- Email: [email protected]
Trends in Authorship of Articles in Major Ophthalmology Journals by Gender, 2002–2014
Trends in Authorship of Articles in Major Ophthalmology Journals by Gender, 2002e2014 Michael Mimouni, MD,1 Shiri Zayit-Soudry, MD,1 Ori Segal, MD,2 Yoreh Barak, MD,1 Arie Y. Nemet, MD,2 Shiri Shulman, MD,3 Noa Geffen, MD2 Purpose: To evaluate trends in the prevalence of women authors in ophthalmology in recent years. Design: Cohort study. Participants: Authors listed in publications of 6 leading ophthalmology journals between January 2002 and December 2014. Methods: Using the PubMed search engine, we conducted an observational study of trends in gender distribution of all authors in 6 leading ophthalmology journals between January 2002 and December 2014. In multiauthored articles, the first listed author often is the lead investigator and the last author is the senior author. Therefore, the full names and positions (first, middle, or last) of all authors in every article were collected. A Google-based name identifier was used to assign the gender of authors. Main Outcome Measures: Proportion of women authors throughout the study period in all journals, general ophthalmology versus subspecialty journals, and basic science versus clinical research journals. Furthermore, we assessed the proportion of women in different authorship positions (first, middle, and last). Results: A total of 102 254 authors from 23 026 published articles were analyzed. There was a significant rise over time in the percentage of women authors, with a steeper slope for first authors than for last authors (P<0.001), although in 2014, women authors were less than the 50% mark in all categories of authorship. The rise in the percentage of women authors was similar in basic and clinical research, but was steeper for first authorship than for last authorship (P<0.001). In all 3 authorship positions (first, middle, or last), women’s contributions consistently were higher in basic research publications. The rise in the percentage of women authors was significantly steeper for general journals than for subspecialty journals (P<0.001). There was no significant rise for last authorship in subspecialty journals. In all 3 authorship positions, the proportion of women was consistently higher in general ophthalmology journals than for subspecialty journals. Conclusions: Despite an overall increase in the contribution of women to the field of ophthalmology, contributions to articles published in subspecialty ophthalmology journals and the proportion of women listed as last authors on overall articles published in ophthalmology journals are still low. Ophthalmology 2016;-:1e6 ª 2016 by the American Academy of Ophthalmology.
Despite the reduction in the disparities between men and women in the workforce that started with the important role that women played in the United States during the First and the Second World Wars,1,2 presently there are still discernible differences between the genders in the fields of science and medicine. For instance, salaries differ significantly between men and women. In 2008, among newly trained physicians in New York State, men were paid on average 17% more than women, compared with a 12.5% difference in 1999.3 The significant gender gap could not be explained by specialty choice, practice setting, work hours, or other characteristics.3 The professional careers of physicians and scientists are still influenced by gender stereotypes and their accompanying expectations,4,5 and women remain underrepresented in leadership positions in academic medicine worldwide.4,5 According to the Association of American Medical 2016 by the American Academy of Ophthalmology Published by Elsevier Inc.
Colleges, women accounted for 32% of associate professors, 20% of full professors, 14% of department chairs, and 11% of deans at United States medical schools in 20126dfar from the near gender parity observed among medical students since 1995. Scholarship is the primary consideration in decisions regarding promotion and tenure in academic medicine and is a major contributor to professional reputation.7 Often, the first author of an article is considered to be the lead investigator and the last author is considered to be the senior investigator. Although gender disparities were found to decrease in the academic world, in terms of grant funding,8 academic hiring,9 and acceptance of research at scholarly journals over the past 2 decades,7 it has been noted that, in certain fields, men predominate in the more prestigious first and last author positions of scientific publications.7 To the best of our knowledge, no http://dx.doi.org/10.1016/j.ophtha.2016.04.034 ISSN 0161-6420/16
1
Ophthalmology Volume -, Number -, Month 2016 corresponding data were reported specifically for the medical field. We therefore conducted this study to test the hypotheses that (1) in ophthalmology, women are underrepresented in the first and last authorship positions of academic publications, similar to trends well described for nonmedical scientific fields; and (2) in recent years, this apparent disparity has decreased. We studied whether these trends, if any, are similar in ophthalmology articles related to basic research and those with a clinical orientation. In addition, we examined whether there were differences in trends according to the orientation of the journal (general ophthalmology or subspecialty).
Methods Because this study did not involve the examination or treatment of patients or a review of patient records, it was exempt from review and approval by our research ethics committees.
Journal Selection Two of the authors (M.M. and N.G.) used the PubMed search engine to analyze all articles published in 6 preselected ophthalmology journals chosen for their impact factor (IF) as listed by the Journal Citation Reports (2014). The first 2 were the general clinical ophthalmology journals that allow original research with the highest IF at the time the study was conducted: Ophthalmology (IF, 6.135) and JAMA Ophthalmology (IF, 4.399). The second 2 journals were the subspecialty clinical ophthalmology journals with the highest IF: Retina (IF, 3.243) and Journal of Glaucoma (IF, 2.106). The last 2 journals were the general basic researchoriented ophthalmology journals with the highest IF: Investigative Ophthalmology and Visual Science (IF, 3.404) and Experimental Eye Research (IF, 2.709).
