Dutch Eating Behavior Questionnaire-Restraint Scale (DEBQ-RS)

Abstract

The Dutch Eating Behavior Questionnaire-Restraint Scale (DEBQ-RS) is an internationally recognized, 10-item psychometric instrument designed to assess dietary restraint—the conscious and deliberate cognitive effort to restrict food consumption for the purpose of controlling body weight or avoiding weight gain. Originally developed by Tatjana van Strien, Jan E. R. Frijters, Gerard P. A. Bergers, and Peter B. Defares in 1986 as one of three orthogonal dimensions of the 33-item Dutch Eating Behavior Questionnaire (DEBQ), the Restraint Scale emerged out of rigorous psychometric critiques directed at earlier measures of restrained eating, specifically C. Peter Herman and Janet Polivy’s Restraint Scale (RS). Early instruments confounded intentional caloric reduction with psychological distress, food cravings, and body weight fluctuations. The DEBQ-RS isolates the behavioral and intentional components of food restriction, yielding a purer construct of dietary restraint.

Each of the 10 items is evaluated along a five-point Likert response format ranging from 1 (Never) to 5 (Always). Extensively validated across clinical, community, adolescent, and cross-cultural cohorts, the DEBQ-RS demonstrates exemplary psychometric robustness. Internal consistency estimates typically yield a Cronbach’s alpha between α = .90 and .95 across sexes and weight classifications, paired with remarkable two-week to one-year test-retest reliability ($r\; >\; .80$). Confirmatory factor analytic investigations consistently confirm a single-factor, unidimensional architecture that exhibits strong measurement invariance across biological sex, clinical eating disorder classifications, and body mass index (BMI) categories. This article offers an exhaustive academic evaluation of the DEBQ-RS, dissecting its historical origin, latent psychological construct, structural models, construct validity, clinical utility, psychometric coefficients, and exact item administration criteria.

Keywords

Dutch Eating Behavior Questionnaire, DEBQ-RS, dietary restraint, restrained eating, psychometrics, factor analysis, construct validity, eating disorders, weight management, self-regulation.

Authors

The Dutch Eating Behavior Questionnaire and its Restraint subscale were conceptualized and psychometrically operationalized by a multidisciplinary team of Dutch researchers in behavioral sciences and nutritional psychology:

• Tatjana van Strien, Ph.D. – Primary investigator and leading authority on human appetite and eating pathology. Professor Emeritus of Eating Behavior at Radboud University Nijmegen (Department of Psychology) and associated with the Department of Health Sciences at Vrije Universiteit Amsterdam, The Netherlands.
• Jan E. R. Frijters, Ph.D. – Professor of Sensory Science and Food Evaluation, Department of Human Nutrition, Wageningen Agricultural University (now Wageningen University & Research), Wageningen, The Netherlands.
• Gerard P. A. Bergers, Ph.D. – Research psychometrician and biostatistician collaborating across behavioral nutrition programs at Wageningen Agricultural University and community health institutions in the Netherlands.
• Peter B. Defares, Ph.D. – Professor of Clinical and Personality Psychology, Department of Clinical Psychology, University of Amsterdam, Amsterdam, The Netherlands.

Purpose

The primary purpose of the DEBQ-RS is to assess the intensity and frequency of an individual’s intentional attempts to suppress or limit food intake to achieve or maintain a lower body mass. In modern psychometric and clinical research, dietary restraint is a critical parameter in the etiology, maintenance, and treatment of eating disorders, obesity, and metabolic conditions. Prior to the development of the DEBQ in 1986, researchers lacked an unconfounded, psychometrically pure tool to assess restriction independent of overeating tendencies.

Overcoming Historic Psychometric Confounders

The seminal Restraint Scale developed by Herman and Mack (1975) and expanded by Herman and Polivy (1980) successfully introduced the scientific community to the concept of the “restrained eater.” However, empirical scrutiny revealed a fatal design limitation: it collapsed two fundamentally distinct behavioral phenomena into a single index. It combined conscious restriction with persistent weight fluctuations, guilt about overeating, and episodic disinhibition. Individuals who scored high on the Herman-Polivy RS often exhibited chaotic, dysregulated eating, alternating between brief fasts and massive binge episodes.

