Citation

Erard RE. Psychol. Inj. Law 2015; 8(1): 40-45.

Copyright

(Copyright © 2015, Holtzbrinck Springer Nature Publishing Group)

DOI

10.1007/s12207-015-9217-3

PMID

unavailable

Abstract

In deciding which instrument to use in making a prediction or classification, forensic psychologists strive to use the most accurate test possible for their purpose. But accuracy in prediction or classification can be measured in many different ways. Choosing the right approach to measuring accuracy requires a basic understanding of various test accuracy statistics including, most fundamentally, sensitivity, specificity, positive predictive power, and negative predictive power. These statistics and their purposes are reviewed, along with related concepts such as base rates, cut scores, and the application of Bayes' theorem to the use of tests in particular circumstances. The advantages of using ROC (receiver operating characteristics) curves and AUC (area under the curve) statistics in choosing tests are also briefly reviewed.

Language: en