Kozlosky M, Schvey NA, ShomakerLB, et al. Laboratory assessment of the food intake of children and adoles-cents with loss of control eating. Am J Clin Nutr. 2009;89(3):738–45.. 59. Byrne ME, LeMay-Russell S, Tanofsky-Kraff M. Loss-of-control eating and obe-sity among children and adolescents. Curr Obes Rep. 2019;8(1):33–42.. 60. Hilbert A, Rief W, Tuschen-Caffier B, de Zwaan M, Czaja J. Loss of controleating and psychological maintenance in children: an ecological momentaryassessment study. Behav Res Ther. 2009;47(1):26–33.. 61. Levin RL, Rawana JS. Attention-deficit/hyperactivity disorder and eating dis-orders across the lifespan: A systematic review of the literature. Clin PsycholRev. 2016;50:22–36.. 62. Dovey TM, Staples PA, Gibson EL, Halford JC. Food neophobia and ‘picky/fussy’ eating in children: A review. Appetite. 2008;50(2–3):181–93.. 63. Sanlier N, Arslan S, Buyukgenc N, Toka O. Are eating behaviors related with bybody mass index, gender and age? Ecol Food Nutr. 2018;57(4):372–87.. 64. Cao YT, Svensson V, Marcus C, Zhang J, Zhang JD, Sobko T. Eating behaviourpatterns in Chinese children aged 12–18 months and association with rela-tive weight - Factorial validation of the Children’s Eating Behaviour Question-naire. Int J Behav Nutr Phys Act. 2012;9:5.. 65. García AC, Pedrero-Tomé R, Alaminos-Torres A, López-Ejeda N, Sánchez Álva-rez M, Martínez CP, et al. Evaluation of the reliability and internal consistencyof the child eating behavior questionnaire (CEBQ) in Spanish schoolchildrenand its association with obesity. Brain Behav. 2025;15(2):e70343.. 66. Bundesamt für Statistik. Anteil der Bevölkerung mit Hochschulabschluss.Atlas der Schweiz. 2025.. Accessed 16 Jan 2025. 67. Statistisches Bundesamt. Educational attainment of the population in Ger-many. Statistisches Bundesamt. 2020.. Accessed 17 Jan 2025. 68. Sirirassamee T, Hunchangsith P. Children’s Eating Behavior Questionnaire: Fac-torial validation and differences in sex and educational level in Thai school-age children. Southeast Asian J Trop Med Public Health. 2016;47(6):1325–34. 69. Quah PL, Cheung YB, Pang WW, Toh JY, Saw SM, Godfrey KM, et al. Valida-tion of the Children’s Eating Behavior Questionnaire in 3 year old childrenof a multi-ethnic Asian population: The GUSTO cohort study. Appetite.2017;113:100–5.. 70. Quah PL, Fries LR, Chan MJ, Fogel A, McCrickerd K, Goh AT, et al. Validation ofthe Children’s Eating Behavior Questionnaire in 5 and 6 year-old children: TheGUSTO Cohort Study. Front Psychol. 2019;10:824.. 71. Spahić R, Pranjić N. Children’s Eating Behaviour Questionnaire: Associationwith BMI in children aged 3–10 years from Bosnia and Herzegovina. PublicHealth Nutr. 2019;22(18):3360–7.. 72. Oyama S, Arslanian KJ, Fidow UT, Naseri T, Soti-Ulberg C, Hawley NL. Factorialvalidation analysis of the Baby and Children’s Eating Behavior Questionnairesin Samoa. Eat Behav. 2021;42:101530.. 73. Al-Hamad A, Al-Naseeb A, Al-Assaf M, Al-Obaid S, Al-Abdulkarim B, OlszewskiP. Preliminary exploration of obesity-related eating behaviour patterns in asample of Saudi preschoolers aged 2–6 years through the Children’s EatingBehaviour Questionnaire. Nutrients. 2021;13(11):4156.. 74. Ayre S, Gallegos D, Nambiar S, Tran C, Do D, Jansen E. Preliminary explorationof the use of the Children’s Eating Behaviour Questionnaire (CEBQ) and Feed-ing Practices and Structure Questionnaire (FPSQ) in Vietnamese mothers. EurJ Clin Nutr. 2022;76(3):442–9.. 75. Manzano MA, Strong DR, Kang Sim DE, Rhee KE, Boutelle KN. Psychometricproperties of the Child Eating Behavior Questionnaire (CEBQ) in school agechildren with overweight and obesity: A proposed three-factor structure.Pediatr Obes. 2021;16(10):e12795.. ### Page 17 76. Zhou N, Sun L. Confirmatory factor analysis of the Children’s Eating BehaviourQuestionnaire in a Chinese urban preschooler sample. Int J Behav Nutr PhysAct. 2021;18:113.. 77. Hunot-Alexander C, Curiel-Curiel CP, Romero-Velarde E, Vásquez-GaribayEM, Mariscal-Rizo A, Casillas-Toral E, et al. Intergenerational transmission ofappetite: Associations between mother-child dyads in a Mexican population.PLoS ONE. 2022;17:0264493.. 78. Malczyk Ż, Kuczka O, Pasztak-Opiłka A, Zachurzok A. Validation of the Chil-dren’s Eating Behaviour Questionnaire in Poland. Nutrients. 2022;14(22):4782.. 79. Mou J, Zhou H, Huang S, Feng Z. Factorial validation of the children’s eatingbehavior questionnaire and the association between the eating behaviorsassessed and BMI Z score in Chinese preschoolers aged 2–5 years. J HealthPopul Nutr. 2024;43:145.. 80. Peuckert MZ, Ayala CO, Mattiello R, Viola TW, Walker MS, Feoli AMP, et al.Validation evidence for the Children’s Eating Behaviour Questionnaire (CEBQ)in Brazil: A cross-sectional study. Nutrients. 2025;17(5):851.. 81. Nigam S, Zhu M, Shoemaker AH, Gross AC, Fox CK, Thaker VV. Novel model ofchildhood appetitive traits in children with obesity. Int J Obes. 2025;49:1084–91.. Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims inpublished maps and institutional affiliations.