The analysis of clinical studies: The analysis of count data, part I

Abstract

Often variables are naturally measured on a nominal or categorical scale. An instance of this is histology type for some cancer site. Another is health status, which can be described on a scale ranging from excellent to poor. In the second example, the scale also has an intrinsic ordering. Nominal data can also arise in situations in which the original data have been measured on a numerical scale, but investigators find it more convenient to collapse the scale into two or more categories. Categorical data require different descriptive and analytical approaches from the ones that have been discussed in previous columns.