tmcomptest

A statistical analysis used for count data is to compare the proportions of a binary outcome of interest for two or more groups.

This m-file it is an implemented alternative to the Marascuillo's test and identifies significant pairwise contrasts that this last one does not. If the previously Chi-square test results in a p-value smaller to an alpha-value, the results are deemed significant, the null hypothesis that all proportions are equal is rejected, and it is concluded that there is a significant difference between at least two of the proportions.

We can look at the pairwise comparisons obtained out of the Tukey's honest significant difference test employed. You will observe the comparisons identified as significant. We can compare up to 20 proportions.

Syntax: tmcomptest(X,alpha)

Input:
X - data matrix size rx2. Rows=number of groups. Column 1=number of successes or interested events. Column 2=number of trials.
alpha - significance level (default = 0.05)

Output:
- Complete Tukey's HSD multiple comparisons test among proportions table