AbstractThere are several criteria to bear in mind when assessing the value of R&D spending, namely: the length of the backward horizon, whether to use cash or real terms, the intrinsic appeal of various denominators, the utility of international comparisons, the necessity of industrial disaggregation, whether or not to ignore defence spending, and the sources of finance employed. It is not possible to be dogmatic about the choice of indicators, but it is possible to indicate where, on a continuum of statistical respectability, most actual choices will lie. This article sets out the arguments for and against a variety of presentations.