Interpreting the t-tests

The following example illustrates several issues and problems associated with the interpretation of p-values arising out of the t-tests. The setting is a very common one where a variable is measured at baseline and then subsequently at the end of the treatment period and the analysis focuses on the relative effects of the two treatments. 

Example 4.1 Comparison of two active treatments for the treatment of Major Depressive Disorder in a randomised control trial 

The primary endpoint in this trial was the 17 point Hamilton Depression Scale (HAMD-17) and the data presented in Table 4.2 correspond to mean (se). 

Unpaired t-test comparing treatment means at baseline, p = 0.73 

---