Supervised by Kamal Uddin Ahmed Chowdhury

Abstract

Aim of the current study was to develop a valid and reliable scale of cognitive distortions which is applicable in Bangladesh. Scrupulous review of the available literature, research, case reports and expert opinions of the mental health professionals about cognitive distortion were used as a basement for the development of the scale. The scale was developed through systematic procedure of item construction, item selection, item analysis and factor analysis. Initial item pool consisted of 193 items which was evaluated by 16 judges in two stages judge evaluation and the item was reduced into 92 considering the appropriateness. For finalizing the first draft, 92 appropriate items with four point Likert scale were presented for 3rd round judge evaluation and 69 items were selected based on 90 % judge agreement. Item analysis was used for selecting those items which were able to fulfill the twofold criterion, discrimination value (a<0.01) and corrected item-total correlation (a<0.01). A total of 168 clinical and non-clinical respondents were used for item and factor analysis. Through this item analysis all 69 items fulfilled these criteria. To increase clinical utility, factor analysis was done for item reduction. By factor analysis 39 items having the factor loading of 0.55 and above were selected for final scale. For ensuring psychometric properties the number of respondents was 478.Content validity was confirmed through multi stage judge evaluation. Criterion related validity was established by calculating concurrent and predictive validity which were (r=0.828,a<0.01) and (r =0.828, r = 0.756, a<0.01) respectively. Construct validity of the current scale came from convergent validation and discriminate validation which were (r=0.670, a<0.01) and (F= 649.564, a<0.01) respectively. Internal consistency reliability was calculated by Cronbach alpha and it was, a = 0.962, that indicates excellent level of internal consistency. Test-retest reliability was also satisfactory (r= 0.890,a<0.01). Finally both screening and severity norms were developed for the current scale. Screening norm was established using sensitivity and specificity calculation and the most preferable cut off point of current scale was 56, with 87% sensitivity and 88% specificity. This cutoff point can discriminate clinical and non clinical level of cognitive distortion. Diagnostic performance of the current scale was measured through ROC curve and the area of ROC curve was 0.949 which was significant at a < 0.01, indicating excellent performance. Severity norms were developed through using percentile method. According to the percentile norm 56 to 72, 73 to 91,92 to 109 and 110 and above score represents mild, moderate, severe and profound level of cognitive distortions respectively.