Psychological Science Can Improve Diagnostic Decisions

Assessing the Risk of Domestic Violence Reoffending: A Validation Study.

“…maximal predictive validity with minimal redundancy" among the items in the scale. This point accurately describes the logic behind the development of statistical prediction rules in general (Swets et al, 2000). Regardless, Kropp and Hart (2000: p. 109) claim that interrater reliability is more important because "if raters cannot agree on the presence of individual risk factors or the implications that can be drawn from them, there is little point in conducting risk assessments.…”

“…Swets et al (2000) have provided a thorough discussion of the history of ROC analysis and a description of the "statistical machinery" upon which it is based. This summary draws from their work.…”

Receiver operating characteristic analysis of eyewitness memory: Comparing the diagnostic accuracy of simultaneous versus sequential lineups.

“…In addition, a witness's identification decision obtained using a procedure associated with a higher diagnosticity ratio is more probative of guilt (i.e., one can be more certain that an identified suspect is, in fact, the perpetrator) compared with an identification decision obtained from a procedure associated with a lower diagnosticity ratio. However, despite these apparent indicators of diagnostic superiority, an inquiry into the nature of ROC analysis, a wellestablished technique grounded in signal-detection theory (Green & Swets, 1966;Swets, Dawes, & Monahan, 2000), reveals that a higher diagnosticity ratio does not actually identify the superior procedure. Indeed, the field of medicine long ago abandoned the use of the diagnosticity ratio (where it is usually referred to as either the likelihood ratio or the positive likelihood ratio) and has come to instead rely almost exclusively on ROC analysis (Lusted, 1971a(Lusted, , 1971bMetz, 1978).…”

Incremental and Predictive Validity of the Antisocial Process Screening Device in a Community Sample of Male and Female Ethnic Minority and Caucasian Youth

“…The AUC of the ROC graph can be taken as an index for interpreting the overall predictive accuracy of an instrument (i.e., the probability that a randomly selected youth who offended at the 6-month follow-up will have a higher APSD score than a randomly selected youth who did not offend at the 6-month follow-up). AUC values can range from 0 (perfect negative prediction), to .50 (chance prediction), to 1.0 (perfect positive prediction; Swets, Dawes, & Monahan, 2000).…”