1 Matched papers per journal

Supplementary Figure S1 shows the percentage of papers per journal included in the analysis. The excluded papers are a combination of missing document types in Web of Science and missing name information. Journals publishing document types which are included in PubMed Medline but not Web of Science (e.g. comments, notes) can account for a large exclusion percentage for many journals. For other journals, first name information is consistently missing for some or all years. Here showing journals with > 50 papers only.

**Supplementary Figure S1.** Percentage of papers per journal included in the analysis.

Supplementary Figure S1. Percentage of papers per journal included in the analysis.

2 Gender determination

The online tool Gender-API was used to estimate the gender of all first-name and country pairings. This pairing is important as the gender connotations for some first names vary by language and culture. As an example, the name Kim is typically male in Danish, female in English-speaking countries, and unisex in Korean. Gender-API uses co-occurrences of names and countries on social media to provide a precision score for each assignment, which we use to calculate a probability of an author being female, f. We exclude all authors from this analysis who only have initials registered in Web of Science, or who are from a country with unreliable gender prediction (See Supplementary Figure S2).

2.1 Country sampling and bias

We calculated a reliability score for each country, by determining the precision score of the Gender-API name assignment for all authors per country. Names with precision scores >= .8 are considered reliable, and the reliability for the country is the average reliability hereof. We use the reliability distribution in Supplementary Figure S2 to heuristically set a cut-off at .9 reliability for inclusion in the analysis. The excluded countries are listed in Supplementary Table S1. For some of the East-Asian countries, the explanation for the low reliability lies in the unisex-naming culture of these countries. For other countries, the probable explanation is the absence of comprehensive social media data from these countries.

**Supplementary Figure S2.** Reliability of gender assignment per country, shown as the rank of countries.

Supplementary Figure S2. Reliability of gender assignment per country, shown as the rank of countries.

Supplementary Table S1. Excluded countries due to unreliable gender assignments from first name.
Country Reliability
Taiwan 0.29
Vietnam 0.35
China 0.38
Mongolia 0.49
Myanmar [Burma] 0.56
Singapore 0.66
South Korea 0.71
Malaysia 0.73
Cambodia 0.74
Laos 0.78
Fiji 0.79
Brunei 0.81
Sri Lanka 0.83
Swaziland 0.83
Botswana 0.83
Seychelles 0.86
Zimbabwe 0.87
Nigeria 0.88
Indonesia 0.88
Burundi 0.89
Madagascar 0.89
Zambia 0.90
Guyana 0.90

2.2 Gender representativity by year

**Supplementary Figure S3.** Proportion of papers with gender assignment for all authors. Reported as function of all sampled papers (p_pubmed) and proportion of all papers matched to Web of Science (p_wos)

Supplementary Figure S3. Proportion of papers with gender assignment for all authors. Reported as function of all sampled papers (p_pubmed) and proportion of all papers matched to Web of Science (p_wos)

3 Specialty disambiguation

3.1 Specialty algorithm

To adjust for medical specialties, papers need to be classified, preferably as one specialty per paper. Such a classification is not readily available in PubMed nor Web of Science. Darmoni et al. (Darmoni et al., 2006) designed an algorithm allowing such classifications from MeSH terms assigned to papers. The algorithm operates on a MeSH-specialty assignment table (Darmoni et al., 2006; Gehanno et al., 2011) which is available through an API at http://www.hetop.eu. For each MeSH term assigned to a paper, the corresponding specialties are counted, so that the paper is classified by the most common specialty. The counts are weighted by whether the MeSH term is a major index term (full weight) or not (half weight). The list of specialties is more detailed than most other such lists, containing a total of 124 specialties. We have summarised this list into five main specialties, based on the expert field knowledge of R.J. These assignments are also available from supplementary Table S2.