Study Period The study was conducted in June 2015. The publication period considered was January 1, 2002, through December 31, 2014. This period was chosen because on January 1, 2002, PubMed started to publish the full names, including first names, of all authors, which allowed us to identify the authors’ genders in most cases. We selected only those articles that also included an abstract to eliminate articles such as letters, editorials, and comments that do not constitute original research. No author or article was excluded because of the country of residence or affiliation of authors.
Author Name Retrieval A computer script was programmed (Visual Studio 2013 C#; Microsoft Corporation, Redmond, WA) that allowed retrieval of all names of all authors in every article and the position of each author in the authors’ list. The position of an author was categorized as either first author or last author. All other authors were designated as middle authors. Whenever a group of authors was defined as a collaborative study group (e.g., Diabetic Retinopathy Clinical Research Network), that group and position (if it was listed first or last) were excluded from the analyses because gender could not be assigned for a group; authors in other positions for that same article were included.
2
Assignment of Author Gender To assign the gender of each author, we used a Google-based program (Baby Name Guesser; available at: http://www.gpeters.com/names/baby-names.php) that uses Google’s database to analyze common patterns involving first names. This program was used to analyze trends in dermatology manuscript authorship over time.10 The program determines whether the name is used more commonly for a man or a woman from popular usage on the Internet and provides the ratio that a given name is used for a specific gender. For instance, “Michael” is 7.527 times more common in men than in women. For the purpose of analysis, we arbitrarily considered a specific name to be that of a man or a woman if the program provided a ratio of more than 3 to 1 (3.0). In case the ratio was less than 3.0, the specific author and his or her position were not included in the analysis; however, other authors listed on that article with a ratio of 3.0 or more remained included. Confirmation of correct gender assignment was performed for the first 100 authors by using the Internet home page of the author’s institutional affiliation and the Google search engine; the assigned gender was found to be correct in all cases. Similarly, manual assignment of 100 random unassigned authors from each journal category in the years 2002 and 2012 was performed separately (n ¼ 100 3 2 ¼ 600). The percentage of manually assigned women authors in each research category in 2002 and 2012 was clinical, 29% and 35%, respectively; basic, 33% and 40%, respectively; and subspecialty, 29% and 30%, respectively.
Statistical Analyses Data were analyzed with Minitab Software version 16 (Minitab, Inc, State College, PA). Chi-square analyses were performed to compare the percentage of women authors among the 3 categories of journals (general, basic, subspecialty) and author positions (first, middle, last). Linear regression was used to analyze the trend in proportion of women authors over time. A separate analysis was conducted for each author position and for each type of journal. To compare slopes of the regression lines, an analysis of covariance (ANCOVA) was performed. A stepwise binary regression analysis was performed to identify factors associated with the gender of the author. A P value of less than 0.05 was considered statistically significant.
Results Overall, 23 026 articles were analyzed throughout the study period, of which 10 291 (44.7%) were from clinical ophthalmology journals, 12 735 (55.3%) were from basic ophthalmology research, and 3761 (16.3%) were from subspecialty clinical journals (Table 1). Of 136 855 author names that were collected for all studies, 102 231 were included in the final analyses after excluding 34 624 (25.3%) because of uncertainty regarding author gender. The distribution of the ambiguous author names was homogeneous across all study years and ranged between 23% and 26%. The percentage of women authors was 34.7% in clinical journals, 36.8% in basic journals, and 30.6% in subspecialty journals. Figure 1 depicts the percentage of women authors (all, first, and last) throughout the study period in all 6 journals combined. There was a significant rise in the percentage of women authors over time, with a steeper slope for first authors (0.75) than for last authors (0.40) or all authors (0.52; P<0.001, ANCOVA). In 2014, women authors still represented less than 50% in all categories of authorship.
Mimouni et al
Trends in Ophthalmic Women Authors
Table 1. Distribution of Journal Issues and Articles Examined and Assignment of Author Gender by Different Categories of Selected Ophthalmology Journals, 2002e2014 Clinical Journals
Journal issues, no. Articles, no. Authors, no. Author gender unassigned, no. Author gender assigned, no. Women authors, no.
General Clinical Journals
Subspecialty Clinical Journals
Basic Science Journals*
436 6530 39 496 9163
251 3761 20 115 4889
849 12 735 77 244 20 572
30 333
15 226
56 672
11 158
4664
20 871
*When comparing general journals with subspecialty journals, basic science journals were categorized as general journals.
Figure 2AeC depicts the percentage of women authors in the general and subspecialty clinical ophthalmology journals versus basic ophthalmology research journals as all authors (Fig 2A), first authors (Fig 2B), and last authors (Fig 2C). The rise in the percentage of all women authors was similar in basic and clinical research, but was steeper for first authorship and less steep for last authorship (Fig 2B, C; P<0.0001, ANCOVA). In all 3 authorship positions, women’s contributions consistently were higher in basic research publications. To identify differences between nonsubspecialty journals and subspecialty journals, we analyzed separately the subspecialty journals (n ¼ 2) versus the rest of the journals (n ¼ 4). Figure 2DeF depicts the percentage of women authors in the general clinical and basic journals versus subspecialty journals as all authors (Fig 2D), first authors (Fig 2E), and last authors (Fig 2F). The rise in the percentage of all women authors was significantly steeper for general journals than for subspecialty journals (Fig 2D; P<0.0001, ANCOVA). The rise in percentage
of women first authors also was steeper for general journals, but there was no significant rise over the years for last authorship in subspecialty journals (Fig 2F). In all 3 figures, the percentage of women authors consistently was higher in general ophthalmology journals.
