Restrained eating: measuring an elusive construct

Area of Science:

• Psychology
• Behavioral Science
• Nutrition Science

Background:

• Herman's Restraint Scale (1978) is used to measure restrained eating and predict food consumption in laboratory settings.
• Previous research indicates the scale reliably predicts food intake in college students, irrespective of weight.
• The scale's generalizability and psychometric properties across diverse adult populations remain unestablished.

Purpose of the Study:

• To evaluate the generality and psychometric properties of Herman's Restraint Scale in a broader adult sample.
• To assess the scale's performance across different weight classifications (obese vs. normal) and dieting statuses (dieting vs. non-dieting).
• To examine the scale's reliability, item correlations, factor structure, and relationship with social desirability in adults.

Main Methods:

• A cross-sectional study involving 136 adult participants (male and female).
• Participants were categorized as obese or normal weight, and as dieting or non-dieting.
• A single questionnaire administered randomized items from the Restraint Scale, Lie Scale, and Social Desirability Scale.

Main Results:

• Significant differences in Restraint Scale scores were observed across adult groups: Obese dieters > Obese non-dieters > Normal weight individuals.
• Alpha reliability coefficients and corrected item-total correlations varied significantly across groups, indicating inconsistent psychometric properties.
• Factor analysis revealed three factors within the ten-item scale, and for obese dieters, the scale was not independent of social desirability.

Conclusions:

• The Restraint Scale demonstrates limited usefulness and questionable psychometric properties when applied to diverse adult populations outside of controlled laboratory settings with college students.
• Findings suggest that weight status and dieting behavior influence scale responses in adults.
• The scale's association with social desirability in obese dieters further questions its validity in non-student adult samples.