3.2 Specialty table

Supplementary Table S2. List of specialty and main specialty designation, and number of papers per specialty for the full sample.
Specialty Main specialty Number of papers
environment and public health basic science 63147
epidemiology basic science 58666
cytology basic science 53840
physiology basic science 44743
virology basic science 34987
bacteriology basic science 32552
information science basic science 13815
nutrition basic science 6058
mycology basic science 4488
toxicology basic science 4326
equipment and supplies basic science 4209
parasitology basic science 3700
pharmacy basic science 3334
economics basic science 3249
education basic science 2437
histology basic science 1867
environnement basic science 1599
organization and administration basic science 1261
evidence-based medicine basic science 891
statistics basic science 861
research basic science 845
phytotherapy basic science 655
embryology basic science 618
microbiology basic science 584
history of medicine basic science 411
law basic science 270
medical informatics basic science 265
disease transmission basic science 224
social affairs basic science 195
pharmacology basic science 162
biochemistry basic science 143
ethics basic science 127
developmental biology and medicine basic science 109
anatomy basic science 31
biomedical engineering basic science 27
molecular biology basic science 20
biophysics basic science 9
neurology hospital based 94856
psychiatry hospital based 19820
dermatology hospital based 6820
veterinary medicine hospital based 5179
addictology hospital based 4301
diagnostic imaging hospital based 2495
nursing care hospital based 1033
anesthesiology hospital based 959
emergency medicine hospital based 880
risk management hospital based 216
nuclear medicine hospital based 189
forensic medicine hospital based 146
pathology hospital based 38
oncology medical 133393
cardiology medical 92712
genetics medical 72929
allergy and immunology medical 50977
rheumatology medical 43453
drugs medical 43306
diagnosis medical 39095
gastroenterology medical 28805
therapeutics medical 24065
pulmonary disease medical 21577
endocrinology medical 18631
hematology medical 17051
infectious disease medicine medical 16778
metabolism medical 11127
geriatrics medical 7902
nephrology medical 4809
hepatology medical 3411
occupational medicine medical 2375
pain medical 1719
preventive medicine medical 1089
internal medicine medical 1080
sleep medicine specialty medical 855
military medicine medical 827
disability medical 711
palliative medicine medical 391
physical medicine and rehabilitation medical 212
ambulatory care medical 188
humanitarian medicine medical 181
critical care medical 139
family medicine medical 136
aerospace medicine medical 131
alternative medicine medical 126
prison medicine medical 89
mountain medicine medical 60
physiotherapy medical 60
venereology medical 56
travel medicine medical 54
home care medical 40
thermal medicine medical 16
homeopathy medical 13
hemobiology-blood transfusion medical 11
naval medicine medical 5
pediatrics pediatric 28814
neonatology pediatric 1209
foetology pediatric 1184
adolescent medicine pediatric 834
school medicine pediatric 62
surgery surgical/procedural 30864
ophthalmology surgical/procedural 22728
urology surgical/procedural 15972
dentistry surgical/procedural 13953
gynecology surgical/procedural 8753
traumatology surgical/procedural 6756
otolaryngology surgical/procedural 6439
vascular medicine and surgery surgical/procedural 3062
reproductive medicine surgical/procedural 3033
obstetrics surgical/procedural 2649
mastology surgical/procedural 1912
transplantation surgical/procedural 1693
sports medicine surgical/procedural 1494
podiatry surgical/procedural 452
orthopedics surgical/procedural 253
burns surgical/procedural 246
oral medicine surgical/procedural 82
periodontics surgical/procedural 60
acupuncture surgical/procedural 57
plastic and esthetic surgery surgical/procedural 46
oral surgical procedures surgical/procedural 30
thoracic and cardiovascular surgery surgical/procedural 16
neurosurgery surgical/procedural 6
orthodontics surgical/procedural 1

4 Geographical regions

4.1 Attention

The attached table is long. To view it, click the tab above.

4.2 Table of geographical regions

Supplementary Table S3. Groupings of countries by geographical region.
Country area_code
Algeria arab
Egypt arab
Jordan arab
Kuwait arab
Lebanon arab
Morocco arab
Oman arab
Qatar arab
Saudi Arabia arab
Syria arab
Tunisia arab
United Arab Emirates arab
Georgia commonwealth_is
Kazakhstan commonwealth_is
Kyrgyzstan commonwealth_is
Russia commonwealth_is
Tajikistan commonwealth_is
Uzbekistan commonwealth_is
Brunei e_asia
Cambodia e_asia
China e_asia
Indonesia e_asia
Japan e_asia
Malaysia e_asia
Mongolia e_asia
Myanmar [Burma] e_asia
Philippines e_asia
Singapore e_asia
South Korea e_asia
Taiwan e_asia
Thailand e_asia
Vietnam e_asia
Argentina lat_am
Bahamas lat_am
Bolivia lat_am
Brazil lat_am
Chile lat_am
Colombia lat_am
Costa Rica lat_am
Cuba lat_am
Dominican Republic lat_am
Ecuador lat_am
El Salvador lat_am
French Guiana lat_am
Guyana lat_am
Haiti lat_am
Honduras lat_am
Jamaica lat_am
Mexico lat_am
Nicaragua lat_am
Panama lat_am
Paraguay lat_am
Peru lat_am
Puerto Rico lat_am
Trinidad and Tobago lat_am
Uruguay lat_am
Venezuela lat_am
Canada north_am
United States north_am
Australia oceania
Fiji oceania
New Zealand oceania
Albania sce_europe
Bosnia and Herzegovina sce_europe
Bulgaria sce_europe
Croatia sce_europe
Czech Republic sce_europe
Estonia sce_europe
Hungary sce_europe
Latvia sce_europe
Lithuania sce_europe
Macedonia sce_europe
Poland sce_europe
Romania sce_europe
Serbia sce_europe
Slovakia sce_europe
Slovenia sce_europe
Turkey sce_europe
Ukraine sce_europe
Benin ss_africa
Botswana ss_africa
Burkina Faso ss_africa
Burundi ss_africa
Côte d’Ivoire ss_africa
Cameroon ss_africa
Congo ss_africa
Congo - Brazzaville ss_africa
Ethiopia ss_africa
Ghana ss_africa
Kenya ss_africa
Lesotho ss_africa
Madagascar ss_africa
Malawi ss_africa
Mali ss_africa
Mauritius ss_africa
Mozambique ss_africa
Namibia ss_africa
Niger ss_africa
Réunion ss_africa
Rwanda ss_africa
Senegal ss_africa
Seychelles ss_africa
Sierra Leone ss_africa
South Africa ss_africa
Swaziland ss_africa
Tanzania ss_africa
Togo ss_africa
Uganda ss_africa
Zambia ss_africa
Zimbabwe ss_africa
Afghanistan sw_asia
Bangladesh sw_asia
India sw_asia
Iran sw_asia
Nepal sw_asia
Pakistan sw_asia
Sri Lanka sw_asia
Austria w_europe
Belgium w_europe
Denmark w_europe
Finland w_europe
France w_europe
Germany w_europe
Greece w_europe
Iceland w_europe
Ireland w_europe
Israel w_europe
Italy w_europe
Luxembourg w_europe
Malta w_europe
Monaco w_europe
Netherlands w_europe
Norway w_europe
Portugal w_europe
Spain w_europe
Sweden w_europe
Switzerland w_europe
United Kingdom w_europe

