Comprehensive computer searches and reporting in systematic reviews


Paul Fehrmann, University Libraries, Kent State University, Kent, Ohio, USA.



Systematic reviews (SR) are a strategic resource for many who may assume that comprehensive computer searches are used to identify the studies that are used in SR. The current study assessed the reports of comprehensive computer searching in SR in psychology. Comprehensive computer search methods listed as basic in SR manuals and publications of major SR organizations (e.g., Cochrane Collaboration) were the “recommended methods” that became items on a checklist used to assess computer search reports. A methodology index search in PsycINFO identified SR in psychology that were compared to SR identified in the Cochrane Database of SR.

Checklist item frequencies supported descriptive analyses, and Mann–Whitney U-test was used to compare the PsycINFO and Cochrane SR. Two recommended computer search methods were significantly more common in Cochrane SR: truncation (z = −5.64, p < .001), controlled vocabulary (z = −5.08, p < .001 ). A third search method (Cited Reference Searching) was virtually absent (SR in psychology: 0/25; and Cochrane SR: 1/25).

Confidence in SR conclusions may be undermined when evidence of recommended or empirically-based search methods is not seen. Results and suggestions might have value for those who use, evaluate, or develop guidelines for SR; research topics are also described. Copyright © 2011 John Wiley & Sons, Ltd.