Sex, Role Models, and Specialty Choices Among Graduates of US Medical Schools in 2006e2008 Reshma Jagsi,
MD, DPhil,
Kent A Griffith,
MS,
Rochelle A DeCastro,
MS,
Peter Ubel,
MD
Undergraduate education studies have suggested instructor sex can influence female students to pursue a discipline. We sought to evaluate a similar hypothesis in medical students. STUDY DESIGN: We obtained Association of American Medical Colleges (AAMC) data about the specialization of 2006e2008 graduates of US medical schools, the sex of their faculty and department chairs, and sex of residents in the residency programs in which they enrolled. We used logistic regression to examine associations between faculty and leadership sex and female students’ pursuit of 5 surgical specialties along with 3 nonsurgical specialties for context. We used Wilcoxon rank-sum tests to evaluate whether women entered residency programs with a higher proportion of female residents. RESULTS: In 2006e2008, US medical school graduates included 23,642 women. Women were substantially under-represented among residents in neurosurgery, orthopaedics, urology, otolaryngology, general surgery, and radiology; women constituted 47.4% of US graduates specializing in internal medicine and 74.9% in pediatrics. We found no significant associations between exposure to a female department chair and selection of that specialty and no consistent associations with the proportion of female full-time faculty. Compared with male students, female students entered residency programs in their chosen specialty that had significantly higher proportions of women residents in the year before their graduation. CONCLUSIONS: Although we did not detect consistent significant associations between exposure to potential female faculty role models and specialty choice, we observed that female students were more likely than males to enter programs with higher proportions of female residents. Sex differences in students’ specialization decisions merit additional investigation. (J Am Coll Surg 2014;218:345e352. 2014 by the American College of Surgeons)
BACKGROUND:
school in certain fields affects medical students’ specialty choices, particularly with regard to surgical specialties. A broad literature has developed on the influence of exposure to role models in education and occupational segregation by sex. Several studies focusing on undergraduate education have sought to assess the impact of instructor sex on female students’ likelihood to pursue additional courses in a particular discipline or to major in that discipline.9-11 Within the field of medicine, it has been suggested that a higher proportion of female faculty might encourage female students to pursue disciplines in which they are under-represented. For example, in one survey study, 49% of female medical students and only 3% of male medical students agreed that having more faculty of the same sex would increase their interest in a surgical career.6 In another survey, 35% of female medical students reported being discouraged from surgical careers by a lack of female role models.12 Still, we are aware of only one small study13 that directly examined the relationship between the proportion of female surgery faculty and the likelihood of a female student from that
For more than a decade, women have constituted nearly half of the medical school student body, but remain substantially under-represented in a number of key specialties, including most surgical specialties. Numerous survey studies1-6 and expert commentaries7,8 have suggested that a lack of exposure to female role models during medical Disclosure Information: Nothing to disclose. Dr Jagsi received a grant from the Burroughs Wellcome Foundation and the Alliance for Academic Internal Medicine. Dr Ubel was supported in part by a Health Policy Investigator Award from the Robert Wood Johnson Foundation. This material is based on data provided by the Association of American Medical Colleges (AAMC). The views expressed herein are those of the authors and do not necessarily reflect the position or policy of the AAMC. Received August 30, 2013; Revised November 11, 2013; Accepted November 14, 2013. From the Department of Radiation Oncology (Jagsi, DeCastro), Center for Cancer Biostatistics, School of Public Health (Griffith), Center for Bioethics and Social Science in Medicine (DeCastro), University of Michigan, Ann Arbor, MI, and Fuqua School of Business and Sanford School of Public Policy, Duke University, Durham, NC (Ubel). Correspondence address: Reshma Jagsi, MD, DPhil, Department of Radiation Oncology, University of Michigan, UHB2C490, SPC 5010, 1500 East Medical Center Dr, Ann Arbor, MI 48109-5010. email: [email protected]
ª 2014 by the American College of Surgeons Published by Elsevier Inc.
345
ISSN 1072-7515/13/$36.00 http://dx.doi.org/10.1016/j.jamcollsurg.2013.11.012
346
Jagsi et al
Sex, Role Models, and Specialty Choices
school pursuing a career in surgery. In that study, 21 of 24 students who pursued surgical specialties came from medical schools with particularly high proportions of women on the surgery faculty. To explore these issues in greater detail, we requested individual data from the Association of American Medical Colleges (AAMC) on individual medical students’ sex, medical school faculty and leadership characteristics, and subsequent residency training. We sought to assess whether female medical students are more likely to pursue residency training in under-represented specialties when they are exposed to female faculty and female leaders in those specialties. We also sought to evaluate whether women are more likely to match in residency programs at institutions that already have a higher proportion of female residents in their field of choice.
