Reliability and factor analysis of the Sinhalese version of the Liebowitz Social Anxiety Scale-Self Reported Version (LSAS-SR)

Abstract

INTRODUCTION AND OBJECTIVES: Despite being one of the most prevalent mental illnesses, lack of awareness and confusion with shyness contributes to under-diagnosis of social phobia. Our objective was to evaluate test-retest reliability, internal consistency and factor analysis of the Sinhalese version of LSAS-SR - which is widely used in the English speaking world as a screening tool for social phobia. METHOD: The established process of translation by bilingual professionals/ back translation to establish consistency was carried out. Face validity and cultural acceptability of the translation was established using the Delphi technique. Pilot evaluation was conducted and standard psychometric procedures were used to evaluate test- retest reliabilityand internal consistency. The subjects were 871 SriLankan university students. RESULTS: Prevalence of social phobia among study population was 9.2%.Pearson's correlation coefficient was 0.998, 0.994 and 0.993 for fear and avoidance subscales and total sore respectively (p<0.01). Cronbach's alpha for fear and avoidance subscales with total score for each subscale was 0.753 and 0.752 respectively. Cronbach's alpha of total score of each sub scale with total score of LSAS-SR was 0.722.Factor analysis resulted in 5 dimensions and these were: factor I - speaking in a group, factor II-activity in public, factor III - social interaction with unknown person, factor IV - attitude of disagreement/ disapproval and factor V - social interaction in leisure activity. CONCLUSION: The findings confirm that this Sinhalese version of LSAS-SR has similar test-retest reliability, internal consistency and factor analysis as the original instrument and can be usedto screen for social phobia in Sri Lanka.