CONCEPTUALIZING LEGALLY RELEVANT FACTORS UNDER GUIDELINES: A REPLY TO ULMER (original) (raw)

In our research note that appears in this issue, we presented the methodological case for including an indicator of the presumptive sentence in analyses of sentencing decisions under guidelines. We argued that this is a convenient way to estimate precisely the prescribed effects of legally relevant variables and that the failure to control for the presumptive sentence may lead to biased estimation of both legal and extralegal effects. Jeffrey Ulmer's thoughtful commentary and reanalysis of the Pennsylvania data (also found in this issue) is precisely the kind of research that we hoped our article would encourage. We are grateful to Professor Ulmer for putting our method to the test in a thorough and even-handed fashion. In this reply, we do two things. First, we offer some additional thoughts on Ulmer's attempt to deal with the model specification problem and the methodological limitations he identifies with our approach. Second, in response to his conceptual reservations, we present a theoretical argument for controlling for the presumptive sentence. We argue that this approach is empirically and conceptually superior to methods that do not do so, is consistent with the sentencing process in the context of guidelines, and is important for testing current theories of sentencing in that context.