New Turkish Eating Questionnaire Passes Its Toughest Scientific Test

Why do some people reach for chocolate when they are stressed, while others eat only when hunger strikes? Scientists have long tried to measure these patterns with short questionnaires, and one of the most practical tools is the Weight-Related Eating Questionnaire, or WREQ. Originally developed for English-speaking populations, the WREQ asks just a handful of questions yet captures four distinct behaviors: routine restraint, compensatory restraint, susceptibility to external food cues, and emotional eating. Now, a team of Turkish researchers has put the instrument through a full psychometric workout, translating it into Turkish and testing whether it still measures what it claims to measure in a completely different language and culture. Their verdict, published in BMC Psychology, is a resounding yes, with a few instructive wrinkles along the way.

The study, led by Elif Çelik of Süleyman Demirel University together with Emine Kocyigit of Ordu University and Özge Cemali of Trakya University, recruited 301 young adults between the ages of 19 and 35 from three regions of Türkiye. The sample was predominantly female, at 85 percent, with a mean age of 20.65 years and a mean body mass index of 22.33 kilograms per square meter, which falls squarely within the normal weight range. That profile matters, because the researchers wanted a tool suited to young adults, a group in which disordered eating patterns often first emerge and in which early identification can change long-term health trajectories. The work was approved by the Ordu University Social Sciences and Humanities Ethics Committee and conducted according to the principles of the Helsinki Declaration.

Translating a psychological questionnaire is far more subtle than running words through a dictionary. Idioms about food, guilt, and self-control do not map neatly across languages, and a poorly adapted item can silently distort responses. The team therefore followed the guidelines of the International Test Commission, the standard framework for cross-cultural adaptation of tests. This process typically involves forward translation, back translation, expert review, and pilot testing, with content validity quantified through item-level and scale-level content validity indices, measures that report how thoroughly each question covers its intended concept. The researchers report that these content validity checks were completed before any statistical testing began, laying the groundwork for the analyses that followed.

The first major statistical hurdle was exploratory factor analysis, a technique that asks whether the pattern of answers naturally clusters into the expected dimensions. The data passed the preliminary tests with flying colors: the Kaiser-Meyer-Olkin measure of sampling adequacy reached 0.88, well above the conventional 0.60 threshold, and Bartlett’s test of sphericity was highly significant at 2424.6 with a p-value below 0.001, confirming that the correlation matrix contained genuine structure worth extracting. Interestingly, the exploratory analysis suggested a three-factor solution explaining 62.15 percent of the total variance, with Emotional Eating accounting for 33.96 percent, Dietary Restraint for 19.32 percent, and Susceptibility to External Cues for 8.87 percent. Communalities, which indicate how much of each item’s variance the factors explain, ranged from 0.335 to 0.814.

Here the story takes a technically interesting turn. The original WREQ distinguishes two kinds of restraint: routine restraint, the habitual effort to limit intake, and compensatory restraint, the deliberate cutting back after overeating or in anticipation of a large meal. The exploratory analysis in the Turkish sample merged these into a single Dietary Restraint factor. But when the researchers ran confirmatory factor analysis, which tests pre-specified models against the data rather than discovering structure from scratch, the original four-factor model fit significantly better. The chi-square difference test yielded a difference of 41.025 with three degrees of freedom and a p-value below 0.001, and the comparative fit index improved by 0.016, a margin considered meaningful in psychometric practice.

The fit statistics for the four-factor model were solid by accepted standards. The ratio of chi-square to degrees of freedom came to 2.46, comfortably below the common cutoff of 3. The root mean square error of approximation, a measure of how well the model reproduces the observed covariances, was 0.070, just inside the threshold of 0.08 that most methodologists accept for adequate fit. The standardized root mean square residual was 0.068, and the comparative fit index and Tucker-Lewis index reached 0.940 and 0.920 respectively, both approaching or exceeding the 0.95 ideal and clearly above the 0.90 minimum. In plain terms, the questionnaire’s original architecture, with its subtle separation of routine and compensatory restraint, survives translation into Turkish better than a simplified three-factor reading of the data would suggest.

Reliability, the question of whether the scale produces consistent measurements, was assessed three ways. Cronbach’s alpha for the overall scale reached 0.864, indicating strong internal consistency, while the subscales ranged from 0.708 to 0.924, spanning acceptable to exceptional. McDonald’s omega, an index less sensitive to violations of equal item loadings, ranged from 0.725 to 0.927 across subscales, corroborating the alpha results. The team also examined test-retest stability using intraclass correlation coefficients, checking that people who answered the questionnaire at one point in time gave substantially similar answers when asked again. Together, these indices show that the Turkish WREQ does not merely hang together conceptually; it delivers stable, repeatable scores of the kind required for both research and clinical screening.

Validity was triangulated against two established instruments: the Three-Factor Eating Questionnaire, known as TFEQ-R21, and the Dutch Eating Behaviour Questionnaire, or DEBQ. If the Turkish WREQ truly measures weight-related eating, its subscales should correlate with the corresponding subscales of these older tools, a property called convergent validity. That is precisely what the researchers found, with correlation coefficients ranging from 0.401 to 0.690, all statistically significant at p below 0.001. Discriminant validity was confirmed through the heterotrait-monotrait ratio, or HTMT, a modern criterion for showing that supposedly distinct constructs are not measuring the same thing in disguise; all HTMT values fell below the 0.85 cutoff. Measurement invariance across sex was partially supported, with changes in the comparative fit index of no more than 0.002, meaning men and women interpret the items in largely comparable ways.

The study also touched on a question of broad public interest: does eating style relate to body weight? In this sample, body mass index showed minor positive correlations with most of the WREQ subscales across the whole group, a pattern consistent with the idea that greater susceptibility to external cues and emotional eating tends to accompany higher body weight, even within a predominantly normal-weight population. The authors are careful not to overstate this, and the modest correlations reflect the young, largely lean sample. Still, the finding hints at the questionnaire’s practical value: identifying people who score high on emotional eating or external eating could flag those at risk of weight gain before it happens, opening a window for early, targeted nutritional counseling.

What does this mean beyond Türkiye? Validated translations of brief behavioral instruments are the connective tissue of international nutrition science. When the same questionnaire is available in dozens of languages with demonstrated equivalence, researchers can pool data across countries, compare eating styles across cultures, and test whether interventions that work in one population generalize to another. For Turkish clinicians, dietitians, and researchers, the WREQ now offers a quick, reliable way to profile eating behavior in young adults, whether in obesity clinics, university health services, or large-scale epidemiological studies. The work, published open access under a Creative Commons license, received no external funding, and the authors declare no competing interests. In an era when eating disorders and obesity are rising in parallel, a short, trustworthy window into why people eat the way they do is a small tool with outsized potential.

Subject of Research: Cross-cultural validation of the Weight-Related Eating Questionnaire in Turkish young adults

Article Title: Validity and reliability of the Turkish version of the weight-related eating questionnaire in young adults

Article References: Çelik, E., Kocyigit, E., & Cemali, Ö. (2026). Validity and reliability of the Turkish version of the weight-related eating questionnaire in young adults. BMC Psychology. https://doi.org/10.1186/s40359-026-05422-4

Image Credits: AI Generated

DOI: 10.1186/s40359-026-05422-4

Keywords: eating behavior, psychometrics, questionnaire validation, young adults, emotional eating, dietary restraint, obesity, Türkiye, factor analysis, reliability, nutrition, public health