Discussion In this study, there was a significant rise in the percentage of women authors over time in original research published in ophthalmology articles, although never passing the 50% mark in any category. The rise was steeper for last authorship than for first authorship and for general journals than for subspecialty journals. The rise in women authors in the field of ophthalmology found in this study is supported by previous studies.11,12 The increased presence of women in medicine is a worldwide phenomenon, but there is lack of a women’s parallel advancement into leadership positions, despite there being no explicit barriers (i.e., the glass ceiling effect). The field of ophthalmology is experiencing an increase in the number of women physicians.13 Buys13 reported that the proportion of women ophthalmologists in Canada increased from 3.1% to 20.5% between 1970 and 2011, although this increase was significantly smaller than that reported for all other medical specialties, where the proportion increased from 7.8% to 36.8%. The reported percentage of women physicians enrolled in ophthalmology residency training programs was 37.7% to 43.3% in France, Germany, Canada, and the United States in 2009.14 The rising numbers of women in ophthalmology may explain our finding that there has been a significant rise in the percentage of women authors over time. Different authorship ordering practices exist depending on the discipline, country, and basic or clinical nature of the research subject or group.15,16 In experimental science
Figure 1. The percentage of women authors (all, first, and last) throughout the study period in all 6 journals combined.
3
Ophthalmology Volume -, Number -, Month 2016
Figure 2. A, The percentage of women authors in the general clinical and basic journals (n ¼ 4) versus subspecialty journals (n ¼ 2) in all positions throughout the study years. B, The percentage of women authors in the general clinical and basic journals (n ¼ 4) versus subspecialty journals (n ¼ 2) as first authors throughout the study years. C, The percentage of women authors in the general clinical and basic journals (n ¼ 4) versus subspecialty journals (n ¼ 2) as last authors throughout the study years. D, The percentage of women authors in the general and subspecialty clinical journals (n ¼ 4) versus basic journals (n ¼ 2) in all positions throughout the study years. E, The percentage of women authors in the general and subspecialty clinical journals (n ¼ 4) versus basic journals (n ¼ 2) as first authors throughout the study years. F, The percentage of women authors in the general and subspecialty clinical journals (n ¼ 4) versus basic journals (n ¼ 2) as last authors throughout the study years.
fields, the most widely accepted convention is that the most important positions are the first and the last, whereby the first-position authors are responsible for most of the experimental work while being supervised by the last-position (senior) author,17 who often is considered to be the research group leader.18 In this study, despite the overall rise in women authors, the rise in women last authors was lower than that of women first authors. This finding seemingly contradicts the recent findings of FrancoCardenas et al,12 who reported a 40% and 48% increase in first and last women authors over a period of 10 years. This difference from our findings may be explained by the fact that 2 of the 6 journals in the current study were subspecialty journals, as opposed to 3 general ophthalmology journals in the study by Franco-Cardenas et al.12 Supporting this explanation is that the subanalysis of subspecialty journals showed that there was no significant rise in women last authors. Shah et al11 reported similar findings to those of the current study, with a significant rise in women first authors and no such rise in women last authors throughout a decade as well. A higher increase in first authors may be explained by differences in leadership roles in ophthalmology.19e21 For instance, in India, fewer women ophthalmologists were reported to be in leadership roles or to have earned in the highest monthly income range when compared with men.19 However, findings from previous studies would contradict this theory. Danesh-Meyer et al20 reported lower monthly
4
income among women ophthalmologists, despite equal numbers of men and women being in ophthalmology leadership roles in New Zealand. In this study, there was a slower rise in the percentage of women authors in subspecialty journals than for general ophthalmology journals. Moreover, there was no significant rise in women last authors in subspecialty journals. FrancoCardenas et al12 reported a significant rise over time in first and last women authors in retina and cataract surgery subspecialties, but not in the other investigated subspecialties. The sample size of the subspecialty groups in their study was relatively small (fewer than 10 in some cases). A higher rise in general journals, as opposed to subspecialty journals, may be explained by differences in working patterns in ophthalmology.19e21 McAlister et al21 investigated the differences in lifestyle and practice patterns of men and women ophthalmologists in Canada. They reported no significant differences between genders in hospital affiliation, university appointment, or rank. However, differences in practice patterns between men and women were found, specifically with regard to surgical time, family patterns surrounding households, and responsibility for childrearing. Perhaps the surgical nature of the field of ophthalmology, combined with different working patterns, may contribute to the ongoing disparity between men and women regarding scholarship and leadership roles in this field. In this study, the basic and clinical vision research journals had a similar rise in the proportion of women
Mimouni et al
Trends in Ophthalmic Women Authors
authors. A recent research study found that, although most vision scientists at different career levels are men, the older generation had a larger percentage of men than the younger generation.22 The investigators reported that the male bias has been decreasing slowly but steadily over the past decade and that women who continue in vision science research beyond graduate school advance in their career at very similar rates as men.22 This study had several limitations. First, it covered a relatively short time span, from 2002 through 2014. An analysis of a wider time span might have led to other conclusions. However, an attempt to retrieve the full names of all authors manually would not have been feasible realistically. An additional limitation of the study is the relatively high rate of author names interpreted by the software as unassigned (25%), which may have affected our results. However, the distribution of such occurrences was even throughout the study period. Nonetheless, manual assignment of unassigned author names yielded proportions of women authors consistent with those estimated based on the automatic assignment throughout the study years. Another limitation of this study is the lack of analysis by country. A potential limitation is that PubMed citations that did not contain an abstract were excluded from this study. Unfortunately, data regarding the gender of the submitting or corresponding author were not analyzed because these data are not available on PubMed. Finally, the cutoff used for gender identification was arbitrary, and thus may have introduced additional bias, despite the manual confirmations that were performed. In conclusion, we found that, despite an overall increase in the contribution of women to the field of ophthalmology, a gap remains between the frequency of women’s contributions to articles published in subspecialty ophthalmology journals and that the proportion of women listed as last authors on overall ophthalmology publications remains low. Nonetheless, the magnitude of this gap in scientific authorship has decreased remarkably over the last decade. We expect additional progress in the future. Further research on this important subject is warranted.