5 Regression results

5.1 Tweedie regression results

Supplementary Table S4. Tweedie regression results
Raw parameters
Standardized parameters
Outcome Model Predictor Estimate Std. Error EE EE.LCL EE.UCL Estimate Std. Error EE EE.LCL EE.UCL
NCS Sample 1 (Intercept) -0.39 0.0020 0.68 0.68 0.68 0.03 0.0010 1.03 1.03 1.03
NCS Sample 1 case -0.02 0.0020 0.98 0.98 0.98 -0.02 0.0020 0.98 0.98 0.98
NCS Sample 1 n_authors -0.01 0.0002 0.99 0.99 0.99 -0.06 0.0018 0.94 0.94 0.94
NCS Sample 1 int_collab 0.11 0.0024 1.12 1.11 1.12 0.11 0.0024 1.12 1.11 1.12
NCS Sample 1 selfcit 0.07 0.0001 1.07 1.07 1.07 0.51 0.0009 1.66 1.65 1.66
NCS Sample 1 mncs_journal 0.26 0.0006 1.29 1.29 1.29 0.50 0.0013 1.65 1.64 1.65
NCS Sample 2 (Intercept) -0.46 0.0025 0.63 0.63 0.64 0.01 0.0012 1.01 1.01 1.01
NCS Sample 2 case -0.01 0.0025 0.99 0.98 0.99 -0.01 0.0025 0.99 0.98 0.99
NCS Sample 2 n_authors 0.00 0.0003 1.00 1.00 1.00 -0.03 0.0023 0.97 0.97 0.98
NCS Sample 2 int_collab 0.10 0.0031 1.11 1.10 1.11 0.10 0.0031 1.11 1.10 1.11
NCS Sample 2 selfcit 0.07 0.0001 1.07 1.07 1.07 0.46 0.0008 1.58 1.58 1.59
NCS Sample 2 mncs_journal 0.30 0.0009 1.35 1.34 1.35 0.57 0.0016 1.76 1.75 1.77
NCS Sample 3 (Intercept) -0.36 0.0033 0.70 0.69 0.70 0.03 0.0016 1.03 1.02 1.03
NCS Sample 3 case -0.04 0.0033 0.96 0.96 0.97 -0.04 0.0033 0.96 0.96 0.97
NCS Sample 3 n_authors 0.00 0.0004 1.00 0.99 1.00 -0.04 0.0029 0.97 0.96 0.97
NCS Sample 3 int_collab 0.13 0.0039 1.14 1.13 1.15 0.13 0.0039 1.14 1.13 1.15
NCS Sample 3 selfcit 0.05 0.0001 1.05 1.05 1.05 0.36 0.0010 1.43 1.42 1.43
NCS Sample 3 mncs_journal 0.26 0.0010 1.30 1.29 1.30 0.49 0.0019 1.63 1.62 1.64
Note:
EE : Exponentiated estimate
EE.LCL : Lower confidence limit of exponentiated estimate
EE.UCL : Upper confidence limit of exponentiated estimate

5.2 Tweedie regression on unmatched data

Supplementary Table S5. Regression results for Tweedie regressions on the full, unmatched sample, using NCS as outcome.
Outcome Model Predictor Estimate Std. Error EE EE.LCL EE.UCL
NCS F_First (Intercept) 0.00 0.0013 1.00 1.00 1.01
NCS F_First f_first -0.01 0.0018 0.99 0.98 0.99
NCS F_First n_authors -0.02 0.0016 0.98 0.98 0.98
NCS F_First int_collab 0.12 0.0022 1.13 1.13 1.14
NCS F_First selfcit 0.42 0.0005 1.52 1.51 1.52
NCS F_First mncs_journal 0.54 0.0011 1.71 1.71 1.72
NCS F_Last (Intercept) 0.00 0.0012 1.00 1.00 1.00
NCS F_Last f_last -0.01 0.0020 0.99 0.99 1.00
NCS F_Last n_authors -0.02 0.0016 0.98 0.98 0.98
NCS F_Last int_collab 0.12 0.0022 1.13 1.13 1.14
NCS F_Last selfcit 0.42 0.0005 1.52 1.51 1.52
NCS F_Last mncs_journal 0.54 0.0011 1.71 1.71 1.72
NCS F_Both (Intercept) 0.00 0.0011 1.00 1.00 1.00
NCS F_Both f_both -0.02 0.0025 0.98 0.97 0.98
NCS F_Both n_authors -0.02 0.0016 0.98 0.98 0.98
NCS F_Both int_collab 0.12 0.0022 1.13 1.13 1.14
NCS F_Both selfcit 0.42 0.0005 1.52 1.51 1.52
NCS F_Both mncs_journal 0.54 0.0011 1.71 1.71 1.72
Note:
EE : Exponentiated estimate
EE.LCL : Lower confidence limit of exponentiated estimate
EE.UCL : Upper confidence limit of exponentiated estimate

