Rheumatoid factor (RF) test results reported in the College of American Pathologists' surveys for 1983-1985 lacked inter-laboratory reliability and mutual validity. Using the 4 most popular commercial kits for RF testing, participating laboratories consistently identified as "positive" or "negative" all but the weakly positive samples. A wide range of titers was reported on qualitative testing, however. One popular kit using a modified sheep red blood cell agglutination technique yielded results that differed markedly from those with other kits. Investigators apparently have paid little attention to these discrepancies. In Arthritis & Rheumatism, from 1983 to 1985, over 50% of the articles that referred directly or indirectly to RFs omitted details of RF methodology. Until a reliable RF test is adopted, it is essential that such methodologic information be specified.In 1987, the major laboratory test still used to strengthen a clinical diagnosis of rheumatoid arthritis (RA), and claimed by some to have prognostic implications, is the detection of rheumatoid factor (RF) (1). The importance of the RF test is underlined by its continued inclusion in the diagnostic criteria proposed and modified by the American Rheumatism Association (2-5); yet, despite its key role, the RF test remains quite crude and lacks standardization.