Impact of Rating Scale Categories on Reliability and Fit Statistics of the Malay Spiritual Well-Being Scale using Rasch Analysis.

BACKGROUND Few studies have employed the item response theory in examining reliability. We conducted this study to examine the effect of Rating Scale Categories (RSCs) on the reliability and fit statistics of the Malay Spiritual Well-Being Scale, employing the Rasch model. METHODS The Malay Spiritual Well-Being Scale (SWBS) with the original six; three and four newly structured RSCs was distributed randomly among three different samples of 50 participants each. RESULTS The mean age of respondents in the three samples ranged between 36 and 39 years old. The majority was female in all samples, and Islam was the most prevalent religion among the respondents. The predominating race was Malay, followed by Chinese and Indian. The original six RSCs indicated better targeting of 0.99 and smallest model error of 0.24. The Infit Mnsq (mean square) and Zstd (Z standard) of the six RSCs were "1.1"and "-0.1"respectively. The six RSCs achieved the highest person and item reliabilities of 0.86 and 0.85 respectively. These reliabilities yielded the highest person (2.46) and item (2.38) separation indices compared to other the RSCs. CONCLUSION The person and item reliability and, to a lesser extent, the fit statistics, were better with the six RSCs compared to the four and three RSCs.