5.3 Tweedie regression on MNCS Journal quantiles

Supplementary Table S6. Tweedie regression of standardized parameters, using MNCS Journal quantiles rather than measurements.
Outcome Model Predictor Estimate Std. Error EE EE.LCL EE.UCL
NCS Sample 1 (Intercept) -0.32 0.0019 0.73 0.72 0.73
NCS Sample 1 case -0.02 0.0020 0.98 0.98 0.99
NCS Sample 1 n_authors -0.07 0.0019 0.93 0.93 0.93
NCS Sample 1 int_collab 0.08 0.0025 1.08 1.08 1.09
NCS Sample 1 selfcit 0.54 0.0009 1.71 1.71 1.72
NCS Sample 1 mncs_j_high 1.19 0.0039 3.28 3.25 3.30
NCS Sample 1 mncs_j_med 0.56 0.0021 1.75 1.74 1.75
NCS Sample 2 (Intercept) -0.34 0.0024 0.71 0.71 0.72
NCS Sample 2 case -0.01 0.0026 0.99 0.99 1.00
NCS Sample 2 n_authors -0.03 0.0024 0.97 0.96 0.97
NCS Sample 2 int_collab 0.08 0.0032 1.08 1.08 1.09
NCS Sample 2 selfcit 0.47 0.0008 1.61 1.60 1.61
NCS Sample 2 mncs_j_high 1.23 0.0051 3.41 3.37 3.44
NCS Sample 2 mncs_j_med 0.57 0.0027 1.76 1.75 1.77
NCS Sample 3 (Intercept) -0.34 0.0033 0.72 0.71 0.72
NCS Sample 3 case -0.03 0.0035 0.97 0.96 0.98
NCS Sample 3 n_authors -0.06 0.0032 0.94 0.93 0.95
NCS Sample 3 int_collab 0.09 0.0042 1.09 1.08 1.10
NCS Sample 3 selfcit 0.40 0.0011 1.50 1.49 1.50
NCS Sample 3 mncs_j_high 1.26 0.0065 3.53 3.48 3.57
NCS Sample 3 mncs_j_med 0.58 0.0037 1.78 1.77 1.79
Note:
EE : Exponentiated estimate
EE.LCL : Lower confidence limit of exponentiated estimate
EE.UCL : Upper confidence limit of exponentiated estimate
mncs_j_high : MNCS Journal scores in and above the 95th percentile
mncs_j_med: MNCS Journal scores from the 50th to the 94th percentile

5.4 Logistic regression

Supplementary Table S7. Logistic regression results
Raw parameters
Standardized parameters
Outcome Model Predictor Estimate Std. Error OR OR.LCL OR.UCL Estimate Std. Error OR OR.LCL OR.UCL
case Sample 1 (Intercept) 0.00 0.0041 1.00 0.99 1.01 0.00 0.0020 1.00 1.00 1.00
case Sample 1 n_authors 0.02 0.0006 1.02 1.01 1.02 0.12 0.0044 1.13 1.12 1.14
case Sample 1 int_collab -0.03 0.0051 0.97 0.96 0.98 -0.03 0.0051 0.97 0.96 0.98
case Sample 1 selfcit -0.02 0.0007 0.98 0.98 0.98 -0.15 0.0047 0.86 0.86 0.87
case Sample 1 mncs_journal -0.03 0.0023 0.97 0.96 0.97 -0.07 0.0044 0.94 0.93 0.94
case Sample 2 (Intercept) 0.12 0.0051 1.13 1.12 1.14 0.00 0.0025 1.00 0.99 1.00
case Sample 2 n_authors -0.01 0.0007 0.99 0.99 0.99 -0.08 0.0055 0.93 0.92 0.94
case Sample 2 int_collab -0.01 0.0064 0.99 0.98 1.01 -0.01 0.0064 0.99 0.98 1.01
case Sample 2 selfcit -0.01 0.0008 0.99 0.98 0.99 -0.10 0.0059 0.91 0.90 0.92
case Sample 2 mncs_journal -0.03 0.0029 0.97 0.96 0.98 -0.06 0.0055 0.94 0.93 0.95
case Sample 3 (Intercept) 0.14 0.0069 1.15 1.13 1.16 0.00 0.0033 1.00 0.99 1.01
case Sample 3 n_authors 0.00 0.0009 1.00 1.00 1.00 -0.02 0.0073 0.98 0.97 1.00
case Sample 3 int_collab 0.01 0.0083 1.01 0.99 1.03 0.01 0.0083 1.01 0.99 1.03
case Sample 3 selfcit -0.03 0.0012 0.97 0.97 0.97 -0.21 0.0082 0.81 0.80 0.82
case Sample 3 mncs_journal -0.06 0.0040 0.94 0.94 0.95 -0.11 0.0076 0.90 0.88 0.91
Note:
OR : Odds ratios
OR.LCL : Lower confidence limit of odds ratios
OR.UCL : Upper confidence limit of odds ratios

