Non-parametric tests for small samples of categorized variables: A study

This paper presents a study on non-parametric tests to verify the similarity between two small samples of variables classified into multiple categories. The study shows that the only tests available for this situation are the chi-square and the exact tests. However, asymptotic tests, such as the chi-square, may not work well for small samples, leaving exact tests as the alternative. Nevertheless, if the number of classes increases, the implementation of these tests can become very difficult, in addition to requiring specific algorithms that may demand considerable computational effort. Therefore, as an alternative to the exact tests, a new test based on the difference between two uniform distributions is proposed. Computational assays are conducted to evaluate the performance of these three tests. Although non-parametric tests present numerous applications in various areas of knowledge, this study was motivated by the need to verify whether the business strategy adopted by a company is a determining factor for its competitiveness.

[1]  E. Pitman.  Significance Tests Which May be Applied to Samples from Any Populations , 1937 .

[2]  M. Kendall.  A NEW MEASURE OF RANK CORRELATION , 1938 .

[3]  Eugenia Stoimenova,et al.  Applied Nonparametric Statistical Methods , 2010 .

[4]  M. Kendall.  Statistical Methods for Research Workers , 1937, Nature.

[5]  E. J. G. Pitman,et al.  Significance Tests Which May be Applied to Samples from Any Populations. II. The Correlation Coefficient Test , 1937 .

[6]  M. Friedman.  The Use of Ranks to Avoid the Assumption of Normality Implicit in the Analysis of Variance , 1937 .

[7]  F. Wilcoxon.  Individual Comparisons by Ranking Methods , 1945 .

[8]  G. H. Freeman,et al.  Note on an exact treatment of contingency, goodness of fit and other problems of significance. , 1951, Biometrika.

[9]  Cristiano Lúcio de Souza.  Competindo pelo futuro , 1995 .

[10]  Hisashi Tanizaki,et al.  Power comparison of non-parametric tests: Small-sample properties from Monte Carlo experiments , 1997 .

[11]  H. B. Mann,et al.  On a Test of Whether one of Two Random Variables is Stochastically Larger than the Other , 1947 .

[12]  Marie Schmidt,et al.  Nonparametrics Statistical Methods Based On Ranks , 2016 .

[13]  J. Wolfowitz,et al.  On a Test Whether Two Samples are from the Same Population , 1940 .

[14]  K. Pearson.  On the Criterion that a Given System of Deviations from the Probable in the Case of a Correlated System of Variables is Such that it Can be Reasonably Supposed to have Arisen from Random Sampling , 1900 .

[15]  W. Kruskal,et al.  Use of Ranks in One-Criterion Variance Analysis , 1952 .

[16]  E. Pitman.  SIGNIFICANCE TESTS WHICH MAY BE APPLIED TO SAMPLES FROM ANY POPULATIONS III. THE ANALYSIS OF VARIANCE TEST , 1938 .

[17]  Herman Chernoff,et al.  ASYMPTOTIC NORMALITY AND EFFICIENCY OF CERTAIN NONPARAMETRIC TEST STATISTICS , 1958 .