Skip to main content

Table 5 False positives from default algorithms for each de-duplication method by publication type

From: Considerations for conducting systematic reviews: evaluating the performance of different methods for de-duplicating references

  Primary research publications Non-primary research publications
  Full-text articles Conference proceedings/abstracts All
Ovid multifile search 0 0 0 0
EndNote X9a 81/208 (39%) 4/208 (2%) 85/208 (41%) 113/208 (54%)
Mendeley 3/17 (18%) 1/17 (6%) 4/17 (24%) 13/17 (76%)
Zotero 1/20 (5%) 10/20 (50%) 11/20 (55%) 9/20 (45%)
Covidence 0/2 (0%) 0/2 (0%) 0/2 (0%) 2/2 (100%)
Rayyanb 11/52 (21%) 5/52 (10%) 16/52 (31%) 34/52(65%)
  1. aUnable to retrieve full-text publication for 10 false-positive duplicate references within EndNote X9
  2. bUnable to retrieve full-text publication for 2 false-positive duplicate references within Rayyan