Longitudinal is usually understood to mean, simply, that a phenomenon is observed for more than one year. In tests and measurements, however, there are two kinds of longitudinal studies. The first is cross-sectional. That is a study that looks at multiple years, but includes all students who participate in any given year, whether those students were present for the other years or not.

A multiple year cross-sectional study helps separate performance that is unique to a specific cohort (i.e. a group of students in a single grade) from performance that is related to teaching and other factors. For example, the district does not know at the present time whether the low fall test scores for first graders will be unique to the first grade class of 2000-01 or whether it will show up with other classes as well. A multiple year cross-sectional study will help to answer that question.

A true longitudinal study eliminates any student who was not present and tested in each of the years of the study. This is very useful in determining what changes in performance are attributable to what is happening in Rochester as opposed to changes that are caused by students moving in and out of the district.