Miller, John Bernard MD, MPH; Schaad, Douglas C. PhD; Crittenden, Robert A. MD, MPH; Oriol, Nancy E. MD; MacLaren, Carol PhD
Each year, medical students apply for residency positions in postgraduate specialty training through the National Resident Matching Program (NRMP). The NRMP briefly outlines the appropriate extent and content of communication between programs and applicants during the residency selection process. NRMP policy states that neither applicants nor representatives of a program should “ask the other prior to the Match to make a commitment as to how each will be ranked.”1
In recent years, the communication that occurs between programs and applicants outside of interviews has raised ethical concerns. A medical school graduate from Canada was among the first to voice concerns about the potential negative effects of the residency selection process. In her commentary, she concluded that “although truthfulness and honesty have long been considered fundamental values within the medical profession, lying and deception have become standard practices within medicine's residency selection process.”2
A recent survey of family practice residency program directors reported that 85% of programs engaged in communication with applicants following the formal interview.3 The primary purpose of this communication is to recruit. Whether this represents an increase from past years is unclear. Among applicants to a competitive family medicine program in the early 1980s, 68% reported receiving feedback on their rank status.4
NRMP policy states that “references to how each (programs and applicants) will rank the other should be avoided and should not be solicited.”1 Nevertheless, many programs ask applicants to communicate their levels of interest at some point during or after the interview day. Carek et al.2 reported that 82% of family practice residency programs told at least some of their applicants to “keep in touch if they are interested in matching” with their program. In a study of the urology Match, 67% of programs reported telling applicants to “keep in touch” with them if they were interested in matching in their program.5 This use of demonstrated interest as a selection factor places considerable pressure on students to offer assurances to programs.
Residency programs further compel applicants to express their levels of interest when they recruit and communicate their levels of interest in applicants. As many as 43% of all students surveyed believed they had received an “informal commitment” from at least one program, mostly from program directors.6 When such informal commitments are given, students may feel compelled to provide assurances to programs, even when they are not explicitly asked to do so.
Many applicants believe that they must communicate an interest in the programs they would like to attend. Failure to do so may jeopardize an applicant's position on a program's rank list that was earned through the strength of applications and interviews. Anderson and colleagues6 found that 55% of students communicated with select programs, while 39% communicated with every program at which they interviewed. Students provided more false assurances to programs they planned to rank less highly. Students misled programs by completing audition electives and second visits at backup programs, and requested special efforts for spousal hires, only to go elsewhere.6
Our study explored the extent and content of communication exchanged informally outside of interviews. It also assessed students' attitudes about the frequency, efficacy, and appropriateness of making misleading statements during the Match process.
In March and April 2001, we made a secure, anonymous questionnaire available online to 1,362 medical students who were graduating from ten U.S. medical schools and who participated in the 2001 Match. Participating schools were Creighton University School of Medicine, Omaha, Nebraska; Harvard Medical School, Boston, Massachusetts; Louisiana State University School of Medicine in New Orleans; New York Medical College, Valhalla, New York; University of Chicago Pritzker School of Medicine, Chicago, Illinois; University of Kentucky College of Medicine, Lexington, Kentucky; University of Rochester School of Medicine and Dentistry, Rochester, New York; Keck School of Medicine of the University of Southern California, Los Angeles, California; University of Washington School of Medicine, Seattle, Washington; and Vanderbilt University School of Medicine, Nashville, Tennessee. We chose these schools non-randomly to obtain a geographically representative sample with both public and private schools, and a broad range of medical schools as ranked by U.S. News & World Report.7
Match participants who had graduated by the time of the Match were not included in the sample, because we believed that their experiences in the Match might be different from those of graduating medical students. The Human Subjects Committee at the University of Washington approved our study.
We administered the questionnaire on the World Wide Web in March and April 2001, just after the Match. Announcements about the questionnaire, including two follow-up requests, were sent to the class e-mail list server at each school. A login and password were provided and were required to access the secure online questionnaire. Aside from response rates, all school-specific information was confidential. An eleventh school was dropped from our study after announcements were not sent to the class in the time frame we outlined.