METHODS After obtaining approval from the University of Michigan Institutional Review Board, we requested data from the AAMC on the characteristics and specialization outcomes of those individuals who graduated from the 126 US fully accredited medical schools in 2006e2008. We selected these years because they were the most recent years for which postgraduate year 2 residency program information was available through the AAMC at the time of the data request in 2012. We chose to consider 3 years of data due to practical considerations of finances and feasibility of the data request, as well as the desire not to extend the study for a long period of time, during which factors influencing specialty choice might have changed. In particular, we chose the class of 2006 as our first year to ensure that the 2003 ACGME work hours regulations were fully implemented before the time when the students in our sample would have been making their specialization decisions. Specifically, in addition to sex information, we sought information from the AAMC’s Faculty Roster System to determine the sex composition of US medical schools’ full-time faculty and the sex of department chairs at each school; its Directory of American Medical Education for information on the sex of division chiefs for medical schools that lacked departments in our specialties of interest; and its GME Track database, a GME resident database and tracking system, for information on the residency programs in which the graduates enrolled, as well as the sex composition of residents in those programs. Of note, because of the relatively unique nature of the extremely small Charles Drew program at University of California, Los Angeles, graduates of this program were excluded from analysis. In addition, because San Juan Bautista graduated its first class of medical students in 2008, graduates
J Am Coll Surg
of this school could only be included for 2008. Therefore, the 2006 and 2007 graduates came from 125 medical schools, and those in 2008 came from 126. For these medical school graduates, we sought to evaluate the correlates of specialization in the 6 largest specialties in which women constitute fewer than one third of the residents (5 of which are surgical): neurosurgery, orthopaedic surgery, urology, otolaryngology, general surgery, and radiology. We also examined the likelihood of specializing in internal medicine and pediatrics for context, as we hypothesized that the proportion of female faculty or female leadership would have less effect in fields in which women are not substantially under-represented or are over-represented. Statistical analysis Statistical analyses were conducted using SAS software, version 9.3 (SAS Institute). We first evaluated the potential association between female leadership and the likelihood that a female student specialized in each specialty of interest. Among medical students at schools with a department or division in the specialty of interest, we evaluated whether the presence of a female department chair or division chief in the third year of that medical student’s training was associated with subsequent specialization in that field. We calculated the mean percentage of female medical students choosing each of the specialties, clustered by medical school. We then reported the mean percentage stratified by whether or not the medical student’s school had a female department chair or division chief. In addition, for those specialties in which at least 5 medical schools had female department chairs or division chiefs in each year, we constructed clustered (by medical school) logistic regression models to examine the magnitude of any possible association for those specialties. Next, we evaluated associations between women’s presence on faculty more generally during the student’s third year of training and the likelihood that a female student specialized in each specialty of interest. We first evaluated the distribution of the percentage of female full-time faculty, as we expected it would not be consistent across the specialties. For those specialties in which nearly all medical schools had at least some female full-time faculty in the department of interest, and the distribution approximated normality, the percentage was treated as a continuous covariate. For those specialties in which a large proportion of medical schools had no female full-time faculty at all, an indicator for having any female full-time faculty was modeled. We developed clustered (by medical school) logistic regression models for these associations. Finally, we explored whether women choosing one of the specialties of interest were more likely to be enrolled
Jagsi et al
Vol. 218, No. 3, March 2014
Table 1.
Sex, Role Models, and Specialty Choices
347
Number of Medical School Graduates in Each Specialty of Interest, by Sex Graduation year 2007, n Women Men
2006, n Variable
Total, all US medical graduates Specialty Diagnostic radiology General surgery Internal medicine Neurological surgery Orthopaedic surgery Otolaryngology Pediatrics Urology
Women
Men
7,734
8,172
7,910
227 440 1,393 23 78 72 1,332 35
599 646 1,600 131 524 184 467 133
256 380 1,407 24 82 77 1,365 37
2008, n Women
Men
8,204
7,956
8,188
630 645 1,565 128 545 189 430 141
237 457 1,406 20 103 94 1,277 51
657 647 1,504 118 543 170 436 127
Data were provided courtesy of the Association of American Medical Colleges.
in residency programs that had a higher proportion of female residents in the year before their medical student graduation. We compared the distributions by their medians, using Wilcoxon rank-sum test statistics.