References
4. 5. 6. 7. 8. 9. 10.
11. 12. 13. 14. 15.
16. 17. 18. 19. 20.
1. Greenwald MW. Women, War, and Work: The Impact of World War I on Women Workers in the United States. Ithaca: Cornell University Press; 1990:xxxiii. 309. 2. Costello J. Virtue under Fire: How World War II Changed Our Social and Sexual Attitudes. 1st ed. Boston: Little, Brown; 1985:xi. 309, 16 [plates]. 3. Lo Sasso AT, Richards MR, Chou CF, Gerber SE. The $16,819 pay gap for newly trained physicians: the unexplained trend of
21.
22.
men earning more than women. Health Aff (Millwood) 2011;30:193–201. Carnes M, Bartels CM, Kaatz A, Kolehmainen C. Why is John more likely to become department chair than Jennifer? Trans Am Clin Climatol Assoc 2015;126:197–214. Carnes M. A piece of my mind. What would Patsy Mink think? JAMA 2012;307:571–2. Rayburn WF, Schrader RM, Fullilove AM, et al. Promotion rates for assistant and associate professors in obstetrics and gynecology. Obstet Gynecol 2012;119:1023–9. West JD, Jacquet J, King MM, et al. The role of gender in scholarly authorship. PLoS One 2013;8:e66212. Boyle PJ, Smith LK, Cooper NJ, et al. Gender balance: women are funded more fairly in social science. Nature 2015;525: 181–3. Williams WM, Ceci SJ. National hiring experiments reveal 2:1 faculty preference for women on STEM tenure track. Proc Natl Acad Sci U S A 2015;112:5360–5. Feramisco JD, Leitenberger JJ, Redfern SI, et al. A gender gap in the dermatology literature? Cross-sectional analysis of manuscript authorship trends in dermatology journals during 3 decades. J Am Acad Dermatol 2009;60:63–9. Shah DN, Huang J, Ying GS, et al. Trends in female representation in published ophthalmology literature, 2000e2009. Digit J Ophthalmol 2013;19:50–5. Franco-Cardenas V, Rosenberg J, Ramirez A, et al. Decadelong profile of women in ophthalmic publications. JAMA Ophthalmol 2015;133:255–9. Buys YM. Aging and feminization of the physician workforce in Canada: comparing ophthalmologists to all other physicians. Can J Ophthalmol 2014;49:291–6. Mansour AM, Shields CL, Maalouf FC, et al. Five-decade profile of women in leadership positions at ophthalmic publications. Arch Ophthalmol 2012;130:1441–6. Costas R, Bordons M. Do age and professional rank influence the order of authorship in scientific publications? Some evidence from a micro-level perspective. Scientometrics 2011;88: 145–61. Savitz DA. What can we infer from author order in epidemiology? Am J Epidemiol 1999;149:401–3. Moed H. Bibliometric indicators reflect publication and management strategies. Scientometrics 2000;47:323–46. Beveridge C, Morris S. Order of merit. Nature 2007;448:508. Saurabh K, Sarkar K, Roy R, Majumder PD. Personal and practice profile of male and female ophthalmologists in India. Indian J Ophthalmol 2015;63:482–6. Danesh-Meyer HV, Deva NC, Ku JY, et al. Differences in practice and personal profiles between male and female ophthalmologists. Clin Experiment Ophthalmol 2007;35: 318–23. McAlister C, Jin YP, Braga-Mele R, et al. Comparison of lifestyle and practice patterns between male and female Canadian ophthalmologists. Can J Ophthalmol 2014;49: 287–90. Cooper EA, Radonjic A. Gender representation in the vision sciences: a longitudinal study. J Vis 2016;16:17.
Footnotes and Financial Disclosures Originally received: February 16, 2016. Final revision: April 16, 2016. Accepted: April 18, 2016. Available online: ---.
