Treatment tests, or treatment direction tests, have been advocated for predicting the usefulness of a technique or intervention prior to its therapeutic application. The application of treatment tests however varies for different approaches: McKenzie considers the response to repeated active movements; Maitland to passive movements during physical assessment; and Mulligan, McConnell & Vicenzino to passive inputs superimposed on an active movement. The response to repeated lumbar spine movements has been shown to predict effective treatment by the McKenzie method, but there is little evidence of validity for other treatment tests. This presentation will consider the content validity of treatment tests by discussing the assumptions and rationale underlying the tests. Examples of patients with adaptive or maladaptive movement patterns will be used to clarify the appropriateness of types of treatment tests when applied to different hypotheses and different patient presentations. For example, the response to passive movement directed to a local symptomatic structure may be more useful in a patient with an adaptive movement pattern, while the response to a remote application of a treatment direction test may be more relevant in the presence of a maladaptive pattern. Pending research on the ability of treatment tests to predict outcomes, it is hoped that an understanding of the assumptions and rationale underlying the tests will assist physiotherapists in their application in the clinical setting.