RESULTS Of the 48,235 individuals who graduated from US accredited medical schools in 2006e2008, 23,642 (49.0%) were women. As shown in Table 1, women were substantially under-represented among US graduates specializing in neurosurgery, orthopaedic surgery, urology, otolaryngology, general surgery, and radiology. By contrast, women constituted 47.4% of those specializing in internal medicine and 74.9% of those specializing in pediatrics. Table 2 details the full-time faculty sex composition and female leadership in the specialties of interest at US medical schools. Very few women served as department chairs or division chiefs in neurosurgery, orthopaedic surgery, otolaryngology, or urology. In contrast to the more
even representation of women and men among US medical graduates specializing in internal medicine, women constituted less than one third of the full-time faculty, and only a small minority of those in leadership positions. Even in pediatrics, although women constituted nearly half of the full-time faculty overall, the proportion of women in these leadership positions was only 15%. The full-time faculty distribution approximated normality for the large specialties of internal medicine and pediatrics, but in neurosurgery and urology, many medical schools had no female full-time faculty at all. Table 3 reports the results of analyses evaluating the association between the presence of a female department chair or division chief during the third year of a female medical student’s training and her subsequent selection of each specialty of interest. As the table details, we found no significant associations between exposure to a female department chair and selection of that specialty. When investigating the association between female fulltime faculty and female medical students’ specialization,
Table 2. Number and Percentage of US Medical Schools with Female Department Chairs or Division Chiefs and the Percentage of Full-Time Female Faculty in the Specialties of Interest
Department
2004e2005* Female department chairs Female faculty, % n/N % Median Minemax
Diagnostic radiology 11/112 9.8 General surgery 7/124 5.6 Internal medicine 11/125 8.8 Neurological surgery 1/94 1.1 Orthopaedic surgery 1/104 1.0 Otolaryngology 1/94 1.1 Pediatrics 19/124 15.3 Urology 1/87 1.1
23.3 17.7 29.6 10.0 10.8 18.7 45.8 7.1
2005e2006* Female department chairs Female faculty, % n/N % Median Minemax
0e68.0 12/112 10.7 037.5 5/124 4.0 10.050.0 13/125 10.4 0100 1/96 1.0 033.3 1/105 1.0 057.1 1/93 1.1 17.373.8 19/124 15.3 071.4 2/89 2.2
22.8 18.1 30.2 8.7 11.1 16.7 45.8 7.1
2006e2007* Female department chairs Female faculty, % n/N % Median Minemax
0e68.2 15/114 13.2 037.9 6/125 4.8 9.145.7 17/126 13.5 050.0 1/96 1.0 033.3 0/106 0 057.1 2/93 2.2 17.373.8 20/125 16.0 071.4 3/88 3.4
24.0 18.8 30.7 9.5 11.7 20.0 46.9 9.8
0e66.7 045.5 050.0 0100 033.3 056.5 20.077.0 066.7
*We examined faculty and leadership characteristics in the academic year corresponding to the third year of medical school for the graduates in our sample.