1
Department of Ophthalmology, Rambam Health Care Campus, Haifa, Israel.
2
Manuscript no. 2016-328.
Department of Ophthalmology, Meir Medical Center, Haifa, Israel.
3
Department of Ophthalmology, Tel Aviv Medical Center, Tel Aviv, Israel.
5
Ophthalmology Volume -, Number -, Month 2016 Financial Disclosure(s):
Obtained funding: none
The author(s) have no proprietary or commercial interest in any materials discussed in this article. Author Contributions:
Overall responsibility: Mimouni, Zayit-Soudry, Segal, Barak, Nemet, Shulman, Geffen
Conception and design: Mimouni, Geffen Analysis and interpretation: Mimouni, Zayit-Soudry, Segal, Barak, Nemet, Shulman, Geffen Data collection: Mimouni, Zayit-Soudry, Segal, Barak, Geffen
6
Abbreviations and Acronyms: ANCOVA ¼ analysis of covariance; IF ¼ impact factor. Correspondence: Michael Mimouni, MD, Department of Ophthalmology, Rambam Health Care Campus, Haifa, Israel. E-mail: [email protected].
Despite the reduction in the disparities between men and women in the workforce that started with the important role that women played in the United States during the First and the Second World Wars,1,2 presently there are still discernible differences between the genders in the fields of science and medicine. For instance, salaries differ significantly between men and women. In 2008, among newly trained physicians in New York State, men were paid on average 17% more than women, compared with a 12.5% difference in 1999.3 The significant gender gap could not be explained by specialty choice, practice setting, work hours, or other characteristics.3 The professional careers of physicians and scientists are still influenced by gender stereotypes and their accompanying expectations,4,5 and women remain underrepresented in leadership positions in academic medicine worldwide.4,5 According to the Association of American Medical 2016 by the American Academy of Ophthalmology Published by Elsevier Inc.
Colleges, women accounted for 32% of associate professors, 20% of full professors, 14% of department chairs, and 11% of deans at United States medical schools in 20126dfar from the near gender parity observed among medical students since 1995. Scholarship is the primary consideration in decisions regarding promotion and tenure in academic medicine and is a major contributor to professional reputation.7 Often, the first author of an article is considered to be the lead investigator and the last author is considered to be the senior investigator. Although gender disparities were found to decrease in the academic world, in terms of grant funding,8 academic hiring,9 and acceptance of research at scholarly journals over the past 2 decades,7 it has been noted that, in certain fields, men predominate in the more prestigious first and last author positions of scientific publications.7 To the best of our knowledge, no http://dx.doi.org/10.1016/j.ophtha.2016.04.034 ISSN 0161-6420/16
1
Ophthalmology Volume -, Number -, Month 2016 corresponding data were reported specifically for the medical field. We therefore conducted this study to test the hypotheses that (1) in ophthalmology, women are underrepresented in the first and last authorship positions of academic publications, similar to trends well described for nonmedical scientific fields; and (2) in recent years, this apparent disparity has decreased. We studied whether these trends, if any, are similar in ophthalmology articles related to basic research and those with a clinical orientation. In addition, we examined whether there were differences in trends according to the orientation of the journal (general ophthalmology or subspecialty).
Methods Because this study did not involve the examination or treatment of patients or a review of patient records, it was exempt from review and approval by our research ethics committees.
Journal Selection Two of the authors (M.M. and N.G.) used the PubMed search engine to analyze all articles published in 6 preselected ophthalmology journals chosen for their impact factor (IF) as listed by the Journal Citation Reports (2014). The first 2 were the general clinical ophthalmology journals that allow original research with the highest IF at the time the study was conducted: Ophthalmology (IF, 6.135) and JAMA Ophthalmology (IF, 4.399). The second 2 journals were the subspecialty clinical ophthalmology journals with the highest IF: Retina (IF, 3.243) and Journal of Glaucoma (IF, 2.106). The last 2 journals were the general basic researchoriented ophthalmology journals with the highest IF: Investigative Ophthalmology and Visual Science (IF, 3.404) and Experimental Eye Research (IF, 2.709).
Study Period The study was conducted in June 2015. The publication period considered was January 1, 2002, through December 31, 2014. This period was chosen because on January 1, 2002, PubMed started to publish the full names, including first names, of all authors, which allowed us to identify the authors’ genders in most cases. We selected only those articles that also included an abstract to eliminate articles such as letters, editorials, and comments that do not constitute original research. No author or article was excluded because of the country of residence or affiliation of authors.
Author Name Retrieval A computer script was programmed (Visual Studio 2013 C#; Microsoft Corporation, Redmond, WA) that allowed retrieval of all names of all authors in every article and the position of each author in the authors’ list. The position of an author was categorized as either first author or last author. All other authors were designated as middle authors. Whenever a group of authors was defined as a collaborative study group (e.g., Diabetic Retinopathy Clinical Research Network), that group and position (if it was listed first or last) were excluded from the analyses because gender could not be assigned for a group; authors in other positions for that same article were included.
