Effects of author contribution disclosures and numeric limitations on authorship trends.

OBJECTIVE To determine whether editorial policies designed to eliminate gratuitous authorship (globally referred to as authorship limitation policies), including author contribution disclosures and/or numeric restrictions, have significantly affected authorship trends during a 20-year period. METHODS We used a custom PERL-based algorithm to extract data, including number of authors, publication date, and article subtype, from articles published from January 1, 1986, through December 31, 2006, in 16 medical journals (8 with explicit authorship guidelines restricting authorship and 8 without formal authorship policies), comprising 307,190 articles. Trends in the mean number of authors per article, sorted by journal type, article subtype, and presence of authorship limitations, were determined using Sen's slope analysis and compared using analysis of variance and matched-pair analysis. Trend data were compared among the journals that had implemented 1 or both of these formal restrictive authorship policies and those that had not in order to determine their effect on authorship over time. RESULTS The number of authors per article has been increasing among all journals at a mean ± SD rate of 0.076±0.057 authors per article per year. No significant differences in authorship rate were observed between journals with and without authorship limits before enforcement (F=1.097; P=.30). After enforcement, no significant change in authorship rates was observed (matched pair: F=0.425; P=.79). CONCLUSION Implementation of authorship limitation policies does not slow the trend of increasing numbers of authors per article over time.