Abstract
The uses of summary statistics are reviewed and demonstrated, with emphasis on helping the reader understand how to choose a reasonable summary statistic function and how to base a test statistic on it. Particular attention is paid to the issues of test size and power, and how they might be affected by missing data. Recent findings which show how stratifying the analysis according to missing data patterns can help minimize the problems caused by missing data are summarized. This stratified approach is demonstrated in detail on data from a two-group clinical trial with repeated measures.
Get full access to this article
View all access options for this article.
