Research Showcase [MPhil] Development of a Scale of Obsessive Compulsive Disorder By Muhammad Kamruzzaman Mozumder
Supervised by Dr. Roquia Begum
Abstract
The present study was conducted to develop a scale for measuring Obsessive Compulsive Disorder in our culture. Rigorous review of the available literatures and case reports along with input from different mental health professionals were used as the base for the development of the scale. The scale was developed through systematic procedure of item construction, item selection and item analysis. After several careful revisions of the initial items pool (187 items), a total of 82 items were presented to 31 judges (Psychiatrists and Clinical Psychologists) for evaluation. Based on their evaluation only 32 items which met the selection criterion were compiled in the first draft of the scale using 5 point Likert format. Final scale of 20 items was devised from item analysis of these 32 items. Only those items that were able to fulfill the dual criterion of discrimination value (a < 0.01) aiid corrected itein-total correlation (a < 0.05) were selected for the final scale. Psychometric properties of the newly developed scale were checked after completing item analysis. The scale was found to possess content validit>', criterion related validity and constmct validity. Criterion related validity' of the scale was ensured by high and significant correlation (r =0.746, a < 0.01) found between the present scale and the Maudsley Obsessive Compulsive Inventory (MOCI; Hodgson and Flachman, 1977), Construct validity evidence came from convergent validation (correlation with MOCI) and discriminant validation (ability to discriminate bistween clinical and non-cliiiical group with F = 336.29, at a < 0.01). Internal consistency reliability measured by split-half (r =0.8913, a < 0.01) and Cronbach alpha method (0.9478) indicates high internal consistency of the present scale. Test-retest reliability done on 76 non diagnosed individual also revealed satisfactory- (r = 0.9J8, a < 0.01) result. Both the screening and severity nonns were developed for the scale. Screening norm was developed using sensitivity and specificity calculation. The ideal cut-off score for screening OCD was found to be 17 with sensitivity of 87% and specificity- of 90%. Diagnostic performance cf the scale was assessed using ROC curve. The area under ROC curve was found to be 0.923 which was significant at a < 0.01. Severity norm was constnicted using percentile method.