International Journal of Contemporary Research in Multidisciplinary

Abstract:

Manual of SF-36 does not support computation of total score of the scale () unlike multidimensional Well-being index, Human development index, etc. where total score are found for an individual or a country/region. The paper gives a method to transform scores of each item to follow normal distribution and find sub-scale scores and as convolution of normally distributed item scores, parameters of which can be estimated from the data. Addition and arithmetic aggregation of item scores is meaningful due to normality of such scores following same distribution with different parameters.  Normally distributed  scores avoid limitations of summative Liker t scores and facilitate parametric statistical analysis including statistical tests of equality of average of  in two groups for cross-sectional as well as longitudinal data.  In addition, helps to find responsiveness of SF-36 by assessing changes in two time periods either for an individual or a sample of individuals which in turn helps to draw progress-paths. Normality of proposed scores also helps to find psychometric features like factorial validity, discriminating value and reliability. The proposed method is illustrated with hypothetical data.

Keywords:

SF 36; Equidistant; Normal distribution; Transformation; Assessment of progress

Aim/Objectives

Methods