Abstract
One of the primary objectives of certain experiments is to determine the impact of various sequences of medications, nutritional regimens, or learning trials. In scenarios where experimental units are scarce or budgetary constraints exist, researchers often subject each unit (subject) to multiple consecutive tests, a design formally known as Repeated Measures Experiments. While data in such studies are frequently quantitative, researchers often encounter cases where variable levels are defined solely on an ordinal basis, focusing on the frequency of observations at each level; this type of information is referred to as Categorical Count Data. This research focuses on repeated measures tests for categorical data, specifically examining the Cochran, McNemar, Ireland & Kullback, Stuart, and Bhapkar tests for two-treatment designs with two levels each. Furthermore, the Stuart, Bhapkar, and Ireland & Ku & Kullback tests are evaluated for two-treatment designs with more than two levels, alongside an analysis of the Cochran and Ireland & Kullback tests, and Weighted Least Squares (WLS) methods for three-treatment designs with two levels. By applying these statistical tests to real-world datasets, this study compares their performance based on the obtained results to derive a comprehensive set of scientific conclusions.
DOI
10.33095/jeas.v16i59.1502
Subject Area
Statistical
First Page
187
Last Page
209
Recommended Citation
Hussein, S. M., & Obeid, H. K. (2010). Statistical Analysis of Repeated-Measures Experiments with Categorical Data Under Two and Three Treatments. Journal of Economics and Administrative Sciences, 16(59), 187-209. https://doi.org/10.33095/jeas.v16i59.1502
