The Use of Nonparametric Methods in Tax Research.

Abstract Nonparametric statistical methods should be used in tax research whenever the assumptions of classical statistics cannot be verified or the data are based on a ranking (an ordinal scale), as with Likert scale measurements. This paper discusses the advantages of using the nonparametric counterparts in place of many popular classical tests. The appropriate uses of three nonparametric procedures-the rank correlation coefficient, Kendall's coefficient of concordance, and the Wilcoxon signed rank test-are explained in detail and are illustrated using data reported by Porcano (1987) concerning the effectiveness of government tax incentives to encourage investment in fixed assets. The results verify and extend Porcano's conclusions.