Consequently, researchers could not discern whether experimental phenomena—such as the “what-the-hell” counter-regulation effect—were driven by caloric restriction itself or by underlying emotional instability and weight cycling. The DEBQ-RS was specifically engineered to isolate the pure behavioral and cognitive dimension of restraint, eliminating questions pertaining to weight volatility, guilt, or binging.

Clinical and Experimental Applications

In clinical contexts, the DEBQ-RS serves several vital diagnostic and therapeutic monitoring functions:

• Differential Diagnosis in Eating Pathology: It distinguishes individuals with high dietary control (e.g., restrictive subtypes of Anorexia Nervosa) from those displaying subjective or objective bulimic episodes (Bulimia Nervosa and Binge Eating Disorder).
• Weight Management Intervention Tracking: It monitors whether patients enrolled in behavioral weight-reduction therapies are adopting controlled, planned nutritional habits versus uncontrolled, maladaptive starvation strategies.
• Experimental Food Ingestion Paradigms: Within behavioral laboratories, the DEBQ-RS classifies participants before preload challenges, taste-test tasks, and cue-reactivity studies, allowing researchers to observe how self-imposed restriction moderates biological hunger and cephalic phase responses.

Psychological Construct

The psychological construct captured by the DEBQ-RS is restrained eating, operationalized as the ongoing, deliberate cognitive suppression of biological appetite and caloric intake to regulate body shape and weight. Unlike physiological hunger, which is governed by complex homeostatic and neuroendocrine signaling (ghrelin, leptin, insulin, and hypothalamic circuitry), dietary restraint operates via top-down executive and cognitive control mechanisms.

Behavioral Manifestations of Restraint

The construct encompasses several behavioral and cognitive modalities, all represented across the 10 scale items:

• Direct Quantitative Reduction: Consciously choosing to consume portions smaller than desired or needed for physiological satiety (e.g., item 2: “Did you try to eat less at meal times than you would like to eat?”).
• Selective Food Avoidance: Systematically declining calorie-dense items and selecting foods based on perceived slimming properties rather than sensory preference or hunger (e.g., item 5: “Did you deliberately eat foods that were slimming?”).
• Compensatory Temporal Restriction: Imposing strict temporal rules, such as eliminating evening consumption or engaging in deliberate post-intake compensation following perceived overindulgence (e.g., item 6: “If you ate too much, did you eat less than usual the next day?”; item 9: “How often in the evenings did you try not to eat because you were watching your weight?”).
• Hypervigilant Caloric Accounting: Maintaining chronic cognitive surveillance over all nutritional choices, ensuring that every dietary decision is filtered through the lens of body weight maintenance (e.g., item 4: “Did you watch exactly what you ate?”; item 10: “Did you take into account your weight in deciding what to eat?”).

Restraint as a Unidimensional Spectrum

Within the DEBQ paradigm, dietary restraint is conceptualized as a continuous, unidimensional spectrum. At the low end of the spectrum are non-restrained eaters who rely primarily on visceral, interoceptive signals of hunger and satiety to guide consumption. At the elevated end are individuals who disregard interoceptive states in favor of rigid cognitive boundaries. When individuals rely solely on cognitive boundaries, they become vulnerable to total disinhibition when those boundaries are breached, making the continuous measurement of this construct central to clinical psychology.

Theoretical Framework

The theoretical bedrock of the DEBQ-RS rests upon Herman and Polivy’s Restraint Theory (1980) and their foundational Boundary Model of Eating Regulation (1984), supplemented by subsequent cognitive and self-regulatory theories of ingestive behavior.

The Boundary Model of Eating

The Boundary Model posits that biological mechanisms establish two distinct physiological boundaries for food intake: an aversive hunger boundary (below which biological hunger compels feeding) and an aversive satiety boundary (above which physiological fullness produces discomfort and terminates consumption). In healthy, non-restrained individuals, an expansive “zone of biological indifference” separates these thresholds, where eating is guided by sensory enjoyment, habit, and social customs.

Restrained eaters, however, superimpose an artificial, cognitively defined diet boundary well below their physiological satiety threshold. This diet boundary represents the maximum caloric allotment the individual permits themselves to consume on any given day. Because this boundary is cognitive rather than biological, it requires continuous self-monitoring and mental exertion.