We collected demographic information, including gender, underrepresented minority status, and specialty. Students were surveyed regarding the communication they had initiated with programs. Closed-ended questions asked about the number of audition rotations, applications, interviews, thank-you letters sent, additional expressions of interest, and additional visits to programs, a.k.a. “second looks.” Respondents were then asked about communication initiated by programs. Closed-ended questions asked about the proportion of programs initiating communication such as thank-you letters and contact beyond thank-you letters. Students also were asked to indicate the themes of those contacts and whether they had offered any misleading statements or assurances to programs. The extent to which students had felt pressured to offer assurances to programs about their level of interest was elicited using a Likert scale. The students were then asked to outline their levels of agreement to statements about Match communication using a Likert scale. They were asked to comment in an open-ended format if they felt they had been misled, whether honest communication had affected their positions in the Match, or if they had been victims of Match rules violations.
We compared respondents and non-respondents on the basis of gender and specialty using the chi-square method. The data were analyzed by school, region, public and private affiliation, institutional ranking,7 specialty, primary care or non-primary care specialty, and specialty fill rate.8 Associations between variables were determined by correlation and analysis of variance (ANOVA) methods as appropriate.
Response rates by school are shown in Table 1 with comparisons of respondents with non-respondents from each school. Overall, 740 graduating medical students responded to the questionnaire, yielding a response rate of 54.3%. Almost half (349, or 47.2%) of the respondents were women. Almost 10% of respondents (71, or 9.6%) identified themselves as an underrepresented minority. In comparing non-respondents with respondents, women were significantly more likely to respond to the questionnaire than men (p = .008). Applicants to primary care and non-primary care specialties were equally likely to respond.
Overall, 62.3% of the respondents matched at the programs they ranked first and 16.2% matched at programs they ranked second. Fourteen percent matched at programs they ranked third to fifth, and 3.8% matched at those they ranked sixth or lower. Six respondents reported scrambling, seven did not match for unspecified reasons, and three reached agreements outside of the match. Nationwide in the 2001 Match, 60.5% matched at programs they ranked first and 16.5% matched at the programs they ranked second.8
There was no significant difference in applicant rank orders on the basis of medical school, either by the school's ranking, region, or public or private affiliation. Nearly equal percentages of respondents from top 20 schools (63.8%) and from lower-ranked schools (62.3%) matched at programs they ranked first. There was also no difference on the basis of the applicant's minority status or gender. Rank orders varied significantly among different specialties (p = .004), with applicants to primary care specialties (p = .002) and specialties with lower fill rates (p = .000) matching at programs they ranked higher.
Communication between Applicants and Programs
Table 2 outlines the associations between various factors and measures of communication in the Match. These are summarized below:
The frequency of applicant-initiated communication by each school's region, ranking, and public or private affiliation is summarized in Table 3. Applicants varied in the numbers of programs at which they did audition rotations, completed applications, interviewed, sent thank-you letters, expressed additional interest, and visited for a “second look.”
Applicants varied by medical school in the amounts of communication they initiated (p = .000). Applicants from schools not ranked among the top 20 by U.S. & News World Report initiated more communication than did their counterparts (p = .000). No significant difference was found on the basis of the school's region, or public or private affiliation, or on the basis of the applicant's gender or minority status.
Applicants to different specialties varied in the amounts of applicant-initiated communication (p = .000), with applicants to non-primary care specialties (p = .000) and to specialties with higher fill rates (p = .000) initiating more communication than their counterparts. Applicants initiating more communication went significantly farther down their lists than did those who initiated less contact (p = .000, R = −.202).
The frequencies of program initiated communication are summarized in Table 4. Themes of discussions with residents or faculty members were identified by 448 of the 740 respondents. The most common theme was encouragement to rank the program highly, noted by 79.5% of respondents. Secondly, 41.7% were given guarantees that they would be highly ranked by programs. Additionally, 23.0% were asked how highly they planned to rank the program; 21.7% reported being told that applicants' level of interest would affect their rankings; and 17.0% were asked which programs they planned to rank number one. Lastly, 11.6% were offered positions outside the Match.
The proportions of programs contacting applicants varied by medical school, (p = .001), the school's rank (p = .006), and region (p = .046). There was no significant difference on the basis of the school's public or private affiliation, or the applicant's gender or minority status.
Applicants to programs in different specialties were contacted by a significantly different proportion of programs (p = .000). Applicants to primary care specialties (p = .000) and specialties with lower fill rates (p = .000) were contacted by a larger proportion of programs than were their counterparts.
