Table 1.
Characteristics of study participants.
Table 2.
Descriptive statistics and Cronbach’ alphas for study variables (N = 2 236).
Table 3.
Pearson’s correlations between the study measures (N = 2 236).
Fig 1.
Path analysis for eating disinhibition and susceptibility to hunger.
Mediation model with standardized estimates and R2 for weight-related negative affect, disinhibition and susceptibility to hunger. Model fit: χ2(8) = 15.134 (p = 0.057); CMIN/DF = 1.892; CFI = 0.999; TLI = 0.997; NFI = 0.999; RMSEA = 0.020, CI = 0.000 to 0.035, p = 1.000). When the effect of depressive symptoms was controlled for, the impact of shame, hated self and reassured self on disinhibition and susceptibility to hunger was fully mediated by their effect on weight-related negative affect. Inadequate self and negative social comparison predict higher disinhibition and susceptibility to hunger directly and partially through increased weight-related negative affect. Note: Depressive symptoms = Depression, Anxiety and Stress Scale-21 Depression subscale; External Shame = Weight-focused External shame Scale; Inadequate Self, Hated Self, Reassured Self = Weight-focused Self-Criticising/Self-Reassuring Scale subscales; Social Comparison = Weight-focused Social Comparison Scale; Weight-Focus. Negative Affect = Weight-focused Feelings Scale Negative Affect subscale; Disinhibition, Susceptibility to hunger = Three Factor Eating Questionnaire subscales.
Fig 2.
Path analysis for BMI changes.
Mediation model with standardized estimates and R2 for weight-related negative affect, disinhibition, susceptibility to hunger and BMI changes. Model fit: χ2(16) = 75.970 (p < 0.001); CMIN/DF = 4.748; CFI = 0.994; TLI = 0.984; NFI = 0.993; RMSEA = 0.042, CI = 0.033 to 0.052, p = 0.897).