“…Several past studies also dichotomized suicide risk as present or absent (e.g., Harris et al, 2009b;Mitchell & Ybarra, 2007;Reynders, Kerkhof, Molenberghs, & Van Audenhove, 2014). However, dichotomizing predictor variables or outcomes reduces validity by limiting the amount of information available on the latent trait and should thus be avoided (Cox et al, 2012;Markon, Chmielewski, & Miller, 2011;Royston, Altman, & Sauerbrei, 2006). Others have used continuous measures of both help seeking and suicidality, as well as including a variety of help sources, providing greater depth and validity to their findings (e.g., Deane et al, 2001;Yakunina et al, 2010).…”