Applicants who initiated more communication with programs through audition rotations, completed applications, interviews, thank-you letters, additional expressions of interest, and visits for a “second look” were contacted by a smaller proportion of programs than were those initiating less contact (p = .000, R = −.185). Applicants contacted by a higher proportion of programs matched at rank orders significantly higher than their counterparts (p = .000, R = .214)
Pressure to offer assurances
Respondents contacted beyond the thank-you letter reported the extents to which they felt pressured to offer assurances to programs, with 55 (7.4%) feeling very pressured, 117 (15.8%) moderately pressured, 158 (21.3%) somewhat pressured, and 282 (31.8%) not at all pressured. A total of 335 respondents identified the modes of communication by which they felt most pressured: 35.5% thought phone contact posed the most pressure; 15.8% felt the mail posed the most pressure; 10.4% named e-mail; and 38.2% thought there was no difference between the modes of communication. Applicants differed by school in the amounts of pressure they felt to offer assurances (p = .001), but there was no significant difference on the basis of the medical school's rank, region, or public or private affiliation, or the applicant's minority status. Women reported feeling more pressure to offer assurances to programs than did men (p = .013)
Applicants differed by specialty in the amounts of pressure felt (p = .003), with applicants to non-primary care specialties (p = .003) and specialties with higher fill rates (p = .014) feeling more pressured. Applicants to surgical subspecialties, general surgery, obstetrics and gynecology, and emergency medicine (in descending order) felt the most pressure to offer assurances to programs, while applicants in combined medicine and pediatrics, pediatrics, and psychiatry felt the least pressure.
Higher levels of communication, whether initiated by applicant's (p = .028, R = .081) or programs (p = .000, R = .181), were associated with applicants' feeling more pressure to offer assurances about their levels of interest.
Making misleading statements
Overall, 12.4% of the 740 respondents admitted making misleading statements or assurances to programs. Applicants from schools not ranked in the top 20 reported making more misleading statements than did those from higher-ranked schools (p = .010). There was no difference by individual medical school, the school's region or public or private affiliation, or the applicant's gender or specialty. Thirteen percent of non-minorities and 5.6% of underrepresented minorities made misleading statements (p = .047).
Applicants who had initiated more communication (p = .003) and applicants who had felt more pressure to make assurances to programs (p = .000) were more likely to make misleading statements (p = .003). There was a trend for applicants who made misleading statements to match at programs they ranked higher (p = .123). The average rank order for those making misleading statements was 1.66, compared with 2.00 for those who did not make misleading statements.
Attitudes about Communication during the Match
Respondents indicated their degrees of agreement or disagreement with three statements about communication between applicants and programs during the Match. Table 5 summarizes their responses.
Misleading statements are often made
Agreement with the statement, “Applicants often make dishonest or misleading assurances or statements to programs about their level of interest” varied by medical school (p = .003), and the school's region (p = .003) and rank, with applicants from schools not ranked among the top 20 schools agreeing more strongly than applicants from higher-ranked programs (p = .005). There were no significant differences on the basis of the school's public or private affiliation, or on the applicant's gender, minority status, and specialty. Overall, applicants to specialties with high fill rates indicated significantly stronger agreement than their counterparts (p = .015).
Applicants initiating more communication agreed more strongly than their counterparts that misleading statements are often made (p = .050, R = .095). There was a trend for applicants contacted by a higher proportion of programs to agree less strongly (p = .062, R = −.069). Applicants feeling more pressure to offer assurances (p = .000, R = .230) and those making misleading statements (p = .000) indicated stronger agreement that misleading statements are often made.
Misleading statements improve an applicant's position in the Match
Agreement with the statement, “Applicants who mislead programs about how strongly they plan to rank them improve their position in the Match” varied significantly by medical school (p = .014) and the school's region (p = .036). There was no significant difference on the basis of the medical school's rank or public or private affiliation, or on the basis of the applicant's gender, minority status, or specialty. Applicants matching at a lower rank order more strongly agreed that misleading statements were beneficial (p = .003, R = −109).
Applicants initiating more communication indicated stronger agreement that misleading statements improve an applicant's position in the Match (p = .000, R = .143). In contrast, applicants contacted by a higher proportion of programs were significantly less likely to agree (p = .011, R = −.094). Applicants feeling more pressure to offer assurances (p = .000, R = .272) and those making misleading statements (p = .000) indicated significantly stronger agreement than did their counterparts.
