Humphry (this issue) has written a thought-provoking piece on the interpretation of item discrimination parameters as scale units in item response theory. One of the key features of his work is the description of an item response theory (IRT) model that he calls the logistic measurement function that combines aspects of two traditions in IRT that have their genesis in the measurement models of Rasch (1980) and Birnbaum (1968). This commentary briefly reviews how item discrimination parameters have been defined previously, and suggests that Humphry's model can also be usefully parameterized as person-response functions (PRFs; Perkins & Engelhard, 2009). The first section briefly examines how "units" have appeared in social science measurement and highlights the connection of units to invariant measurement. Next, the authors briefly describe how item discrimination parameters have been conceptualized. The following section stresses group response and person response functions within the context of Humphry's model. Finally, the implications for future measurement research in the social sciences are discussed. (Contains 2 figures.)