Manual Evaluation of Robot Performance in Identifying Open Access Articles

Abstract

Antelman et al. (2005) hand-tested the accuracy of the algorithm that Hajjem et al.'s (2005) software robot used to to trawl the web and automatically identify Open Access (OA) and Non-Open-Access (NOA) articles (references derived from the ISI database). Antelman et al. found much lower accuracy than Hajjem et al. Had reported. Hajjem et al. have now re-done the hand-testing on a larger sample (1000) in Biology, and demonstrated that Hajjem et al.'s original estimate of the robot's accuracy was much closer to the correct one. The discrepancy was because both Antelman et al. And Hajjem et al had hand-checked a sample other than the one the robot was sampling. Our present sample, identical with what the robot saw, yielded: d' 2.62, bias 0.68, true OA 93%, false OA 12%. We also checked whether the OA citation advantage (the ratio of the average citation counts for OA articles to the average citation counts for NOA articles in the same journal/issue) was an artifact of false OA: The robot-based OA citation Advantage of OA over NOA for this sample [(OA-NOA)/NOA x 100] was 70%. We partitioned this into the ratio of the citation counts for true (93%) OA articles to the NOA articles versus the ratio of the citation counts for the false (12%) "OA" articles. The "false OA" advantage for this 12% of the articles was 33%, so there is definitely a false OA Advantage bias component in our results. However, the true OA advantage, for 93% of the articles, was 77%. So in fact, we are underestimating the true OA advantage.