Misleading statements may be justified
Agreement with the statement, “Applicants may be justified in making dishonest or misleading assurances or statements to programs” varied by medical school (p = .000) and medical school rank, with applicants from schools not ranked among the top 20 schools agreeing more strongly than applicants from higher-ranked programs (p = .000). There was no significant difference by the school's region or public or private affiliation, or by the applicant's gender or minority status.
Agreement with this statement also varied by specialty (p = .000), with applicants to specialties with high fill rates indicating significantly stronger agreement (p = .000). Agreement did not vary based on the level of communication. Applicants feeling more pressure to offer assurances (p = .000, R = .236) and those making misleading statements indicated significantly stronger agreement than those who did not make misleading statements (p = .000).
The extent and content of communication between applicants and programs in the Match vary significantly by medical school and the applicants' specialties. Applicants from lower-ranked schools and those applying to specialties with higher fill rates are more likely to initiate contact with programs. Programs are more likely to contact applicants from higher-ranked medical schools and applicants applying to specialties with lower fill rates. Applicants who communicate with programs feel greater pressure to offer assurances to programs, regardless of who initiated the communication. In different instances, the pressure to offer assurances may serve as both the impetus for applicant-initiated communication and the result of communication initiated by programs. Applicants who feel greater pressure to offer assurances to programs are more likely to make misleading statements or assurances.
Applicants commonly encounter situations in which telling the truth about their preferences harms their applications at programs where their interest is not as great. Making misleading statements appears to help an applicant's position in the Match. Applicants who make misleading statements match at a higher rank average than do those who do not make misleading statements, even though they are from lower-ranked medical schools.
Many of the misleading statements and violations of NRMP policy arise from the uncertainty of the process. Students often have difficulty judging how highly they are going to match. In many cases, attempts to ensure that they will match at a program they will want to attend, even if it is not their first choice, may be prudent. In other studies, most applicants reported satisfaction with the Match process overall.4,6 One might argue that the communication between programs and applicants relieves some of the uncertainty about the outcome and helps both parties create a good Match. This may indeed be the case for honest communication, but remains unlikely with pressure-filled communication.
Whereas students' unfamiliarity with the Match process may lead to unethical statements or seemingly unnecessary actions, residency programs understand the process and the rules. While some program directors abstain from recruitment altogether, others appear to construct their Match lists with the goal of “matching well,” i.e., not having to go too far down their lists. To achieve this, knowing where applicants plan to rank them is a high priority.
While some programs face uncertainty about filling residency slots because of past failures, the only uncertainty for many of the programs is the number of applicants on their rank list they will need to fill their class. One of our respondents, an applicant to dermatology, observed:
Many programs are obsessed with bragging rights about how highly they stayed on their rank lists. The selection process was often skewed in favor of those who had a personal connection to the program or who had strong personal reasons for being in that particular geographic area. Even if a particular program thought you were the best applicant, they wouldn't necessarily rank you highly if they thought they couldn't get you.
Students believe that programs often rank stronger applicants behind those who have demonstrated an interest, as one respondent, an applicant to psychiatry, noted:
I definitely got the feeling that some programs might not rank me if they thought I wasn't going to rank them extremely highly. I don't know how it becomes public how far down their rank list programs had to go in their Match, but this seems to be the crux of the issue. Programs would rather rank 20 less desirable candidates who are 100% sure of coming, than 40 candidates who are 50% sure of coming, just because of the loss of status that comes with having to go down to number 40, as opposed to number 20. If this were kept completely confidential, programs would have no incentive to take someone off their list who they would like to have, but they don't think they can get. I think not being ranked because you are “too good” is pretty appalling.
The varied patterns of communication between programs and applicants raise ethical dilemmas for many applicants. Applicants in primary care specialties may negotiate a larger volume of communication, although with fewer personal, direct appeals. Applicants to non-primary care specialties with high fill rates face a relative dearth of general feedback but may be subjected to more direct, pressure-filled communication from programs that want them. Many applicants report conflicted feelings about the steps that must be taken to protect their places in the Match. One of our respondents, an applicant to radiology, wrote:
I was told by a staff member at one of the programs that I was interested in that “…the Match is a game. You have to play it.” There were several things that the interviewers asked me that may have been illegal and were certainly inappropriate. I had to answer them because if I refused, I would have ruined my chances of matching at that program. I was therefore justified in giving the answer that I thought they wanted. In my honest opinion, the Match has very little to do with honesty and more to do with telling people what they want to hear. The interview process was a very uncomfortable and disheartening experience for me.
