The aim of this study was to investigate and validate the use of a computerized testing program for the diagnosis of arithmetic difficulties experienced by primary school children. The basic research question was whether a microcomputer could be used to diagnose difficulties in addition, subtraction and multiplication as well as a paper-and-pencil test can. Variables considered were convenience and ease of use, time taken, the accuracy of the diagnosis in comparison with that of an experienced remedial teacher and of a regular classroom teacher, and the usability of the information provided. An introduction, aim of evaluation, general procedure, analysis of results, formative assessment, and conclusions are provided. A trial version and revised version of the Seville Diagnostic Arithmetic Test, error scatterplots by tester, and error comparisons by levels are appended. (KR)