All analyses were performed with SPSS 14.0 (31 ). Data were screened for univariate and multivariate normality. There were no influential outliers. Logarithmic transformations were made for total energy intake and fat-free mass, and arcsine transformations were conducted for the percentage of macronutrient content (fat, protein, and carbohydrate) intake. Pre- and post-meal mood states, as assessed by the BRUMS, were mean rank transformed to achieve normality.
Primary hypotheses were examined with a series of linear mixed models with repeated measures. The dependent variables were total energy intake (kcal), percentage of energy intake from protein, carbohydrate, and fat, food groups (dairy, desserts/snacks, meats, vegetables, fruits, bread, condiments, drinks), duration of eating (minutes), post-meal state Anxiety, and post-meal state Anger, Confusion, Depression, Fatigue, Tension, and Vigor. The fixed-factor main effects in the model were LOC (presence or absence of at least one episode in the month preceding assessment) and meal type (normal, binge). We also tested the two-way interaction between LOC and meal type to determine whether any observed effects of LOC significantly varied between the normal meal and the binge meal. Each model included sex, age, race (coded non-Hispanic Caucasian and all other racial/ethnic minorities), fat-free mass, and percent fat mass as covariates. For the models predicting duration of eating and post-meal state anxiety and mood, total energy intake also was included as a covariate. Pre-test meal anxiety or mood state was included in the models predicting post-test meal anxiety and mood states. Randomization order was also considered as a covariate in all analyses, but was removed because it did not significantly contribute to any model. Reported means and standard errors are adjusted for all variables included in each model. Differences were considered significant when p values were ≤0.05. All tests were two-tailed.
Primary hypotheses were examined with a series of linear mixed models with repeated measures. The dependent variables were total energy intake (kcal), percentage of energy intake from protein, carbohydrate, and fat, food groups (dairy, desserts/snacks, meats, vegetables, fruits, bread, condiments, drinks), duration of eating (minutes), post-meal state Anxiety, and post-meal state Anger, Confusion, Depression, Fatigue, Tension, and Vigor. The fixed-factor main effects in the model were LOC (presence or absence of at least one episode in the month preceding assessment) and meal type (normal, binge). We also tested the two-way interaction between LOC and meal type to determine whether any observed effects of LOC significantly varied between the normal meal and the binge meal. Each model included sex, age, race (coded non-Hispanic Caucasian and all other racial/ethnic minorities), fat-free mass, and percent fat mass as covariates. For the models predicting duration of eating and post-meal state anxiety and mood, total energy intake also was included as a covariate. Pre-test meal anxiety or mood state was included in the models predicting post-test meal anxiety and mood states. Randomization order was also considered as a covariate in all analyses, but was removed because it did not significantly contribute to any model. Reported means and standard errors are adjusted for all variables included in each model. Differences were considered significant when p values were ≤0.05. All tests were two-tailed.