The Counter-Regulation Phenomenon (“What-the-Hell” Effect)

A central tenet derived from this framework is the phenomenon of counter-regulation. In laboratory experiments, when a non-restrained eater consumes a high-calorie “preload” (such as a large milkshake), their physiological satiety boundary is approached, leading them to consume significantly less food during a subsequent ad libitum taste-test. In contrast, when a restrained eater is induced to consume a preload that visibly violates their subjective diet boundary, their cognitive control collapses. Experiencing what Herman and Polivy termed the “what-the-hell” effect, the restrained eater consumes substantially more food than if no preload had been consumed, eating until they reach the far higher physiological satiety boundary.

Integration with Three Eating Styles

Van Strien and colleagues integrated this model into a tripartite taxonomy of human eating behavior, contrasting restrained eating directly with:

1. Emotional Eating: Consumption elicited by internal negative emotional states (such as anxiety, boredom, or depression), rooted in Kaplan and Kaplan’s psychosomatic theory.
2. External Eating: Ingestion driven by salient environmental and sensory food stimuli (such as smell, visual appeal, or availability) regardless of internal hunger states, derived from Stanley Schachter’s externality theory of obesity.

By establishing that the DEBQ-RS correlates independently of the DEBQ Emotional and External scales, the authors demonstrated that cognitive dietary restraint operates as a distinct psychological mechanism, even if it often interacts with external or emotional triggers to precipitate dysregulated binging.

Validity

Extensive psychometric investigations have established the validity of the DEBQ-RS across clinical cohorts, non-clinical populations, varied age groups, and different nationalities.

Construct and Structural Validity

Construct validity is substantiated by the scale’s capacity to represent intentional food limitation without confounding elements of overeating. In contrast to Herman and Polivy’s RS, which repeatedly demonstrates a two-factor structure (“Dietary Concern” and “Weight Fluctuation”), confirmatory factor analyses of the DEBQ-RS unequivocally confirm a single, homogeneous latent factor. The items display exceptionally high factor loadings, reflecting pure restraint behaviors.

Convergent Validity

The DEBQ-RS demonstrates strong convergent validity when compared against other psychometric assessments of eating behavior:

• Three-Factor Eating Questionnaire (TFEQ): Correlations between the DEBQ-RS and the Cognitive Restraint scale (Factor I) of the TFEQ (Stunkard & Messick, 1985) are consistently high, ranging between $r\; =\; .78$ and $r\; =\; .89$, confirming that both scales capture the same underlying intentional construct.
• Eating Disorder Inventory (EDI): The DEBQ-RS displays strong positive correlations with the Drive for Thinness subscale of the EDI ($r\; =\; .65$ to $.75$), illustrating that extreme cognitive restraint is intimately linked with thinness pressures and body image dissatisfaction.

Discriminant Validity

Discriminant validity is evidenced by near-orthogonal relationships with scales measuring emotional instability and reactive eating. In the original standardization studies by van Strien et al. (1986), the DEBQ-RS shared minimal variance with the DEBQ Emotional Eating Scale ($r\; =\; -.08$ to $.12$) and the DEBQ External Eating Scale ($r\; =\; -.15$ to $.05$). This confirms that intentional restraint is structurally and empirically distinct from cue-driven or emotion-induced overconsumption.

Predictive and Behavioral Validity

The behavioral validity of the DEBQ-RS has been verified through laboratory-based food consumption designs and nutritional diary assessments:

• In prospective dietary studies, individuals scoring high on the DEBQ-RS consume fewer daily total calories, lower percentages of dietary fat, and elevated proportions of diet-branded and low-energy foods compared to low scorers.
• Under experimentally induced cognitive loads, high scorers display marked attention biases toward calorie-dense nutritional cues during modified Stroop tasks and visual probe paradigms, confirming the psychological salience of dietary restriction in cognitive processing.

Reliability

The DEBQ-RS has consistently demonstrated high reliability across diverse demographic samples, language translations, and clinical environments.

