In our never-ending struggle to evaluate scientific data, an important study component is the critique of the steps used to collect or to interpret data. If the steps or tests used to differentiate data points are appropriate, then the difference or similarities between data become apparent. If tests are not chosen or designed appropriately, subtle differences become indistinct or blurred. The more efficiently that a test can separate similar but different data points, the better the test. Obviously, the more a test lumps dissimilar data points, the less helpful the test.

The accuracy of a test addresses this issue of differentiation. Accuracy reflects the number of data points or viariates that are correctly identified by a test. An accurate test is not "fooled" by similar-but-different variates; it can differentiate positives and negatives. The Lachman test for ACL insufficiency can be used to examine the issue of test accuracy. If, for instance, the Lachman test can be performed in such a way that those knees with excessive secondary restraint laxity, hypermobility, or the upper ranges of normal anterior tibial translation can be differentiated from those with a torn ACL, then the Lachman test would be described as an accurate test for a torn ACL. If, however, the test cannot separate those knees with a torn ACL and excessive muscle tightness, intact secondary restraints, and minimal increases in anterior tibial translation from normal, loose knees, the test would not be considered accurate for this lesion.