5.5 Negative binomial regression results

Supplementary Table S8. Regression results for the three negative binomial regressions with times cited (CS) as outcome.
Outcome Model Predictor Estimate Std. Error IRR IRR.LCL IRR.UCL
CS Sample 1 (Intercept) 1.44 0.0019 4.22 4.21 4.24
CS Sample 1 case 0.00 0.0017 1.00 0.99 1.00
CS Sample 1 n_authors 0.00 0.0002 1.00 1.00 1.00
CS Sample 1 int_collab 0.04 0.0021 1.04 1.04 1.05
CS Sample 1 selfcit 0.14 0.0002 1.15 1.15 1.15
CS Sample 1 mncs_journal 0.43 0.0009 1.53 1.53 1.54
CS Sample 2 (Intercept) 1.42 0.0024 4.13 4.11 4.15
CS Sample 2 case -0.01 0.0021 0.99 0.98 0.99
CS Sample 2 n_authors 0.00 0.0003 1.00 1.00 1.00
CS Sample 2 int_collab 0.04 0.0027 1.05 1.04 1.05
CS Sample 2 selfcit 0.14 0.0003 1.15 1.15 1.15
CS Sample 2 mncs_journal 0.44 0.0011 1.55 1.55 1.56
CS Sample 3 (Intercept) 1.41 0.0032 4.11 4.08 4.14
CS Sample 3 case -0.01 0.0028 0.99 0.98 0.99
CS Sample 3 n_authors 0.00 0.0004 1.00 1.00 1.00
CS Sample 3 int_collab 0.04 0.0035 1.04 1.03 1.05
CS Sample 3 selfcit 0.14 0.0004 1.15 1.15 1.15
CS Sample 3 mncs_journal 0.44 0.0015 1.55 1.55 1.56
Note:
IRR : Incidence rate ratios
IRR.LCL : Lower confidence limit of incidence rate ratios
IRR.UCL : Upper confidence limit of incidence rate ratios

5.6 Model fit

Supplementary Table S9. Dispersion parameters for all regressions.
Regression model Sample 1 Sample 2 Sample 3
Tweedie regression, raw params 1.002 1.010 1.003
Tweedie regression, unmatched data 1.027 1.028 1.027
Tweedie regression, MNCS Journal quantiles 1.058 1.098 1.137
Logistic regression, raw params 1.000 1.000 1.000
Negative binomial regression 1.000 1.000 1.000

