<![CDATA[This study aims to analyze the performance of tests of independence for ordinal data. Tests of independence is one of the most frequently used statistical tools in econometrics. Researchers are often interested about the independence of variable summarized in Contingency Tables. Many tests are available in literature to test independence in Contingency Tables, however, there is no clarity about the choice of tests which are incapable to provide comparison of large number of tests. A central problem and question facing researcher is to decide which tests of independence is most powerful for ordinal data. Most of the studies make pairwise comparison of tests and such studies are unable to guide about optimal test among a wide set of tests. This study used Monte Carlo Simulations (MCS) and compares seven popular tests of independence for ordinal data namely, Spearman coefficient of correlation, Kendall’s , Kendall’s coefficient, Goodman and Kruskal , Sumer’s D and Novel Phi_k . For ordinal data the stringency criteria cannot be applied to compare tests. Therefore, the comparison was made on the basis of power analysis and our MCS concludes based on solid estimations using power criterion that the most powerful test is Novel in k CT’s for ordinal data. ]]>
Copyrights © 2024
