Because the relationship between computer use and achievement is still puzzling, there is a need to prepare and analyze good quality datasets on computer use and achievement. Such a dataset can be derived from TIMSS data. This paper describes how this dataset can be prepared. It also gives an example of how the dataset may be analyzed. The analysis evidenced that (1) computers should not be used frequently, and (2) they should be used as a learning tool more often than as an information tool.