The pressure applicants feel may affect their attitudes about whether dishonest communication may be justified.
Hundert and colleagues9 introduced the concept of the “informal curriculum” to illustrate the many settings in which students learn professional behavior. They argued that there is an “enormous range of influences on a medical trainee's development of ethical and professional standards of behavior.” These influences take place inside and outside the classroom, in both formal and informal settings, as the “informal curriculum is enmeshed in the culture in which students and residents live.”9
The residency selection process lasts up to nine months, from application through interviews to Match day, consuming almost the entire final year of medical school. It is arguably one of the most influential periods in the professional development of medical students. There is growing concern that many applicants find that telling the truth harms them in the Match, while others benefit from lying or from misleading the programs. Applicants may “learn to lie” during this process to advance their careers. What effect this will have on subsequent clinical interactions and professional relationships is unclear.
The current patterns of Match communication may distort applicants' understanding of their own difficulties in the Match. For example, in our study, applicants matching at lower-rank-order programs were more likely to agree that misleading statements were beneficial in the Match. Some applicants seemed to attribute their unexpectedly poor Match performances to the Match process and to other people's misleading statements rather than to their own qualifications. One respondent, an applicant to orthopedics, wrote:
I was asked multiple times by the program directors of other institutions what number I would rank their program. I thought that this was unfair by the rules of the Match. I have to admit, I answered their questions honestly because I was in the spotlight and felt like I was being pressured by getting the same question multiple times from the same institutions. I felt uneasy about answering, but told them the truth, which was that their program was not first on my list, because my home program was. Now that I didn't Match, I wonder if this played a role in the Match process.
In their study of the urology Match, Teichman and colleagues5 outlined numerous ideas for improving the Match, including “an ombudsman to address code violations without fear of recrimination.” This study supported the establishment of a well-publicized, centralized system for reporting Match violations to deter violations of NRMP policy. A standing committee of the Association of American Medical Colleges or the NRMP that had resident and student representation could review the complaints and sanction applicants and programs as appropriate.
Results from our study and others strongly suggest that communication initiated by programs after interviews should be prohibited, because applicants may feel pressure to respond. NRMP policy regarding the acceptable content of communication between programs and applicants must be expanded and widely published. Applicants should be allowed to contact programs, but should be limited to informational questions and answers. However, such communication should not be solicited by the program, nor should programs inquire about an applicant's rank order list.
Perhaps the most important change for improving the Match involves changing the way residency programs report Match results to deans, hospital CEOs, their own residents, and prospective applicants. Medical schools have stopped summarizing the performances of their graduating classes each year by the percentages who matched at their first choice, second choice, and so on. Similarly, residency programs should be prohibited from summarizing their Match results in reference to their rank lists.
Our study was limited by the overall response rate and the significant difference in response rates among schools. Graduating medical students are difficult, however, to engage in survey. The use of a login and password, while having the benefit of securing the online questionnaire, may have served as an obstacle to responding, negatively affecting the response rate. A sample substantial in size, with geographic diversity, and inclusion of a mixture of public and private institutions helped to offset the response rate. In addition, comparisons of respondents and non-respondents suggest that the sample may be representative of applicants in the 2001 Match. We do not know whether a survey with a higher response rate would produce different results.
Another limitation of our study is the inability to identify the direction of the relationships between factors affecting Match communication. Does communication shape attitudes or do attitudes predict the communication that occurs between programs and applicants? The results of our study suggest that the answer is both. They bring us closer to a general understanding of the extent and possible effects of communication between programs and applicants and serve as the basis for subsequent research. Further studies are needed to better understand the effects of Match communication on the ethical and professional development of medical students.
Communication between applicants and programs in the Match varies significantly by medical school and specialty. Applicants who communicate with programs feel greater pressure to offer assurances of their interest to programs, regardless of who initiates the communication. Applicants feeling greater pressure to offer assurances to programs are more likely to make misleading statements, which may improve their positions in the Match. NRMP policy should be amended to strictly limit such communication, as it violates the spirit if the Match, and may have undesired effects on the professional and ethical development of medical students.