Research News

Supervised by Dr. Roquia Begum

Abstract

The aim of the present study was to develop an anxiety scale for Bangladeshi population, which would measure the presence and severity of anxiety in an individual. Following the sequential system model or rational method of the scale construction (Jackson 1970; Wiggins, 1973) the objective of the present research was attained in three consecutive phases. In order to make the scale sensitive for the Bangladeshi population samples of this study were taken from those hospitals and clinics of Dhaka city where patients from almost all over the country usually attend. In the first phase, a total of 98 items referring anxiety symptoms were written in Bengali. These items were constructed on the basis of the statements of some real anxiety patients, on the opinions of experts on this area and from some foreign scales. After judging the face validity of the items by 22 judges 47 items were retained. These items were again given to 6 judges for second time judge evaluation. After editing the items according to the judges’ suggestions, 39 items were finalized as the final first draft of the scale. For scoring of each item of the scale, subjects had to respond on a 5-point rating scale, such as “never occurs'’, “mildly occurs", "moderately occurs”, “severely occurs”, and “profoundly occurs”. 1, 2, 3, 4 and 5 were assigned to these responses, respectively. The sum of the values was the total score of a respondent to the scale. In the second phase the scale was administered on a sample of 204 respondents (102 anxious subjects and 102 non-anxious subjects). Selected sample was purposive and the two groups of subjects were tried to match in respect to gender, age, marital status, education and income. The item analysis was then carried out on the items by computing their item total correlation and discriminating indices. Based on the results of item analysis, 36 items were selected for the proposed scale. In the third phase, reliability and validity of the final scale was estimated and two tentative norms were developed for the target population. The reliability of the scale was estimated by examining internal consistency (through conducting split-half reliability and coefficient alpha) and test- retest reliability. Split-half reliability was computed among 410 subjects and Guttman split-half correlation was found to be 0.7608. The test-retest reliability for the final form of the current anxiety scale was estimated on a sample of 123 non-anxious subjects with an interval of 2-3 weeks. The estimated reliability coefficient was found to be 0.699 (significant at a = .01). Concurrent validity and construct validity were estimated for the current scale. Three criteria were used for the estimation of concurrent validity: rating of anxiety by the psychiatrists, patients' self-rating of anxiety, and scores on Anxiety sub-scale of the Bengali version of Hospital Anxiety and Depression Scale (Chowdhury, 2000) on a group of 105 adult anxiety subjects. Present anxiety scale was found to be positively correlated with the psychiatrists' ratings of anxiety (r = 0.317), patients' self-rating of anxiety (r = 0.591) and the HADS anxiety sub-scale (r = 0.628). All the values were significant at a = 0.01 level. For the construct validity determination of the scale two groups of subjects (50 anxious and 50 non-anxious) were matched in respect to gender, age, marital status, education and income. Findings suggest that the scale has the ability to sharply discriminate between anxious and non-anxious subjects. Thus reliability and validity of the current scale were ensured. Finally, two different tentative norms were developed. At first a percentile norm assessing severity of anxiety was developed based on responses of 207 clinical subjects. Later on a screening norm for anxiety identification was developed based on responses of the total 410 subjects (207 clinical and 203 non-clinical subjects). All the above mentioned procedures were adopted for the development of an anxiety scale for Bangladeshi population.