Likert Data: What to Use, Parametric or Non-Parametric?
Jacqueline Murray

Abstract
The objective of the study was to determine whether the type of statistical tests conducted on Likert scale data affect the conclusions. Pearson, Spearman rho and Kendall tau_b analyses conducted on actual scale data revealed that there was a positive relationship between all the permuted pairs of the variables at the p<.05 level. However the relationship between the variables indicated a weak relationship for all of the tests except for the relationship between the constructs academic self-regulation and learning styles for which the Pearson and Spearman rho lead to the conclusion ofa moderate relationship. The coefficient of determination calculated to ascertain the amount of variability between the permuted pairs of the variables revealed similar variability forall of the variables except for the variables academic self-regulation and learning styles where yet again the conclusions for the Pearson and Spearman rho were similar and that forKendall tau_b different.

Full Text: PDF