Predictive strategies in diagnostic tasks.

Abstract

A concept attainment model of diagnosis was used to study the influence of inferential ability and restricted-unrestricted information conditions on 60 nurses' hypothesis-scanning strategies, diagnostic accuracy, and confidence in two diagnostic tasks. A selection paradigm, similar to the game of Twenty Questions, was employed, and it was assumed that verbal reports of how information was being utilized were reliable indicators of scanning procedures. Analysis of variance, t tests, chi-square analyses, and the Cochrane Q test were used in data analyses. Subjects' behavior approximated a model strategy of predictive hypothesis testing in diagnostic concept attainment tasks involving surgical complications. Neither the interaction between inferential ability and information conditions nor the effect of inferential ability alone influenced the dependent variables, although prolonged predictive hypothesis testing and unrestricted information conditions were associated with greater inaccuracies. Inplications for nursing history taking and diagnosis are discussed.