Internal Consistency

Internal consistency metrics for the DEBQ-RS are consistently robust:

• In the original validation cohorts comprising Dutch adult men and women, Cronbach’s alpha coefficients were reported at α = .95 for women and α = .93 for men (van Strien et al., 1986).
• In subsequent cross-cultural adaptations—including English, Spanish, German, French, and Chinese populations—Cronbach’s alpha values have routinely ranged between α = .90 and .95 in non-clinical samples, and exceeded α = .92 in clinical cohorts diagnosed with Bulimia Nervosa or Binge Eating Disorder.
• Corrected item-total correlations for each of the 10 items consistently surpass $r\; =\; .60$, with items such as “Did you watch exactly what you ate?” and “Did you deliberately eat less in order not to become heavier?” frequently demonstrating item-total correlations exceeding $r\; =\; .75$.

Test-Retest Stability

Temporal stability assessments reveal high test-retest reliability across multiple temporal intervals:

• Across a two-week interval, test-retest correlation coefficients have been documented at $r\; =\; .92$.
• Over extended spans of six months to one year in untreated community controls, temporal coefficients remain high ($r\; =\; .80$ to $.85$), indicating that cognitive dietary restraint functions as an enduring, trait-like behavioral disposition rather than a volatile affective state.

Factor Analysis

Extensive exploratory (EFA) and confirmatory factor analyses (CFA) have affirmed the structural properties of the DEBQ-RS, both in isolation and within the 33-item DEBQ.

Exploratory Factor Structure

During the original development by van Strien and colleagues (1986), principal component analysis with varimax and oblimin rotations of the full 33-item inventory extracted three primary, distinct factors:

1. Restrained Eating (Factor I, accounting for the single largest proportion of total scale variance, with 10 items).
2. Emotional Eating (Factor II, comprising 13 items).
3. External Eating (Factor III, comprising 10 items).

All 10 Restraint items loaded unequivocally onto Factor I, with salient factor loadings ranging from .62 to .86, and negligible cross-loadings (< .20) on the Emotional and External eating dimensions.

Confirmatory Factor Analytic (CFA) Fit Indices

Subsequent modern CFA studies have assessed the goodness-of-fit for the unidimensional 10-item Restraint model. Across male, female, adolescent, and adult clinical populations, structural equation modeling yields excellent fit indices:

• Comparative Fit Index (CFI): Routinely exceeds $.95$ (frequently $>\; .97$).
• Tucker-Lewis Index (TLI): Consistently maintains values $>\; .96$.
• Root Mean Square Error of Approximation (RMSEA): Consistently falls below $.06$ (typically between $.035$ and $.054$), with narrow 90% confidence intervals.
• Standardized Root Mean Square Residual (SRMR): Remains exceptionally low, typically $\le \; .04$.

Measurement Invariance

Multi-group CFA studies demonstrate strict measurement invariance (configural, metric, scalar, and residual invariance) across:

• Gender: Establishing that the latent construct of dietary restraint has identical psychological meaning and item functioning in both men and women.
• Weight Status: Invariant across normal weight, overweight, and class I–III obesity populations.
• Age Demographics: Validated invariance spanning late childhood, adolescence, young adulthood, and geriatric cohorts.

Instrument / Measurement Tool

• Tool Name: Dutch Eating Behavior Questionnaire-Restraint Scale (DEBQ-RS)
• Associated Inventory: Dietary Restraint Subscale of the 33-item Dutch Eating Behavior Questionnaire (DEBQ)
• Authors: Tatjana van Strien, Jan E. R. Frijters, Gerard P. A. Bergers, and Peter B. Defares (1986)
• Construct Assessed: Cognitive and behavioral dietary restraint; conscious, intentional restriction of food intake to control body weight
• Target Population: Adolescents and adults (ages 12 years through geriatric cohorts); applicable to general, athletic, and psychiatric/eating disorder populations
• Administration Format: Paper-and-pencil self-report inventory, computerized clinical testing battery, or remote web-based psychometric assessment
• Number of Items: 10 items
• Response Format: 5-point Likert scale (1 = Never, 2 = Seldom, 3 = Sometimes, 4 = Often, 5 = Always)
• Completion Time: Approximately 3 to 5 minutes
• Scoring Methodology:
  • All 10 items are positively keyed; there are no reverse-scored items.
  • Item responses are assigned numerical values from 1 to 5 corresponding to the chosen category.
  • The overall scale score is calculated by computing the mean of all 10 completed items: $text\{DEBQ-RS\; Score\}\; =\; frac\{sum\_\{i=1\}^\{10\}\; text\{Item\}\_i\}\{10\}$.
  • Scores range continuously from 1.00 to 5.00, where higher scores reflect greater levels of cognitive dietary restraint.
• Normative Reference Values (General Community Benchmarks):
  • Adult Females (Community): Mean ≈ 2.50 to 2.80 (SD ≈ 0.85)
  • Adult Males (Community): Mean ≈ 1.80 to 2.10 (SD ≈ 0.65)
  • Clinical Restrictive Cohorts (e.g., Anorexia Nervosa): Mean ≥ 3.80 to 4.50