348
Jagsi et al
J Am Coll Surg
Sex, Role Models, and Specialty Choices
Table 3. Association Between Exposure to a Female Department Chair or Division Chief and Selection of that Specialty for Female Medical Students Female medical students choosing specialty, mean % (95% CI) Female department chair Male department chair
Specialty
Diagnostic radiology General surgery Internal medicine Neurological surgery Orthopaedic surgery Otolaryngology Pediatrics Urology
3.3 4.3 18.5 0 0.6 1.7 16.8 0.5
(2.14.5) (3.25.3) (15.721.2)
2.9 5.4 17.6 0.3 1.1 1.0 17.2 0.5
(0.60.6) (1.02.4) (14.519.0) (0.30.7)
(2.63.2) (5.15.7) (16.718.5) (0.20.4) (0.91.2) (0.81.1) (16.517.9) (0.40.6)
Odds ratio*
95% CI
p Valuey
1.11 0.78 1.06
0.791.55 0.601.03 0.881.26
0.54 0.08 0.55
z
z
z
z
z
z
z
z
z
0.93
0.811.07
0.30
z
z
z
*Models account for the clustering of medical students within medical schools. The odds ratio compares female medical students that have a female department chair in that specialty with female medical students that have a male department chair. The sample sizes range from 21,650 for diagnostic radiology to 23,600 for internal medicine. y Significance of odds ratio. z Model not run because
MD, DPhil,
Kent A Griffith,
MS,
Rochelle A DeCastro,
MS,
Peter Ubel,
MD
Undergraduate education studies have suggested instructor sex can influence female students to pursue a discipline. We sought to evaluate a similar hypothesis in medical students. STUDY DESIGN: We obtained Association of American Medical Colleges (AAMC) data about the specialization of 2006e2008 graduates of US medical schools, the sex of their faculty and department chairs, and sex of residents in the residency programs in which they enrolled. We used logistic regression to examine associations between faculty and leadership sex and female students’ pursuit of 5 surgical specialties along with 3 nonsurgical specialties for context. We used Wilcoxon rank-sum tests to evaluate whether women entered residency programs with a higher proportion of female residents. RESULTS: In 2006e2008, US medical school graduates included 23,642 women. Women were substantially under-represented among residents in neurosurgery, orthopaedics, urology, otolaryngology, general surgery, and radiology; women constituted 47.4% of US graduates specializing in internal medicine and 74.9% in pediatrics. We found no significant associations between exposure to a female department chair and selection of that specialty and no consistent associations with the proportion of female full-time faculty. Compared with male students, female students entered residency programs in their chosen specialty that had significantly higher proportions of women residents in the year before their graduation. CONCLUSIONS: Although we did not detect consistent significant associations between exposure to potential female faculty role models and specialty choice, we observed that female students were more likely than males to enter programs with higher proportions of female residents. Sex differences in students’ specialization decisions merit additional investigation. (J Am Coll Surg 2014;218:345e352. 2014 by the American College of Surgeons)
BACKGROUND:
school in certain fields affects medical students’ specialty choices, particularly with regard to surgical specialties. A broad literature has developed on the influence of exposure to role models in education and occupational segregation by sex. Several studies focusing on undergraduate education have sought to assess the impact of instructor sex on female students’ likelihood to pursue additional courses in a particular discipline or to major in that discipline.9-11 Within the field of medicine, it has been suggested that a higher proportion of female faculty might encourage female students to pursue disciplines in which they are under-represented. For example, in one survey study, 49% of female medical students and only 3% of male medical students agreed that having more faculty of the same sex would increase their interest in a surgical career.6 In another survey, 35% of female medical students reported being discouraged from surgical careers by a lack of female role models.12 Still, we are aware of only one small study13 that directly examined the relationship between the proportion of female surgery faculty and the likelihood of a female student from that
For more than a decade, women have constituted nearly half of the medical school student body, but remain substantially under-represented in a number of key specialties, including most surgical specialties. Numerous survey studies1-6 and expert commentaries7,8 have suggested that a lack of exposure to female role models during medical Disclosure Information: Nothing to disclose. Dr Jagsi received a grant from the Burroughs Wellcome Foundation and the Alliance for Academic Internal Medicine. Dr Ubel was supported in part by a Health Policy Investigator Award from the Robert Wood Johnson Foundation. This material is based on data provided by the Association of American Medical Colleges (AAMC). The views expressed herein are those of the authors and do not necessarily reflect the position or policy of the AAMC. Received August 30, 2013; Revised November 11, 2013; Accepted November 14, 2013. From the Department of Radiation Oncology (Jagsi, DeCastro), Center for Cancer Biostatistics, School of Public Health (Griffith), Center for Bioethics and Social Science in Medicine (DeCastro), University of Michigan, Ann Arbor, MI, and Fuqua School of Business and Sanford School of Public Policy, Duke University, Durham, NC (Ubel). Correspondence address: Reshma Jagsi, MD, DPhil, Department of Radiation Oncology, University of Michigan, UHB2C490, SPC 5010, 1500 East Medical Center Dr, Ann Arbor, MI 48109-5010. email: [email protected]
ª 2014 by the American College of Surgeons Published by Elsevier Inc.
345
ISSN 1072-7515/13/$36.00 http://dx.doi.org/10.1016/j.jamcollsurg.2013.11.012
346
Jagsi et al
Sex, Role Models, and Specialty Choices
school pursuing a career in surgery. In that study, 21 of 24 students who pursued surgical specialties came from medical schools with particularly high proportions of women on the surgery faculty. To explore these issues in greater detail, we requested individual data from the Association of American Medical Colleges (AAMC) on individual medical students’ sex, medical school faculty and leadership characteristics, and subsequent residency training. We sought to assess whether female medical students are more likely to pursue residency training in under-represented specialties when they are exposed to female faculty and female leaders in those specialties. We also sought to evaluate whether women are more likely to match in residency programs at institutions that already have a higher proportion of female residents in their field of choice.
