Zinbarg (1998) revisits the widely accepted triple-response mode conceptualization for assessing and treating anxiety and fear. The scientific validity of these constructs has been undermined considerably by empirical findings of Lang (1978) and others that they do not exist all of a piece, and are made up of distinguishable responses broadly categorizable as cognitive, motor, and physiological. Zinbarg argues that these constructs can still be useful if viewed as unitary latent traits at a higher level, coexisting with several distinguishable response systems at lower levels of a hierarchical model. Two major difficulties with his argument are discussed: (a) methodological and (b) conceptual. The first deals with the highly complex multicontent-multimethod-multibehavior analyses needed to validate triple response mode logic. The hierarchical model appears to require the addition of multisystem comparisons, further complicating the analysis. Given that there are as yet no studies showing the separate and joint contributions of method and behavior to response system relationships, it is questionable whether even more difficult analyses will be undertaken. Conceptually, the hierarchical model looks like a return to abstract, highly inferential, medical model logic, which behavioral assessors and therapists have historically found distasteful. More parsimonious behavioral approaches are available for dealing with "anxiety" and "fear." Practitioners and researchers are encouraged to consider them before adopting a model that takes us even further from the actual behavior of clients.
