Psychometric Properties of the Turkish Version of the Flexible or Rigid Control of Eating Scale (FORCES) in Women

Abstract

Background: Eating control is a central construct in eating disorder research; however, existing measures often insufficiently capture its multidimensional nature and rarely distinguish rigid and flexible regulatory styles. The Flexible or Rigid Control of Eating Scale (FORCES) was developed to address this limitation; however, no culturally adapted instrument assessing rigid and flexible eating control has previously been validated in Turkish populations. Methods: This study aimed to translate, culturally adapt, and psychometrically evaluate the FORCES in a non-clinical sample of Turkish women. Two independent cross-sectional online samples were included. Exploratory factor analysis (EFA; n=416) was conducted to examine whether the original multidimensional structure would be replicated in a Turkish context, followed by confirmatory factor analysis (CFA; n=417). Internal consistency, test–retest reliability, convergent validity, construct validity, and Differential Item Functioning (DIF) were evaluated. Item Response Theory (IRT)-based Rasch analyses examined item-level psychometric properties. Construct validity was assessed using Three-Factor Eating Questionnaire (TFEQ-R21), Dichotomous Thinking in Eating Disorders Scale (DTEDS), State-Trait Anxiety Inventory (STAI), and Body Mass Index (BMI). Test–retest reliability was evaluated in a subsample (n=53) over a two-week interval. Results: EFA supported the original five-factor structure, explaining 62.85% of the variance, with factor loadings ranging from 0.589 to 0.921. CFA confirmed this structure and demonstrated good model fit (χ²/df=2.655, CFI=0.951, TLI=0.940, RMSEA=0.063, SRMR=0.035). The overall scale demonstrated high internal consistency (α=0.874; ω=0.880), while subscale reliability coefficients ranged from 0.766 to 0.917 for Cronbach's alpha and from 0.775 to 0.918 for McDonald's omega. Rasch analyses supported satisfactory item functioning and substantial measurement equivalence across BMI groups. Test–retest reliability coefficients indicated satisfactory temporal stability (ICC=0.726–0.821). Rigid control dimensions were moderately associated with emotional eating, dichotomous thinking, anxiety, and BMI (r=0.20–0.49, p<0.01), whereas flexible control dimensions showed weaker but significant associations (r=0.10–0.25, p<0.01), supporting the distinction between rigid and flexible eating control. Conclusions: The Turkish version of the FORCES is a reliable and valid instrument for assessing flexible and rigid eating control in women. By distinguishing maladaptive rigid control from adaptive flexible regulation, the scale provides a multidimensional framework for investigating eating regulation and may support future cross-cultural and behavioral research in Turkish populations.