Permissions & Fee and Test Year

The Dutch Eating Behavior Questionnaire and its Restraint Scale were first published in 1986:

• Publication Year: 1986
• Copyright & Intellectual Property: The original psychometric scale was published within the peer-reviewed scholarly literature by van Strien et al. in the International Journal of Eating Disorders (John Wiley & Sons, Inc.).
• Research & Non-Commercial Use: The scale items and scoring algorithms are freely available for scientific research, academic dissertations, and educational training, provided appropriate citation is given to Dr. Tatjana van Strien and the original 1986 publication.
• Commercial / Diagnostic Publishing: Standardized diagnostic paper testing forms, normative scoring manuals, and computerized diagnostic versions are commercially managed in certain jurisdictions by psychometric publishing companies (e.g., Hogrefe Publishing, Boom test uitgevers in the Netherlands). Researchers or clinicians planning wide-scale commercial applications or revenue-generating software implementations must obtain formal permissions or licenses from the copyright holders.

References

• Herman, C. P., & Mack, D. (1975). Restrained and unrestrained eating. Journal of Personality, 43(4), 647–660. https://doi.org/10.1111/j.1467-6494.1975.tb00727.x
• Herman, C. P., & Polivy, J. (1980). Restrained eating. In A. J. Stunkard (Ed.), Obesity (pp. 208–225). W.B. Saunders.
• Herman, C. P., & Polivy, J. (1984). A boundary model for the regulation of eating. Research Publications – Association for Research in Nervous and Mental Disease, 62, 141–156.
• Herbozo, S. (2007). The Effects of Ambiguous Appearance-related Feedback on Body Image, Mood States, and Intentions to Use Body Changes Strategies in College Women (Doctoral dissertation). Department of Psychology, College of Arts and Sciences, University of South Florida. Available online at: http://usf.sobek.ufl.edu/content/SF/S0/02/68/04/00001/E14-SFE0002487.pdf
• Stunkard, A. J., & Messick, S. (1985). The Three-Factor Eating Questionnaire to measure dietary restraint, disinhibition and hunger. Journal of Psychosomatic Research, 29(1), 71–83. https://doi.org/10.1016/0022-3999(85)90010-8
• van Strien, T., Frijters, J. E., Bergers, G. P., & Defares, P. B. (1986). The Dutch Eating Behavior Questionnaire (DEBQ) for assessment of restrained, emotional, and external eating behavior. International Journal of Eating Disorders, 5(2), 295–315. https://doi.org/10.1002/1098-108X(198602)5:2<295::AID-EAT2260050209>3.0.CO;2-T
• Wardle, J. (1987). Eating style: A validation study of the Dutch Eating Behaviour Questionnaire in normal subjects and women with eating disorders. Journal of Psychosomatic Research, 31(2), 161–169. https://doi.org/10.1016/0022-3999(87)90072-9
• Westenhoefer, J., Broeckmann, P., Münch, A. K., & Pudel, V. (1994). Cognitive control of eating behaviour and the disinhibition effect. Appetite, 23(1), 27–41. https://doi.org/10.1006/appe.1994.1032
• Westenhoefer, J., Stunkard, A. J., & Pudel, V. (1999). Validation of the flexible and rigid control dimensions of dietary restraint. International Journal of Eating Disorders, 26(1), 53–64. https://doi.org/10.1002/(sici)1098-108x(199907)26:1<53::aid-eat7>3.0.co;2-8

Items of the Scale