An attempt to replicate a study of the accuracy of test results of medical laboratories was unsuccessful. Limitations of the obtained data prevented the research from having satisfactory internal validity, so no formal report was published. External validity of the study was also limited because the systematic random sample of 78 licensed laboratories (of 222) in Illinois did not include hospital laboratories, which were not licensed. The independent variable was the proportion of individuals of a given certification affiliation to the total number of personnel employed in a given laboratory. The dependent variable was reported accuracy on the College of American Pathologists proficiency tests. Data were collected from the Illinois Department of Public Health. Proficiency test data, from 1986, were available for only 27 of the laboratories in the original random sample. Personnel data were from September 1987. The time lag between proficiency test and personnel data was of primary concern regarding the internal validity of the study, as it was possible that currently employed laboratory personnel were not responsible for testing performed at a given laboratory 18 months earlier. In addition, certification status on employment records was not required by the State of Illinois. Solid evidence for the accuracy of laboratory results should be derived from comparisons made at the level of the individual, since individuals actually perform tests. (Two references are included in the document.) (CML)