6 Literature review

reference Field Study type Nation N Period Data Indicator Statistical approach Key result
(Böhm et al., 2015) Cardiology Observation GE 1905 abstracts, 366 abstracts by women 2006-2010 Authors of abstracts submitted to the annual meetings of the German Cardiac Society Journal impact factor Mann–Whitney-U-Test. Only mean values are reported as results. On average, women had published in journals with higher impact factor scores than men (women: 5.1 ± 0.2, Men: 4.4 ±, p= 0.000).
(Choi et al., 2009) Radiation oncology Observation US 826 authors, 234 women 1997-2007 Faculty at 78 US Radiation oncology departments H-index. Not specified When stratified by academic rank no notable differences were found between women’s and men’s H-indices at the assistant professor, associate professor and full professor level. On average, women department chairs had lower H-indices than men department chairs. Results for full sample, men: N=592, Mean= 9.4 (95% CI: 8.7-10.01), women: N=234, Mean= 6.4 (95% CI: 5.5-7.4). Results for assistant professors, men: N= 188, Mean= 4 (95% CI: 3.4-4.6), women: N= 102 Mean= 4 (95% CI: 3.0-4.9). Results for associate professors, men: N= 131, Mean= 9.7 (95% CI: 8.6-10.8), women: N=46, Mean= 8 (95% CI: 6.2-9.8) Results for full professors, men: N=94, Mean= 17 (95% CI: 14.7-18.8), women: N=23, Mean= 17 (95% CI: 12.3-20.1). Results for department chairs, men: N=68, Mean= 18 (95% CI: 15.7-20.8), women: N= 11, Mean= 16 (95% CI: 12.7-19.8).
(Eloy et al., 2013) 34 specialties Observation US 9,952 authors, 3,133 women 2012 25 institutions from the AMA’s Fellowship and Residency Electronic H-index Mann-Whitney U-test and Kruskall -Wallis Rank sum test. Only mean values are reported as results. Women had lower average H-indices at all academic ranks from assistant professor level to chair/chief level. Interactive database. Results for full sample, women: N=3133, Mean= 5.59, men: N=6819, Mean= 10.25, p<0.0001. Results for assistant professors, women: N=1882, Mean= 3.77, men: N=2650, Mean= 2.60, p<0.0005. Results for associate professors: women: N=721, Mean= 7.14, men: N=1525, Mean= 8.76, p<0.0005. Results for full professors, women: N= 430, Mean= 14.65, men: N=2057 Mean= 17.22, p<0.0005. Results for department chairs, women: N= 100, Mean= 11.72, men: N=587, Mean= 18.98, p<0.0005.
(Frandsen et al., 2015) Clinical research Observation DK 134 researchers, 73 women Five year period PhDs enrolled at the Institute of Clinical Research, University of Southern Denmark. Cumulative citation impact Student’s t-test In a comparison of male and female PhDs matched on sub-discipline, education, age and enrollment year, no notable average gender difference was found with respect to citation impact (women: Mean= 99.11, men: Mean= 105.95, p=0.798).
(Holliday et al., 2014) Radiation oncology Observation US 1,031 authors, 293 women 1996-2012 Faculty at 82 US academic radiation oncology departments H-index and m-index Mann-Whitney-U test On average, women had slightly lower median m-indices than men (women: 0.47, men: 0.58, p<.05). On average, women had lower H-indices than men (women: 5, men: 8, p<.05). When stratified by rank, average differences in H-indices in favor of men were shown for all ranks with the exception of the assistant professor level. When stratified by rank, no statistically significant gender differences were shown for the m-quotient.
(Housri et al., 2008) Academic surgery Observation GE 994 abstracts, 96 with women authors 2000-2004 Authors of abstracts presented at the annual meetings of the German Association for Academic Surgery (GAAS) and the Society of University Surgeons (SUS) Citation-rates per paper and journal impact factor Student’s t-test Results for SUS (N= 37 women and 300 men): No notable gender differences were observed in citation-rates per paper (women: 12.10 ± 4.47, men: 9.48 ± 0.60, p= 0.255). Gender differences in average. Journal impact factors were Gender differences in Journal Impact Factor scores were statistically insignificant (women: 3.27 ± 0.43, men: 2.67 ± 0.1, p= 0.063). The inconsequential results may be due to the small number of women included in the comparisons. Results for GAAS (N= 59 women and 590 men): Gender differences in average citation-rates per paper were statistically insignificant (women: 5.80 ± 0.98, men: 4.910 ± 0.35, p= 0.389). statistically insignificant (women: 4.741 ± 0.99, men: 3.348 ± 0.14, p= 0.063).
(Ingram, 2015) Pediatric pulmonology Observation US 85 authors, 35 women 2014-2015 10 top-ranked departments in Pediatric pulmonology H-index and m-quotient Kruskal–Wallis rank sum test Women had notably lower median H-indices than men (women: 3, men: 11, p= 0.002). Median-based gender differences for m-quotients were smaller and statistically insignificant (women: 0.41, men: 0.57, p=.09). This inconsequential result may be due to the samples used in the comparisons.
(Klimo et al., 2014) Neurosurgery Observation US/CA 312 authors, 52 women 2008-2013 Database of all neurosurgeons in North America H-index and m-quotient Not specified Women had lower average H-indices than men (women: Mean= 8, men: Mean= 14, p= 0.001). Women had slightly lower m-quotients than men (women: Mean= 0.66, men: Mean= 0.52, p: 0.013)
(Larivière et al., 2011) Health (broad) Observation CA 6,231 authors (also includes other fields) 2000-2008 University and Clinical Professors at the universities in Quebec Specialty-normalized journal impact factors and citations per paper Comparison of mean values On average, women published in slightly less prestigious journals (women: 1.17, men: 1.27) and had lower citation rates per paper (women: 1.23, men: 1.47).
(Martinez et al., 2015) Musculoskeletal tumor research Observation US 505 authors, 28 women 2013 Members of the Musculoskeletal Tumor Society H-index Multiple regression analysis In a regression analysis adjusting for academic rank and experience, author gender was a statistically insignificant (p = 0.48) predictor of the H-index (No beta coefficient provided).
(Mueller et al., 2016) Surgery Observation, cohort study US 978 faculty, 234 women 1950-2009 Full-time faculty members of surgery departments of three academic centers H-index, Cumulative citations Student’s T-test At the assistant professor level women had lower average H-indices than men (women: Mean= 8.15 (SD=6.41), men: Mean= 11.42 (SD=7.93), p= 0.002). No statistically significant gender differences in H-indices were identified for associate and full professors (numerical results and stratified sample sizes not reported). Likewise, no statistically significant differences were detected in the cumulative citation impact of women and men across the three ranks (numerical results and stratified sample sizes not reported). The inconsequential results may be explained by the small samples employed in each of the sub-group analyses.
(Mirnezami et al., 2016) Medical science Observation CA 1270 (Gender composition not specified) 2000-2012 Database of university funding in Quebec, disambiguated by gender Discipline-normalized citation rates per paper Random effect 2SLS regressions Adjusting for multiple covariates including Journal Impact Factor, age and research funding, no statistically significant gender difference was detected in average citation rates per paper. Main predictor (0=male, 1=female): β= 0.0095, p>0.05.
(Nielsen, 2016) Medical sciences (broad) Observation DK 1,714 authors, 568 women 2009 Medical researchers at Aarhus University. Data retrieved from Web of Science Self-citations, field-normalized citations per paper, Source normalized impact per paper Mann-Whitney U-test Women accrued lower average field-normalized citations per paper than men (women: Median= 0.72, men: Median= 0.87, p= 0.015). Women had lower source normalized impact per paper than men (women: Median: 1.26, men: Median: 1.17, p =0.01 )
(Okhovati et al., 2015)Okhovati et al. 2016 Epidemiology Observation IR 91 authors, 14 women 2013 Web of Science, Researchers in Iran H-index, AR-index and G-index Multivariate linear regressions Adjusting for scientific age and rank, the main predictor Gender (0= women, 1= men) was found to be an insignificant predictor of H-index scores (β= 1.36, p= 0.17) AR-index (β= 2.35, p= 0.22) and G-index (β= 0.34, p= 0.27). The inconsequential results may be explained by extremely small samples employed in the analyses.
(Pagel and Hudetz, 2011) Anesthesiology Observation US 1630 authors, 510 women 1996-2011 Faculty members from 24 US academic anesthesiology departments H-index and citations per paper Mann-Whitney U-test Women had lower average H-indices than men. Women and men had similar citation rates per paper (numerical results not reported).
(Pagel and Hudetz, 2015) Anaesthesia Education Observation US 397 authors, 82 women 2014-2015 Grant recipients of the Foundation for Anesthesia Education and Research (FAER) grant program since 1987 Citation rates, H-index Mann-Whitney U-test Women had lower average citations rates per paper than men (women: Median=18, men: Median= 23, p= 0.039). Women had lower average cumulative citation performance than men (women: Median= 327, men: Median = 827, p = 0.000). Women had lower average H-indices than men (women: Median= 10, men: Median= 14, p= 0.002).
(Paik et al., 2014) Plastic surgery Observation US 505 authors, 79 women 2012 AMA’s Fellowship and Residency Interactive Database. H-index Student’s t-test At the assistant and associate professor level, women had lower average H-indices than men (women: N= 67, Mean= 5.1, men: N=254, Mean= 6.4, p= 0.04). The number of women at the full professor and department chair level was too small for meaningful statistical comparison.
(Pashkova et al., 2013) Anesthesiology Observation US 645 authors, 198 women 2012 Faculty at 25 US anesthesiology departments H-index Mann-Whitney U-test When stratified by academic, rank no discernable difference was found between women’s and men’s H-indices at the assistant professor and associate professor level. Male full professors had notably larger average H-indices than female full professors (no numerical specifications, results are only presented in figures)
(Raj et al., 2016) Medical science Observation, cohort study US 1244 authors 1995-2012 Medical faculty from 24 medical schools H-index Negative binomial regression models In a regression adjusting for race/ethnicity, specialty, setting and years since first appointment, women’s average H-index relative to men’s was = 0.81 (95% CI =.73-.90), P<0.0001).
(Rana et al., 2013) Radiation Oncology Observation US 607 authors, 203 women 1996-2012 Domestic radiation oncology residency-training institutions H-index Simple comparison of mean and median values Women’s average H-index was 2.1 (95% CI: 1.7–2.4) and men’s was 2.7 (95% CI: 2.4–3.1).
(Susarla et al., 2015) Oral and Maxillofacial Surgeons Observation US 325 authors, 38 women ? American Association of Oral and Maxillofacial Surgeons (AAOMS) database H-index Bivariate analysis (means and SD) No notable gender differences were detected in average H-indices (Women: Mean= 6.6 ± 8.0; Men: Mean= 6.6 ± 7.6).
(Winnik et al., 2012) Cardiovascular research Observation US 590 authors, 96 women 2006 Abstracts submitted to the European Society of Cardiology Congress in 2006 Papers cited more than 10 times within 2 years after publication Logistic regression Both the gender of first authors (Male N: 217, Female N: 71) and last authors (Male N: 259, Female N: 25) were found to be insignificant predictors of producing papers with +10 citations. First authors (male = 0, female=1): Odds ratio: 1.34 (95% CI=0.066-2.73). Last authors (male= 0, female=1): Odds ratio: 0.22 (95% CI: 0.003-1.66). These inconsequential result may be explained by the small sample of women included in the analyses.

