Abstract
Basic data on six components of Logistics Performance Index (LPI) were obtained from the survey using six Likert items. The scale can be regarded as a Likert scale consisting of six items where each component was rated by respondents on a scale of 1–5. LPI Reports use a single item to measure a complex construct and do not indicate reliability of the entire scale or an item (or question). Most of the existing methods of finding reliability of Likert scales with different sets of assumptions have been found to violate one or more features of such scales. The paper proposes two non-parametric measures based on directional statistics to find reliability of items and the scale used in LPI considering empirical probabilities of Item – Response categories without making any assumptions for the observed variables or the underlying variable being measured. Properties and advantages of the proposed methods are discussed along with empirical verification with a hypothetical data. Use of Non-parametric reliability is recommended for Likert-type data for clear theoretical advantages and easiness in calculations.
Key words: Angular Association, Bhattacharyya’s measure, Likert scale, Most preferred direction, Polychoric correlations