How Should We Analyze Likert Item Data?
ISSN 2158-9429
Volume 12, Issue 2 - December 2019
Editor: Bindu Poudel-Ward
Abstract
- Program evaluation in extension and outreach often relies on the assessment of Likert item data. Appropriate statistical approaches for this task include ordinal regression, traditional nonparametric tests, and aligned rank transformation analysis of variance. Here, these methods are compared for simulated data, and p values, type I errors, and power are presented. t tests and other traditional parametric approaches may be inappropriate for this kind of data. Researchers should not rely on p values as the sole determiner of a meaningful effect. Instead, analysis should also include effect size statistics and an assessment of practical importance.
-
View/Download Full Article