METHODS After obtaining approval from the University of Michigan Institutional Review Board, we requested data from the AAMC on the characteristics and specialization outcomes of those individuals who graduated from the 126 US fully accredited medical schools in 2006e2008. We selected these years because they were the most recent years for which postgraduate year 2 residency program information was available through the AAMC at the time of the data request in 2012. We chose to consider 3 years of data due to practical considerations of finances and feasibility of the data request, as well as the desire not to extend the study for a long period of time, during which factors influencing specialty choice might have changed. In particular, we chose the class of 2006 as our first year to ensure that the 2003 ACGME work hours regulations were fully implemented before the time when the students in our sample would have been making their specialization decisions. Specifically, in addition to sex information, we sought information from the AAMC’s Faculty Roster System to determine the sex composition of US medical schools’ full-time faculty and the sex of department chairs at each school; its Directory of American Medical Education for information on the sex of division chiefs for medical schools that lacked departments in our specialties of interest; and its GME Track database, a GME resident database and tracking system, for information on the residency programs in which the graduates enrolled, as well as the sex composition of residents in those programs. Of note, because of the relatively unique nature of the extremely small Charles Drew program at University of California, Los Angeles, graduates of this program were excluded from analysis. In addition, because San Juan Bautista graduated its first class of medical students in 2008, graduates
J Am Coll Surg
of this school could only be included for 2008. Therefore, the 2006 and 2007 graduates came from 125 medical schools, and those in 2008 came from 126. For these medical school graduates, we sought to evaluate the correlates of specialization in the 6 largest specialties in which women constitute fewer than one third of the residents (5 of which are surgical): neurosurgery, orthopaedic surgery, urology, otolaryngology, general surgery, and radiology. We also examined the likelihood of specializing in internal medicine and pediatrics for context, as we hypothesized that the proportion of female faculty or female leadership would have less effect in fields in which women are not substantially under-represented or are over-represented. Statistical analysis Statistical analyses were conducted using SAS software, version 9.3 (SAS Institute). We first evaluated the potential association between female leadership and the likelihood that a female student specialized in each specialty of interest. Among medical students at schools with a department or division in the specialty of interest, we evaluated whether the presence of a female department chair or division chief in the third year of that medical student’s training was associated with subsequent specialization in that field. We calculated the mean percentage of female medical students choosing each of the specialties, clustered by medical school. We then reported the mean percentage stratified by whether or not the medical student’s school had a female department chair or division chief. In addition, for those specialties in which at least 5 medical schools had female department chairs or division chiefs in each year, we constructed clustered (by medical school) logistic regression models to examine the magnitude of any possible association for those specialties. Next, we evaluated associations between women’s presence on faculty more generally during the student’s third year of training and the likelihood that a female student specialized in each specialty of interest. We first evaluated the distribution of the percentage of female full-time faculty, as we expected it would not be consistent across the specialties. For those specialties in which nearly all medical schools had at least some female full-time faculty in the department of interest, and the distribution approximated normality, the percentage was treated as a continuous covariate. For those specialties in which a large proportion of medical schools had no female full-time faculty at all, an indicator for having any female full-time faculty was modeled. We developed clustered (by medical school) logistic regression models for these associations. Finally, we explored whether women choosing one of the specialties of interest were more likely to be enrolled
Jagsi et al
Vol. 218, No. 3, March 2014
Table 1.
Sex, Role Models, and Specialty Choices
347
Number of Medical School Graduates in Each Specialty of Interest, by Sex Graduation year 2007, n Women Men
2006, n Variable
Total, all US medical graduates Specialty Diagnostic radiology General surgery Internal medicine Neurological surgery Orthopaedic surgery Otolaryngology Pediatrics Urology
Women
Men
7,734
8,172
7,910
227 440 1,393 23 78 72 1,332 35
599 646 1,600 131 524 184 467 133
256 380 1,407 24 82 77 1,365 37
2008, n Women
Men
8,204
7,956
8,188
630 645 1,565 128 545 189 430 141
237 457 1,406 20 103 94 1,277 51
657 647 1,504 118 543 170 436 127
Data were provided courtesy of the Association of American Medical Colleges.
in residency programs that had a higher proportion of female residents in the year before their medical student graduation. We compared the distributions by their medians, using Wilcoxon rank-sum test statistics.