6.1 Search strategy

Databases: PubMed and Google Scholar

Years: 2006 through 2016.

Search terms (all fields):

("citation impact" OR "scientific impact" OR "scientific quality" OR "publication quality" OR "publication impact" OR "research impact" OR "citation performance" OR "citation rate*" OR "research performance" OR "scientific performance" OR "publication performance" OR citations) AND (Gender OR Sex) AND (health OR Medicine)

Inclusion criteria: +Quantitative study, + numerical specifications on gender analysis of scientific performance, citation-related indices

References

Böhm M, Papoutsis K, Gottwik M, Ukena C. 2015. Publication performance of women compared to men in German cardiology. International Journal of Cardiology 181:267–269. doi:10.1016/j.ijcard.2014.12.018

Choi M, Fuller CD, Thomas CR. 2009. Estimation of Citation-Based Scholarly Activity Among Radiation Oncology Faculty at Domestic Residency-Training Institutions: 1996-2007. International Journal of Radiation Oncology Biology Physics 74:172–178. doi:10.1016/j.ijrobp.2008.07.030

Darmoni SJ, Névéol A, Renard J-M, Gehanno J-F, Soualmia LF, Dahamna B, Thirion B. 2006. A MEDLINE categorization algorithm. BMC Medical Informatics and Decision Making 6:7–7. doi:10.1186/1472-6947-6-7

Eloy JA, Svider PF, Cherla DV, Diaz L, Kovalerchik O, Mauro KM, Baredes S, Chandrasekhar SS. 2013. Gender disparities in research productivity among 9952 academic physicians. Laryngoscope 123:1865–1875. doi:10.1002/lary.24039

Frandsen TF, Jacobsen RH, Wallin JA, Brixen K, Ousager J. 2015. Gender differences in scientific performance: A bibliometric matching analysis of Danish health sciences graduates. Journal of Informetrics 9:1007–1017. doi:10.1016/j.joi.2015.09.006