2
Assignment of Author Gender To assign the gender of each author, we used a Google-based program (Baby Name Guesser; available at: http://www.gpeters.com/names/baby-names.php) that uses Google’s database to analyze common patterns involving first names. This program was used to analyze trends in dermatology manuscript authorship over time.10 The program determines whether the name is used more commonly for a man or a woman from popular usage on the Internet and provides the ratio that a given name is used for a specific gender. For instance, “Michael” is 7.527 times more common in men than in women. For the purpose of analysis, we arbitrarily considered a specific name to be that of a man or a woman if the program provided a ratio of more than 3 to 1 (3.0). In case the ratio was less than 3.0, the specific author and his or her position were not included in the analysis; however, other authors listed on that article with a ratio of 3.0 or more remained included. Confirmation of correct gender assignment was performed for the first 100 authors by using the Internet home page of the author’s institutional affiliation and the Google search engine; the assigned gender was found to be correct in all cases. Similarly, manual assignment of 100 random unassigned authors from each journal category in the years 2002 and 2012 was performed separately (n ¼ 100 3 2 ¼ 600). The percentage of manually assigned women authors in each research category in 2002 and 2012 was clinical, 29% and 35%, respectively; basic, 33% and 40%, respectively; and subspecialty, 29% and 30%, respectively.
Statistical Analyses Data were analyzed with Minitab Software version 16 (Minitab, Inc, State College, PA). Chi-square analyses were performed to compare the percentage of women authors among the 3 categories of journals (general, basic, subspecialty) and author positions (first, middle, last). Linear regression was used to analyze the trend in proportion of women authors over time. A separate analysis was conducted for each author position and for each type of journal. To compare slopes of the regression lines, an analysis of covariance (ANCOVA) was performed. A stepwise binary regression analysis was performed to identify factors associated with the gender of the author. A P value of less than 0.05 was considered statistically significant.
Results Overall, 23 026 articles were analyzed throughout the study period, of which 10 291 (44.7%) were from clinical ophthalmology journals, 12 735 (55.3%) were from basic ophthalmology research, and 3761 (16.3%) were from subspecialty clinical journals (Table 1). Of 136 855 author names that were collected for all studies, 102 231 were included in the final analyses after excluding 34 624 (25.3%) because of uncertainty regarding author gender. The distribution of the ambiguous author names was homogeneous across all study years and ranged between 23% and 26%. The percentage of women authors was 34.7% in clinical journals, 36.8% in basic journals, and 30.6% in subspecialty journals. Figure 1 depicts the percentage of women authors (all, first, and last) throughout the study period in all 6 journals combined. There was a significant rise in the percentage of women authors over time, with a steeper slope for first authors (0.75) than for last authors (0.40) or all authors (0.52; P<0.001, ANCOVA). In 2014, women authors still represented less than 50% in all categories of authorship.
Mimouni et al
Trends in Ophthalmic Women Authors
Table 1. Distribution of Journal Issues and Articles Examined and Assignment of Author Gender by Different Categories of Selected Ophthalmology Journals, 2002e2014 Clinical Journals
Journal issues, no. Articles, no. Authors, no. Author gender unassigned, no. Author gender assigned, no. Women authors, no.
General Clinical Journals
Subspecialty Clinical Journals
Basic Science Journals*
436 6530 39 496 9163
251 3761 20 115 4889
849 12 735 77 244 20 572
30 333
15 226
56 672
11 158
4664
20 871
*When comparing general journals with subspecialty journals, basic science journals were categorized as general journals.
Figure 2AeC depicts the percentage of women authors in the general and subspecialty clinical ophthalmology journals versus basic ophthalmology research journals as all authors (Fig 2A), first authors (Fig 2B), and last authors (Fig 2C). The rise in the percentage of all women authors was similar in basic and clinical research, but was steeper for first authorship and less steep for last authorship (Fig 2B, C; P<0.0001, ANCOVA). In all 3 authorship positions, women’s contributions consistently were higher in basic research publications. To identify differences between nonsubspecialty journals and subspecialty journals, we analyzed separately the subspecialty journals (n ¼ 2) versus the rest of the journals (n ¼ 4). Figure 2DeF depicts the percentage of women authors in the general clinical and basic journals versus subspecialty journals as all authors (Fig 2D), first authors (Fig 2E), and last authors (Fig 2F). The rise in the percentage of all women authors was significantly steeper for general journals than for subspecialty journals (Fig 2D; P<0.0001, ANCOVA). The rise in percentage
of women first authors also was steeper for general journals, but there was no significant rise over the years for last authorship in subspecialty journals (Fig 2F). In all 3 figures, the percentage of women authors consistently was higher in general ophthalmology journals.
Discussion In this study, there was a significant rise in the percentage of women authors over time in original research published in ophthalmology articles, although never passing the 50% mark in any category. The rise was steeper for last authorship than for first authorship and for general journals than for subspecialty journals. The rise in women authors in the field of ophthalmology found in this study is supported by previous studies.11,12 The increased presence of women in medicine is a worldwide phenomenon, but there is lack of a women’s parallel advancement into leadership positions, despite there being no explicit barriers (i.e., the glass ceiling effect). The field of ophthalmology is experiencing an increase in the number of women physicians.13 Buys13 reported that the proportion of women ophthalmologists in Canada increased from 3.1% to 20.5% between 1970 and 2011, although this increase was significantly smaller than that reported for all other medical specialties, where the proportion increased from 7.8% to 36.8%. The reported percentage of women physicians enrolled in ophthalmology residency training programs was 37.7% to 43.3% in France, Germany, Canada, and the United States in 2009.14 The rising numbers of women in ophthalmology may explain our finding that there has been a significant rise in the percentage of women authors over time. Different authorship ordering practices exist depending on the discipline, country, and basic or clinical nature of the research subject or group.15,16 In experimental science
Figure 1. The percentage of women authors (all, first, and last) throughout the study period in all 6 journals combined.
