Statistical analysis

Data were analyzed with Statistical Package for Social Sciences (SPSS) software, version 14. Descriptive data are presented as %s or as mean ± SDs. Significance of associations was tested using Chi square for categorical variables and Student’s t-test for continuous variables. For the purpose of a logistic regression analysis, participants were classified into 2 sub-groups depending on examination results; (1) ‘Pass’ category: participants who have passed the CS, ApSS or IBSS examinations in the respective groups, and (2) ‘Fail’ category: those who have not passed the above examinations in the respective groups. A binary logistic-regression analysis was performed in all patients with ‘Examination Results’ as the dependent variable (0=’Fail’; 1=’Pass’) and EI score, PSS score, gender (0 = male; 1 = female), satisfaction about selecting medicine (0 = unsatisfied; 1 = satisfied) and extra-curricular activities (0 = no; 1 = yes) as the continuous or dichotomous independent variables and confounding factors. A backward elimination procedure was used to select the explanatory independent variables, where a p-value of 0.10 was considered as the cut-off for variable exclusion. In all analyses a p-value ≤ 0.005 was considered statistically significant (a Bonferroni type adjustment was made to reduce the Type I error in multiple analysis).