“…present a very useful discussion of the merits and demerits of two-part models (modeling incarceration and sentence length separately), tobit, and Heckman two-step corrections with ordinary least squares (OLS) in addressing problems of censoring and/or selection surrounding the incarceration and sentence length decisions. Several researchers have also demonstrated the usefulness of using multinomial logistic regression to predict different types of incarceration, such as county jail vs. state prison, either in a individual-level (Holleran & Spohn, 2004), or multilevel context (Kramer & Ulmer, 2009;Wang & Mears, 2010). In addition, Britt (2009) proposes quantile regression as an interesting alternative for assessing variation in the effects of predictors of interest (legally relevant, extralegal, case processing, etc.)…”