Gehanno J-F, Ladner J, Rollin L, Dahamna B, Darmoni SJ. 2011. How are the different specialties represented in the major journals in general medicine? BMC Medical Informatics and Decision Making 11. doi:10.1186/1472-6947-11-3

Holliday EB, Jagsi R, Wilson LD, Choi M, Thomas CR, Fuller CD. 2014. Gender Differences in Publication Productivity, Academic Position, Career Duration, and Funding Among U.S. Academic Radiation Oncology Faculty. Academic Medicine 89:767–773. doi:10.1097/ACM.0000000000000229

Housri N, Cheung MC, Koniaris LG, Zimmers TA. 2008. Scientific Impact of Women in Academic Surgery. Journal of Surgical Research 148:13–16. doi:10.1016/j.jss.2008.02.015

Ingram DG. 2015. Publication Measures of Scholarly Productivity: Norms for Academic Pediatric Pulmonologists. Pediatric Allergy, Immunology, and Pulmonology 28:183–186. doi:10.1089/ped.2015.0511

Klimo PJ, Venable GT, Khan NR, Taylor DR, Shepherd BA, Thompson CJ, Selden NR. 2014. Bibliometric evaluation of pediatric neurosurgery in North America. Journal of neurosurgery Pediatrics 14:695–703. doi:10.3171/2014.8.PEDS1488

Larivière V, Vignola-Gagné E, Villeneuve C, Gélinas P, Gingras Y. 2011. Sex differences in research funding, productivity and impact: an analysis of Québec university professors. Scientometrics 87:483–498. doi:10.1007/s11192-011-0369-y

Martinez M, Lopez S, Beebe K. 2015. Gender Comparison of Scholarly Production in the Musculoskeletal Tumor Society Using the Hirsch Index. Journal of Surgical Education 72:1172–1178. doi:10.1016/j.jsurg.2015.06.020

Mirnezami SR, Beaudry C, Larivière V. 2016. What determines researchers’ scientific impact? A case study of Quebec researchers. Science and Public Policy 43:262–274. doi:10.1093/scipol/scv038

Mueller CM, Gaudilliere DK, Kin C, Menorca R, Girod S. 2016. Gender disparities in scholarly productivity of US academic surgeons. Journal of Surgical Research 203:28–33. doi:10.1016/j.jss.2016.03.060

Nielsen MW. 2016. Gender inequality and research performance: moving beyond individual-meritocratic explanations of academic advancement. Studies in Higher Education 41:2044–2060. doi:10.1080/03075079.2015.1007945

Okhovati M, Bazrafshan A, Zare M, Moradzade M, Mokhtari AM. 2015. Research Performance Measures and the Moderating Role of Faculty Characteristics in Epidemiology. Global Journal of Health Science 8:72. doi:10.5539/gjhs.v8n5p72

Pagel PS, Hudetz JA. 2015. Scholarly Productivity and National Institutes of Health Funding of Foundation for Anesthesia Education and Research Grant Recipients: Insights from a bibliometric analysis. Anesthesiology 123:683–691.

Pagel PS, Hudetz JA. 2011. An analysis of scholarly productivity in United States academic anaesthesiologists by citation bibliometrics. Anaesthesia 66:873–878. doi:10.1111/j.1365-2044.2011.06860.x

Paik AM, Mady LJ, Villanueva NL, Goljo E, Svider PF, Ciminello F, Eloy JA. 2014. Research productivity and gender disparities: A look at academic plastic surgery. Journal of Surgical Education 71:593–600. doi:10.1016/j.jsurg.2014.01.010

Pashkova AA, Svider PF, Chang CY, Diaz L, Eloy JA, Eloy JD. 2013. Gender disparity among US anaesthesiologists: Are women underrepresented in academic ranks and scholarly productivity? Acta Anaesthesiologica Scandinavica 57:1058–1064. doi:10.1111/aas.12141

Raj A, Carr PL, Kaplan SE, Terrin N, Breeze JL, Freund KM. 2016. Longitudinal analysis of gender differences in academic productivity among medical faculty across 24 medical schools in the United States. Academic Medicine 91:1074–1079. doi:10.1097/ACM.0000000000001251

Rana S, Holliday EB, Jagsi R, Wilson LD, Choi M, Thomas CR, Fuller CD. 2013. Scholastic activity among radiation oncology residents at us academic institutions: A benchmark analysis. Journal of Cancer Education 28:541–546. doi:10.1007/s13187-013-0500-2

Susarla SM, Swanson EW, Lopez J, Peacock ZS, Dodson TB. 2015. Does self-citation influence quantitative measures of research productivity among academic oral and maxillofacial surgeons? Journal of Oral and Maxillofacial Surgery 73:1981e1–1981e7. doi:10.1016/j.joms.2015.05.018

Winnik S, Raptis DA, Walker JH, Hasun M, Speer T, Clavien PA, Komajda M, Bax JJ, Tendera M, Fox K, Van De Werf F, Mundow C, Lüscher TF, Ruschitzka F, Matter CM. 2012. From abstract to impact in cardiovascular research: Factors predicting publication and citation. European Heart Journal 33:3034–3045. doi:10.1093/eurheartj/ehs113