Abstract
This study describes the byline order of gender-discordant co-first authors in 10 clinical and basic science journals with high impact factors.
The practice of acknowledging equal contributions of 2 or more co-first authors is increasingly common. However, first author position may carry the most prestige and be important for hiring and promotion. In this study, we investigated whose name is listed first when co-first authors are of different genders.
Methods
We performed a cross-sectional analysis of all original research articles with co-first authors of different genders published in 10 journals between January 1, 2005, and December 31, 2014 (Table 1). Journals with the highest impact factors were identified using the 2013 Journal Citation Reports rankings, from which the top 5 journals from the General and Internal Medicine category and the top 5 journals from 3 basic science categories (Multidisciplinary Sciences, Biochemistry and Molecular Biology, and Research and Experimental Medicine) were selected. We identified author gender using the author’s professional web page, the United States Social Security Administration’s database (for names with 95% or more of an association with a single gender), or an online database, NamepediA (used for 2% of the authors).
Table 1. Bivariable Predictors of Female-First Byline Placement for Co-first Authors of Different Genders Among 10 Journals With High Impact Factors.
| No. of Studies With a Female First in Byline/No. of Studies With Co-first Authors of Different Genders | Proportion of Studies With a Female First in Byline (95% CI) | P Valuea | |
|---|---|---|---|
| Total | 429/862 | 0.50 (0.46-0.53) | .92 |
| Basic science journals | 362/682 | 0.53 (0.49-0.57) | .12 |
| Cell | 142/259 | 0.55 | |
| JEM | 79/155 | 0.51 | |
| Nature | 41/92 | 0.45 | |
| Nature Medicine | 13/21 | 0.62 | |
| Science | 87/155 | 0.56 | |
| Clinical research journals | 67/180 | 0.37 (0.30-0.45) | <.001 |
| Annals of Internal Medicine | 8/21 | 0.38 | |
| BMJ | 8/15 | 0.53 | |
| JAMA | 4/17 | 0.24 | |
| Lancet | 8/37 | 0.22 | |
| NEJM | 39/90 | 0.43 | |
| Gender of last author | |||
| Male | 351/725 | 0.48 (0.45-0.52) | .41 |
| Female | 78/137 | 0.57 (0.48-0.65) | .12 |
| Year of publication | |||
| 2005-2009 | 232/457 | 0.51 (0.46-0.55) | .78 |
| 2010-2014 | 197/405 | 0.49 (0.44-0.54) | .62 |
| Location of corresponding author | |||
| North America | 255/498 | 0.51 (0.47-0.56) | .62 |
| Europe | 128/273 | 0.47 (0.41-0.53) | .33 |
| Other | 46/91 | 0.51 (0.40-0.61) | >.99 |
Abbreviations: JEM, Journal of Experimental Medicine; NEJM, New England Journal of Medicine.
We tested whether the proportion of women in the first author position was different from 0.50 using an exact binomial probability test and calculated a 2-sided P value and a Clopper-Pearson CI, with statistical significance set at .05.
The dependent variable was placement of the female co-first author first in the byline. Independent variables included journal genre (clinical vs basic science), last author gender, year of publication (dichotomized into 2 equal time intervals of 5 years each), geographic location of the corresponding author, and total number of authors. For each categorical variable, we tested whether the proportion of females in the first-author position differed from 0.50 using a binomial test. For the continuous variable, we performed a Wilcoxon rank-sum test.
A multivariable modified Poisson regression analysis assessed independent relationships between covariates of interest and the dependent variable. We reported associations using adjusted risk ratios (aRRs) and CIs. Analyses were performed using Stata (StataCorp), version 13.1. Statistical significance was set at a P value less than .05 (2-sided).
Results
Of the 32 829 research articles published from 2005 through 2014 among 10 journals with high impact factors, 3706 (11.3%) credited multiple authors as equal contributors. After inspecting the bylines of these articles, 2250 (6.9%) had exactly 2 co-first authors. We excluded 1260 with co-first authors of the same gender and 128 with at least 1 co-first author of uncertain gender, leaving 862 eligible articles with co-first authors of different genders.
Overall, the proportion of female authors listed first in the byline was 0.50 (95% CI, 0.46 to 0.53; P = .92). However, among articles published in clinical journals, the proportion of female authors listed first was 0.37 (95% CI, 0.30 to 0.45; P < .001). Table 1 summarizes the tests of proportions for all categorical independent variables. The median (interquartile range [IQR]) total number of authors did not differ between articles with a female author listed first (9 [IQR, 6-14]) or a male author listed first (10 [IQR, 6-14]; P = .56).
In the adjusted analysis (Table 2), compared with publication in a basic science journal, publication in a clinical research journal was inversely associated with female first authorship (aRR, 0.69 [95% CI, 0.56-0.85]; P < .001). Also, a female (vs male) last author was associated with a female co-first author listed first in the byline (aRR, 1.18 [95% CI, 1.00-1.39]; P = .04).
Table 2. Multivariable Analysis of Female-First Byline Placement for Co-first Authors of Different Genders Among 10 Journals With High Impact Factors.
| Adjusted Risk Ratio (95% CI)a | P Value | |
|---|---|---|
| Genre of journal | ||
| Basic science | 1 [Reference] | 1 [Reference] |
| Clinical medicine | 0.69 (0.56-0.85) | <.001 |
| Gender of last author | ||
| Male | 1 [Reference] | 1 [Reference] |
| Female | 1.18 (1.00-1.39) | .04 |
| Year of publication | ||
| 2005-2009 | 1 [Reference] | 1 [Reference] |
| 2010-2014 | 1.00 (0.87-1.14) | .94 |
| Location of corresponding author | ||
| North America | 1 [Reference] | 1 [Reference] |
| Europe | 0.96 (0.83-1.12) | .61 |
| Other | 0.98 (0.79-1.22) | .87 |
| Total No. of authors | 1.00 (1.00-1.01) | .23 |
The adjusted risk ratio represents the probability of a female (vs male) co-first author being listed first, controlling for the other covariates in the Table.
Discussion
Overall and among articles published in basic science journals, co-first author gender was not associated with byline position. However, female co-first authors of articles published in clinical journals were less likely than their male counterparts to be listed first in the byline. Also, female co-first authors were more likely to be listed first when the last author was a woman.
It is unclear why differences were seen between clinical and basic science journals. Possible factors that were not measured include differences in author seniority or specialty, in the proportion of the total authors who were female, or gender bias. Another limitation is the inclusion of only 10 journals with high impact factors, limiting generalizability to other journals.
This study suggests a need for investigation of what factors influence byline position among co-first authors and what professional consequences, if any, result from differences in byline position of equally contributing co-first authors.
Section Editor: Jody W. Zylke, MD, Deputy Editor.
References
- 1.Akhabue E, Lautenbach E. “Equal” contributions and credit. Ann Epidemiol. 2010;20(11):868-871. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Wuchty S, Jones BF, Uzzi B. The increasing dominance of teams in production of knowledge. Science. 2007;316(5827):1036-1039. [DOI] [PubMed] [Google Scholar]
- 3.NamepediA Name database including given and family names. http://www.namepedia.org/. Accessed October 7, 2016.
- 4.Social Security Administration Popular baby names. https://www.ssa.gov/oact/babynames/limits.html. Accessed November 18, 2016.
- 5.Zou G. A modified Poisson regression approach to prospective studies with binary data. Am J Epidemiol. 2004;159(7):702-706. [DOI] [PubMed] [Google Scholar]
