Who are we testing? Self-selection bias in laboratory-based eating behaviour studies

• Published in: Appetite Vol. 141; p. 104330
• Main Authors: Haynes, Ashleigh, Robinson, Eric
• Format: Journal Article
• Language: English
• Published: England Elsevier Ltd 01.10.2019
• Subjects: Adolescent; advertising; anxiety; Behavioral Research; body mass index; college students; Eating behaviour; eating habits; Feeding Behavior - psychology; Female; females; Humans; Laboratory; Male; males; Methodology; motivation; Patient Selection; questionnaires; Recruitment; Research Design; Research Subjects - psychology; Selection Bias; Self-selection bias; Students - psychology; Surveys and Questionnaires; Young Adult; Eating behaviour; Laboratory; Methodology; Self-selection bias; Recruitment
• ISSN: 0195-6663; 1095-8304; 1095-8304
• DOI: 10.1016/j.appet.2019.104330

Research conducted in controlled laboratory settings plays an important role in understanding human eating behaviour. However, participant characteristics may affect motivation to participate in laboratory eating behaviour research and result in self-selection biases. First-year psychology students (n = 207, 90.8% female, m age = 18.8, m BMI = 21.9) were led to believe that they were choosing between enrolling in one of two laboratory-based studies: an ‘advertising’ study, or an ‘advertising and food’ study; the latter specified that participants would consume food and answer questions about their eating behaviour. Participants subsequently reported demographic information and completed a series of questionnaires assessing psychological characteristics hypothesised to predict self-selection in laboratory eating studies. The majority of participants preferred to participate in a study unrelated to eating over an eating-related study (70.5%, χ2 (1) = 34.90, p < .001). Participants choosing the eating-related study over the non-eating study had higher self-reported external eating tendencies and positive preoccupations with food, lower negative preoccupations with food and social appearance anxiety. Participants choosing the eating related study were also more likely to be male than female, although the number of males in the study was small (n = 19) and this finding should therefore be interpreted with caution. The present study provides some evidence that in university based samples of participants there are likely self-selection biases in laboratory-based eating behaviour studies and researchers may benefit from considering the potential influence of self-selection biases when conducting eating behaviour research.