Applying One Reason Decision - making : The Prioritisation of Literature Searches

Abstract = findings" Key Phrase "Key Phrase = group" Age Group "Age Group = Adulthood" Population "Population = Human" Population Location "Population Location = US" Publication Type "Publication Type = Empirical-Study" mean performance of both models averaged over the datasets. For example, once a user has read the first 60% of the prioritised articles using the one reason model, they have seen about 90% of the relevant articles. Because the models have a stochastic element, arising from breaking ties when two or more articles have equal evidence, the bestand worse-case performance is indicated by error bars (where large enough to be visible). The chance level of effort-reward performance, where each extra 10% of reading yields another 10% of the relevant articles, is shown by the thin solid line, and the performance obtained by reading the articles in the default reverse chronological order used by PsycINFO is shown by the thick broken line. The thin broken lines show the best and worst