3
Ophthalmology Volume -, Number -, Month 2016
Figure 2. A, The percentage of women authors in the general clinical and basic journals (n ¼ 4) versus subspecialty journals (n ¼ 2) in all positions throughout the study years. B, The percentage of women authors in the general clinical and basic journals (n ¼ 4) versus subspecialty journals (n ¼ 2) as first authors throughout the study years. C, The percentage of women authors in the general clinical and basic journals (n ¼ 4) versus subspecialty journals (n ¼ 2) as last authors throughout the study years. D, The percentage of women authors in the general and subspecialty clinical journals (n ¼ 4) versus basic journals (n ¼ 2) in all positions throughout the study years. E, The percentage of women authors in the general and subspecialty clinical journals (n ¼ 4) versus basic journals (n ¼ 2) as first authors throughout the study years. F, The percentage of women authors in the general and subspecialty clinical journals (n ¼ 4) versus basic journals (n ¼ 2) as last authors throughout the study years.
fields, the most widely accepted convention is that the most important positions are the first and the last, whereby the first-position authors are responsible for most of the experimental work while being supervised by the last-position (senior) author,17 who often is considered to be the research group leader.18 In this study, despite the overall rise in women authors, the rise in women last authors was lower than that of women first authors. This finding seemingly contradicts the recent findings of FrancoCardenas et al,12 who reported a 40% and 48% increase in first and last women authors over a period of 10 years. This difference from our findings may be explained by the fact that 2 of the 6 journals in the current study were subspecialty journals, as opposed to 3 general ophthalmology journals in the study by Franco-Cardenas et al.12 Supporting this explanation is that the subanalysis of subspecialty journals showed that there was no significant rise in women last authors. Shah et al11 reported similar findings to those of the current study, with a significant rise in women first authors and no such rise in women last authors throughout a decade as well. A higher increase in first authors may be explained by differences in leadership roles in ophthalmology.19e21 For instance, in India, fewer women ophthalmologists were reported to be in leadership roles or to have earned in the highest monthly income range when compared with men.19 However, findings from previous studies would contradict this theory. Danesh-Meyer et al20 reported lower monthly
4
income among women ophthalmologists, despite equal numbers of men and women being in ophthalmology leadership roles in New Zealand. In this study, there was a slower rise in the percentage of women authors in subspecialty journals than for general ophthalmology journals. Moreover, there was no significant rise in women last authors in subspecialty journals. FrancoCardenas et al12 reported a significant rise over time in first and last women authors in retina and cataract surgery subspecialties, but not in the other investigated subspecialties. The sample size of the subspecialty groups in their study was relatively small (fewer than 10 in some cases). A higher rise in general journals, as opposed to subspecialty journals, may be explained by differences in working patterns in ophthalmology.19e21 McAlister et al21 investigated the differences in lifestyle and practice patterns of men and women ophthalmologists in Canada. They reported no significant differences between genders in hospital affiliation, university appointment, or rank. However, differences in practice patterns between men and women were found, specifically with regard to surgical time, family patterns surrounding households, and responsibility for childrearing. Perhaps the surgical nature of the field of ophthalmology, combined with different working patterns, may contribute to the ongoing disparity between men and women regarding scholarship and leadership roles in this field. In this study, the basic and clinical vision research journals had a similar rise in the proportion of women
Mimouni et al
Trends in Ophthalmic Women Authors
authors. A recent research study found that, although most vision scientists at different career levels are men, the older generation had a larger percentage of men than the younger generation.22 The investigators reported that the male bias has been decreasing slowly but steadily over the past decade and that women who continue in vision science research beyond graduate school advance in their career at very similar rates as men.22 This study had several limitations. First, it covered a relatively short time span, from 2002 through 2014. An analysis of a wider time span might have led to other conclusions. However, an attempt to retrieve the full names of all authors manually would not have been feasible realistically. An additional limitation of the study is the relatively high rate of author names interpreted by the software as unassigned (25%), which may have affected our results. However, the distribution of such occurrences was even throughout the study period. Nonetheless, manual assignment of unassigned author names yielded proportions of women authors consistent with those estimated based on the automatic assignment throughout the study years. Another limitation of this study is the lack of analysis by country. A potential limitation is that PubMed citations that did not contain an abstract were excluded from this study. Unfortunately, data regarding the gender of the submitting or corresponding author were not analyzed because these data are not available on PubMed. Finally, the cutoff used for gender identification was arbitrary, and thus may have introduced additional bias, despite the manual confirmations that were performed. In conclusion, we found that, despite an overall increase in the contribution of women to the field of ophthalmology, a gap remains between the frequency of women’s contributions to articles published in subspecialty ophthalmology journals and that the proportion of women listed as last authors on overall ophthalmology publications remains low. Nonetheless, the magnitude of this gap in scientific authorship has decreased remarkably over the last decade. We expect additional progress in the future. Further research on this important subject is warranted.
