34 Chapter 2 Data analysis Treatment outcome (binge eating pathology) was defined as the score on the bulimia subscale of the EDI-1 at post-treatment and at six months follow-up. To compare characteristics of treatment completers with treatment dropouts, we performed Student’s t-tests to test for differences between the two groups on eating disorder psychopathology (EDI-1 and BAT), general psychopathology (SCL-90 and BDI) and personality (NEO-PI-R), BMI and age. χ2 tests were conducted to test for differences on gender, social embedding, level of education, and employment. For the above described tests, α was set at .05. Treatment effects (pretreatment versus post-treatment and post-treatment versus follow-up) were tested using repeated measures analyses of variance (ANOVAs) with time as the within-subject variable and bulimia scale scores or BMI as the outcome variable. For treatment effects, to correct for multiple testing, α was set at .01. Pretreatment predictors for post-treatment bulimia scale scores were determined using hierarchical linear regression analysis (using n = 304: treatment completers that also completed the post-treatment assessment). Initial predictors were pretreatment bulimia scale scores, the remaining pretreatment EDI subscales, the subscales and total scores of the SCL-90 and the BAT, as well as the NEO-PI-R scales, BDI total score, and background variables (gender, age, BMI, social embedding and level of education), which were selected using a backward manual stepwise elimination using p > .10. Outcome measure was the post-treatment bulimia scale score. Similar analyses were conducted to predict follow-up bulimia scale scores based on the post-treatment predictors (using n = 190: treatment completers that also completed the follow-up measurement). Missing data were left missing; no data imputations were conducted. For each individual test, listwise deletion was performed, resulting in a somewhat different n between tests. Results Differences between treatment completers and treatment dropouts Treatment completers did not differ from treatment non-completers with respect to eating disorder psychopathology or general psychopathology. Some significant differences in personality existed: treatment completers are more agreeable and more conscientious than non-completers. In addition, treatment completers have significantly
RkJQdWJsaXNoZXIy MjY0ODMw