Recontextualising dieting: defining, assessing, and exploring select psychological outcomes and determinants

Abstract

In this thesis I review the historical conceptualisations and operationalisations of dieting, propose a broadened definition of dieting designed to better reflect the range of dieters’ lived experiences, and present a new questionnaire developed to align with this updated conceptualisation (Chapter 3). Data from two longitudinal survey studies conducted in the United Kingdom with community samples of adults (i.e., the COVID and HADES datasets; described in Chapter 2) were used to complete all analyses including the psychometric tests of the dieting scale. An exploratory factor analysis of the seven-item dieting scale revealed a unifactorial structure which also demonstrated construct validity, internal consistency, and two-month test-retest reliability (COVID dataset). To address limitations identified by the previous study, the original scale was revised such that the response scale and wording of the original items were modified, and five original items were added. The revised 12-item scale maintained a unifactorial structure (as assessed by a confirmatory factor analysis) with good psychometric properties (construct validity, internal consistency, six-week test-retest reliability; HADES dataset). The remaining empirical chapters utilise the dieting scale to explore various correlates and determinants of dieting.

Specifically, in Chapter 4 I used two random intercept cross-lagged panel models to explore the reciprocal associations between appearance satisfaction and dieting using the COVID dataset. Results partially supported the presence of bidirectional within-person associations and suggested that longer intervals (≥ 4 months) may be required to identify within-person effects.

Chapter 5 explores psychological inflexibility related to eating as a potential attribute that could help refine the precision of dieting as a predictor of eating disorder psychopathology. Because the Inflexible Eating Questionnaire (used to assess psychological inflexibility related to eating) had never been validated in a British sample, I conducted a confirmatory factor analysis with the 11-items using the COVID dataset. The previous unifactorial model was not supported. An exploratory factor analysis identified 3 problematic items which were removed for subsequent analyses. The revised scale demonstrated additional psychometric properties (construct validity, internal consistency, two-month test-retest reliability) and established full measurement invariance between participants who did and did not report a current or past eating disorder diagnosis. Using robust linear regression, psychological inflexibility related to eating was found to moderate the association between current levels of dieting and current and future levels of eating disorder psychopathology, but not change in eating disorder psychopathology (i.e., not controlling for baseline levels).

In Chapter 6 I discuss the development and validation of a measure designed to assess healthism attitudes related to food, eating, and weight, and explore associations between these attitudes and (mal)adaptive eating practices using correlations and hierarchical regression analyses (HADES data). An exploratory factor analysis identified a three-factor solution with factors representing healthism: Beliefs (9 items), Negative Emotions (12 items), and Positive Emotions (6 items). The healthism scale achieved acceptable model fit and the subscales demonstrated construct validity, internal consistency, and six-week test-retest reliability. Each subscale was significantly associated with dieting, eating disorder psychopathology, and intuitive eating levels at the same timepoint and six weeks later. However, compared to the other subscales, the Beliefs subscale was the best predictor of dieting and anti-fat attitudes, and the Negative Emotions subscale was the best predictor of body dissatisfaction and intuitive eating.

Taken together, these studies represent an important step towards improving our conceptual understanding of dieting, increasing our knowledge about the psychological risks posed by dieting, and more accurately predicting eating disorder psychopathology. I discuss the overall implications and directions for future research in the final chapter.