Participants and Study Design
We performed a cross-sectional survey of fourth-year medical students applying during the 1999 National Residency Matching Plan (NRMP) in internal medicine. Potential participants were chosen from the lists of all medical students applying for placement in one or more of four internal medicine residency programs, including the University of California-San Francisco, University of California-Los Angeles, Massachusetts General Hospital,and New York Presbyterian Hospital. These four institutions were selected because of similar quality, and thus, would be expected to attract a similar pool of applicants.
Lists of applicants were provided by each institution and were cross-referenced for duplicate applications. One institution (UCLA) mailed their own surveys to assure complete anonymity of their applicant pool. NRMP numbers were provided by this institution to allow cross-references for duplicate applications. Only one survey was mailed to each student, and follow-up surveys were not sent to nonresponders in an attempt to limit response bias from students who did not match at their desired location. All applicants with mailing addresses outside of the United States and Canada were excluded from the study in an attempt to limit international medical graduates (IMG) from responding. We limited IMG for several reasons: 1) these applicants tend to be less competitive in the NRMP at these four institutions, 2) because of extremely low acceptance and interview rates of these students, we felt that their inclusion could result in significant response bias, and 3) they may have significantly different motivating factors that influence their residency choice. Because of the sampling procedure described above, the exact number of international medical graduates excluded is unknown. The remaining 2,908 students were mailed a questionnaire. Only one mailing was possible in the one month between receiving the students’ addresses and the results of the match. This timeframe was adhered to in order to decrease possible bias introduced by the students’ match results. The institutional review board of University of California-San Francisco approved this study.
viagra soft tabs online
The questionnaire obtained demographic information, including age, gender, ethnicity, relationships with significant others, medical school and class standing. Ethnicity was described as white/Caucasian, African-American, Latin American/Hispanic, Native American, Pacific Islander, Asian or other. URM were classified as African-American, Latin American/Hispanic, Native American and Pacific Islander. Class standing was determined by applicant’s self-ranking in the top quarter, middle half or bottom quarter of the class, and whether or not they anticipated nomination into their local chapter of Alpha Omega Alpha (AOA), if applicable.
Information on the importance of 46 factors in choosing a particular residency program was obtained. Items chosen for evaluation were derived from prior studies, discussion with program directors in internal medicine and focus group discussions with current residents. Focus group discussions were used to generate evaluation items but did not address the cultural validity of the survey. The surveys were then pretested for completeness and clarity. Factors assessed included 14 on location characteristics, 20 on program features, six on recruitment strategies, three on future plans and three reflecting advice received. Location characteristics included questions on gender and ethnic diversity of residents, faculty, and patients and geographic aspects of area. Questions also addressed programmatic factors, such as academic reputation, program size, program description, benefits/financial incentives and emphasis on primary care as well as interview techniques and process opportunities for future training or jobs in the area and advice received from a role model, friend or dean. Data were gathered using a five-point Likert scale, with 1 indicating that the factor was not important and 5 indicating that the factor was very important in the applicant’s selection of a particular residency program. A factor was described as important if it received a mean response of 3 or greater. cialis soft tabs online
Data were analyzed using SAS version 8.2. Descriptive statistics, including means, standard deviations and percents, were computed for each of the variables by gender and ethnicity (URM, Caucasian and Asian). Initial exploratory analyses for race and gender differences included Chi-squared tests for categorical variables and Student’s t test for continuous variables. We further developed logistic regression models to explore the association of ethnicity (URM vs. non-URM) with the demographic measures and the factors in selecting a residency program. Our primary measures of outcomes were the factors in choosing a residency program. The major hypotheses were evaluated using analysis of covariance models (ANCOVA) to examine any gender or ethnic/racial group differences. For each outcome variable, we investigated the interactions among ethnicity, gender, age, marital status and class rank; we also compared Akaike’s information criterion (AIC) and the residual log likelihood of the models with and without the interaction terms. Since the main effects models tended to be more parsimonious with smaller AIC values, we decided to base our results on the main effects models. Estimates of adjusted means and standard errors (SE) were obtained by gender and by race, controlling for ethnicity, age, marital status and class rank. A significance level of 0.05 was used for all statistical tests.
cetirizine 10 mg