References
4. 5. 6. 7. 8. 9. 10.
11. 12. 13. 14. 15.
16. 17. 18. 19. 20.
1. Greenwald MW. Women, War, and Work: The Impact of World War I on Women Workers in the United States. Ithaca: Cornell University Press; 1990:xxxiii. 309. 2. Costello J. Virtue under Fire: How World War II Changed Our Social and Sexual Attitudes. 1st ed. Boston: Little, Brown; 1985:xi. 309, 16 [plates]. 3. Lo Sasso AT, Richards MR, Chou CF, Gerber SE. The $16,819 pay gap for newly trained physicians: the unexplained trend of
21.
22.
men earning more than women. Health Aff (Millwood) 2011;30:193–201. Carnes M, Bartels CM, Kaatz A, Kolehmainen C. Why is John more likely to become department chair than Jennifer? Trans Am Clin Climatol Assoc 2015;126:197–214. Carnes M. A piece of my mind. What would Patsy Mink think? JAMA 2012;307:571–2. Rayburn WF, Schrader RM, Fullilove AM, et al. Promotion rates for assistant and associate professors in obstetrics and gynecology. Obstet Gynecol 2012;119:1023–9. West JD, Jacquet J, King MM, et al. The role of gender in scholarly authorship. PLoS One 2013;8:e66212. Boyle PJ, Smith LK, Cooper NJ, et al. Gender balance: women are funded more fairly in social science. Nature 2015;525: 181–3. Williams WM, Ceci SJ. National hiring experiments reveal 2:1 faculty preference for women on STEM tenure track. Proc Natl Acad Sci U S A 2015;112:5360–5. Feramisco JD, Leitenberger JJ, Redfern SI, et al. A gender gap in the dermatology literature? Cross-sectional analysis of manuscript authorship trends in dermatology journals during 3 decades. J Am Acad Dermatol 2009;60:63–9. Shah DN, Huang J, Ying GS, et al. Trends in female representation in published ophthalmology literature, 2000e2009. Digit J Ophthalmol 2013;19:50–5. Franco-Cardenas V, Rosenberg J, Ramirez A, et al. Decadelong profile of women in ophthalmic publications. JAMA Ophthalmol 2015;133:255–9. Buys YM. Aging and feminization of the physician workforce in Canada: comparing ophthalmologists to all other physicians. Can J Ophthalmol 2014;49:291–6. Mansour AM, Shields CL, Maalouf FC, et al. Five-decade profile of women in leadership positions at ophthalmic publications. Arch Ophthalmol 2012;130:1441–6. Costas R, Bordons M. Do age and professional rank influence the order of authorship in scientific publications? Some evidence from a micro-level perspective. Scientometrics 2011;88: 145–61. Savitz DA. What can we infer from author order in epidemiology? Am J Epidemiol 1999;149:401–3. Moed H. Bibliometric indicators reflect publication and management strategies. Scientometrics 2000;47:323–46. Beveridge C, Morris S. Order of merit. Nature 2007;448:508. Saurabh K, Sarkar K, Roy R, Majumder PD. Personal and practice profile of male and female ophthalmologists in India. Indian J Ophthalmol 2015;63:482–6. Danesh-Meyer HV, Deva NC, Ku JY, et al. Differences in practice and personal profiles between male and female ophthalmologists. Clin Experiment Ophthalmol 2007;35: 318–23. McAlister C, Jin YP, Braga-Mele R, et al. Comparison of lifestyle and practice patterns between male and female Canadian ophthalmologists. Can J Ophthalmol 2014;49: 287–90. Cooper EA, Radonjic A. Gender representation in the vision sciences: a longitudinal study. J Vis 2016;16:17.
Footnotes and Financial Disclosures Originally received: February 16, 2016. Final revision: April 16, 2016. Accepted: April 18, 2016. Available online: ---.
1
Department of Ophthalmology, Rambam Health Care Campus, Haifa, Israel.
2
Manuscript no. 2016-328.
Department of Ophthalmology, Meir Medical Center, Haifa, Israel.
3
Department of Ophthalmology, Tel Aviv Medical Center, Tel Aviv, Israel.
5
Ophthalmology Volume -, Number -, Month 2016 Financial Disclosure(s):
Obtained funding: none
The author(s) have no proprietary or commercial interest in any materials discussed in this article. Author Contributions:
Overall responsibility: Mimouni, Zayit-Soudry, Segal, Barak, Nemet, Shulman, Geffen
Conception and design: Mimouni, Geffen Analysis and interpretation: Mimouni, Zayit-Soudry, Segal, Barak, Nemet, Shulman, Geffen Data collection: Mimouni, Zayit-Soudry, Segal, Barak, Geffen
6
Abbreviations and Acronyms: ANCOVA ¼ analysis of covariance; IF ¼ impact factor. Correspondence: Michael Mimouni, MD, Department of Ophthalmology, Rambam Health Care Campus, Haifa, Israel. E-mail: [email protected].