Better laboratory evaluations of instruments and kits are required.

There are many published protocols for evaluating instruments and reagent-kit sets, and a plethora of papers describe the results of evaluations performed. It might be assumed, therefore, that this facet of clinical chemistry has no problems, but examination of evaluations of instruments published in the first half of 1984 shows flaws in some aspects of experimental design and execution, statistical analysis, and interpretation of data. We believe that clinical chemists, manufacturers and suppliers, and journal editors and referees can all contribute to improvement of this situation, and we suggest some guidelines for each of these groups.