“…For example, in the criminal justice field, they are often used to index the relationship between risk factors or risk scales with dichotomously defined recidivism (e.g., Andrews & Bonta, 2010;Olver, Wong, Nicholaichuk, & Gordon, 2007;Rice, Harris, & Lang, 2013). Similarly, meta-analyses on the prediction of recidivism or other dichotomous outcomes frequently use correlations as their effect size (Andrews et al, 1990;Bonta, Rugge, Scott, Bourgon, & Yessine, 2008;Campbell, French, & Gendreau, 2009;French & Gendreau, 2006;Gendreau, Little, & Goggin, 1996;Gonçalves, Gonçalves, Martins, & Dirkzwager, 2014). Correlations with dichotomous or ordinal data also appear as intermediate steps in other analyses, such as factor analyses of measures with dichotomous or ordinal scale items.…”