RESULTS Of the 48,235 individuals who graduated from US accredited medical schools in 2006e2008, 23,642 (49.0%) were women. As shown in Table 1, women were substantially under-represented among US graduates specializing in neurosurgery, orthopaedic surgery, urology, otolaryngology, general surgery, and radiology. By contrast, women constituted 47.4% of those specializing in internal medicine and 74.9% of those specializing in pediatrics. Table 2 details the full-time faculty sex composition and female leadership in the specialties of interest at US medical schools. Very few women served as department chairs or division chiefs in neurosurgery, orthopaedic surgery, otolaryngology, or urology. In contrast to the more
even representation of women and men among US medical graduates specializing in internal medicine, women constituted less than one third of the full-time faculty, and only a small minority of those in leadership positions. Even in pediatrics, although women constituted nearly half of the full-time faculty overall, the proportion of women in these leadership positions was only 15%. The full-time faculty distribution approximated normality for the large specialties of internal medicine and pediatrics, but in neurosurgery and urology, many medical schools had no female full-time faculty at all. Table 3 reports the results of analyses evaluating the association between the presence of a female department chair or division chief during the third year of a female medical student’s training and her subsequent selection of each specialty of interest. As the table details, we found no significant associations between exposure to a female department chair and selection of that specialty. When investigating the association between female fulltime faculty and female medical students’ specialization,
Table 2. Number and Percentage of US Medical Schools with Female Department Chairs or Division Chiefs and the Percentage of Full-Time Female Faculty in the Specialties of Interest
Department
2004e2005* Female department chairs Female faculty, % n/N % Median Minemax
Diagnostic radiology 11/112 9.8 General surgery 7/124 5.6 Internal medicine 11/125 8.8 Neurological surgery 1/94 1.1 Orthopaedic surgery 1/104 1.0 Otolaryngology 1/94 1.1 Pediatrics 19/124 15.3 Urology 1/87 1.1
23.3 17.7 29.6 10.0 10.8 18.7 45.8 7.1
2005e2006* Female department chairs Female faculty, % n/N % Median Minemax
0e68.0 12/112 10.7 037.5 5/124 4.0 10.050.0 13/125 10.4 0100 1/96 1.0 033.3 1/105 1.0 057.1 1/93 1.1 17.373.8 19/124 15.3 071.4 2/89 2.2
22.8 18.1 30.2 8.7 11.1 16.7 45.8 7.1
2006e2007* Female department chairs Female faculty, % n/N % Median Minemax
0e68.2 15/114 13.2 037.9 6/125 4.8 9.145.7 17/126 13.5 050.0 1/96 1.0 033.3 0/106 0 057.1 2/93 2.2 17.373.8 20/125 16.0 071.4 3/88 3.4
24.0 18.8 30.7 9.5 11.7 20.0 46.9 9.8
0e66.7 045.5 050.0 0100 033.3 056.5 20.077.0 066.7
*We examined faculty and leadership characteristics in the academic year corresponding to the third year of medical school for the graduates in our sample.
348
Jagsi et al
J Am Coll Surg
Sex, Role Models, and Specialty Choices
Table 3. Association Between Exposure to a Female Department Chair or Division Chief and Selection of that Specialty for Female Medical Students Female medical students choosing specialty, mean % (95% CI) Female department chair Male department chair
Specialty
Diagnostic radiology General surgery Internal medicine Neurological surgery Orthopaedic surgery Otolaryngology Pediatrics Urology
3.3 4.3 18.5 0 0.6 1.7 16.8 0.5
(2.14.5) (3.25.3) (15.721.2)
2.9 5.4 17.6 0.3 1.1 1.0 17.2 0.5
(0.60.6) (1.02.4) (14.519.0) (0.30.7)
(2.63.2) (5.15.7) (16.718.5) (0.20.4) (0.91.2) (0.81.1) (16.517.9) (0.40.6)
Odds ratio*
95% CI
p Valuey
1.11 0.78 1.06
0.791.55 0.601.03 0.881.26
0.54 0.08 0.55
z
z
z
z
z
z
z
z
z
0.93
0.811.07
0.30
z
z
z
*Models account for the clustering of medical students within medical schools. The odds ratio compares female medical students that have a female department chair in that specialty with female medical students that have a male department chair. The sample sizes range from 21,650 for diagnostic radiology to 23,600 for internal medicine. y Significance of odds ratio. z Model not run because
