Variation in the assessment of basic clinical data gathered by emergency medical technicians and emergency department nurses was studied. Prior to testing, precise definitions, categories and procedures were developed and tolerance limits for the quantitative variables were created. Each participant evaluated four consecutive patients simultaneously with another evaluator setting the standard. The results indicate that the error rate is low to moderate for the different variables. For the quantitative variables, the error rate is over 20% and, when an error does occur, it is often very large. This indicates a need for ongoing emphasis on accurate measurement and, in cases where highly accurate data is essential, the use of multiple observers.