INTRODUCTION
In Brazil, students that begin higher education at universities come directly from high school through a very competitive process, based mainly on assessment ofknowledge. This process has generated disadvantages for students from secondary public schools coming mainly from lower socioeconomic strata. In order to improve this selection process, federal government and institutions have launched several inclusion policies, including bonuses and quotas for public high school attendance, students of black/indigenous race and low socioeconomic background1,2, as well as providing loans to pay forattending private universities. With the inclusion policies growing, questions about student performance and dropout rates due to lack of financial support have been raised3,4. In public universities, different scholarship strategies are available, not always for the purpose of socioeconomic assistance, since initially the scholarships were earmarked for the best students with the intention of starting in the world of research, guided by their professors; it was later found that some of the difficulties mentioned above could be surpassed, not only for the participants of research projects, but also for those who worked within the university for approximately 10 hours a week.
Previous studies have shown conflicting results regarding the performance of students who receive scholarships. However the number of students, years of follow up and type of undergraduate courses could have influenced these differences3,5,6,8. Some studies have shown that students receiving a scholarship showed a more supportive attitude to colleagues outside the classroom, participated more in community service activities, engaged in more extracurricular activities6, and tended to persist in the academic career, seeking master’s and doctorate degrees9.
In this study, we aim to evaluate the performance of students with the scholarship during the six-year course of medical school and during the posterior selection process for Medical Residency at a large and highly competitive public university in Brazil.
METHODS
Analysis was conducted of socioeconomic and performance data of students that graduated between 2010 and 2013 at the University of Campinas (Unicamp) Medical School.The Ethics Committee of the Faculty of Medical Science approved the study (IRB 515.297). Data on the socio-demographic profile, obtained from each student when applying for an undergraduate position, included current age, sex, self-declared race or color, household income, educational level of father and mother and number of family members depending on that income. Per capita income was calculated in monthly minimum wages divided by the number of people depending on the household income, calculating for minimum wages of 115.3 dollars in 2005, 159 dollars in 2006, 190 dollars in 2007 and 197 dollars in 2008.The parents’ level of education was classified as fundamental school for the first 9 years, followed by 3 years corresponding to high school, both grouped as basic education, in accordance with the National Curricular Guidelines for Basic Education, and the 2014 National Education Plan. Basic schooling was considered either complete or incomplete. Higher education associated with undergraduate degree education was also reported as complete or incomplete.
Two categories of scholarships are available: those associated with research initiation projects, and those exclusively for social support. Social support scholarships for housing, food and transportation, are dependent of a detailed social selection process. Various programs support different amount of scholarships for research, including the Institutional Program for Scientific and Technological Initiation Grants (Pibic/Pibiti), financed by National Council for Scientific and Technological Development (CNPq) and Unicamp, and from the Foundation for Research Supportof the State of São Paulo (Fapesp).
The information about the cumulative undergraduate performance at six (middle) and twelve (final) semesters of medical school (GPA6 and GPA12respectively), and scholarship received, was obtained from the database of the Academic Directorate. The residency admission results were obtained from the Permanent Commission for Entrance Exams (Comvest).The five components of the assessment system for residency admission analysed included multiple-choice questions, short essay questions, clinical skills assessment (CSA), interview, and total residency admission score.
Statistical analysis
A database was set up in SPSS 18.0 and the tests to ward off the effects of possible collinearity between the variables were performed along with the cleaning of the database. For all analyses, a p-value less than 0.05 was considered statistically significant.
Categorical data were compared by Fisher’s exact test. The Student’s t-test and the Mann-Whitney test were used to analyze the receipt of scholarships with per capita income. Linear regression was used to analyze the academic performance coefficient, here named Grade Point Average (GPA) in the sixth and twelfth semesters, and the performance at the medical residence selection exam as dependent variables; required age, socioeconomic level, received bonuses for having studied at a public institution for the entirety of high school education, and the educational level or schooling of the parents were control variables.
RESULTS
Demographic data
The study included 417 medical students enrolled from 2005 to 2008. The mean age was 20 years (range 17 to 33), with a predominance of female 230 (55.1%) students. Three hundred and thirty-nine (81.2%) students declared themselves white, 33 (7.9%) yellow, 31(7.4%) brown, 2 (0.5%) black and 2 (0.5%) indigenous. Regarding household income, 35 (8.4%) students reported receiving up to 5 minimum wages; 125 (29.9%) students reported higher than 10; 90 (21.5%) students reported a household income between 5 and 10 minimum wages. A total of 78 (19.3%) households had one minimum wage per capita, with the mean greater than three in both groups. Almost half the students had parents who had completed higher education (208 or 49.8% for fathers and 194 or 46.4% for mothers), while 79 (18.9%) fathers and 59 (14.1%) mothers had completed a graduate degree.
Scholarship and its impact
Among 243 (58.2%) students with a scholarship, 217 (89.3%) were awarded exclusively for research activities and 20 (8.2%) exclusively for social support, while 6 (2.5%) received the scholarship for both purposes. A significant increase in the number of students benefited by scholarships was observed (Table 1).
Year of admission | ||||
---|---|---|---|---|
|
||||
Scholarship | 2005 | 2006 | 2007 | 2008 |
N (%) | N (%) | N (%) | N (%) | |
Receiving | 48 (47.1%) | 58 (53.7%) | 63 (59.4%) | 73 (72.3%) |
Notreceiving | 54 (52.9%) | 50(46.3%) | 43 (40.6%) | 28 (27.7%) |
Total | 102 | 108 | 106 | 101 |
X2 = 16.45, p = 0.002.
No difference was observed between the father’s (p=0.33) or the mother’s (p=0.40) schooling, sexes (p=0.21), age (p=0.71), race (p=0.69), type (private or public) of primary (p=0.32) or high school (p=0.29) and household income (p=0.19) comparing students with and without scholarships, although there was a lower per capita income among those receiving a scholarship (p<0.01) (Table 2).
Scholarship | N | Average | SD | |
---|---|---|---|---|
Per capita income | Yes | 224 | 3.68 | 2.39 |
No | 162 | 4.39 | 2.64 |
Student’s t Test (T = 2,6/p<0.01).
Students who received scholarships performed better in the middle (p<0.01) and at the end of the medical course (p<0.01) (Table 3). In the selection for medical residency, no difference was detected for any phase of the exam, including multiple choice, short essay questions, clinical skills assessment, interview and total score (Table 4).
Scholarship | N | Average score | SD | t | p | |
---|---|---|---|---|---|---|
GPA* 6 | No | 175 | 0.78 | 0.04 | -3.31 | <0.01 |
Yes | 242 | 0.80 | 0.04 | |||
GPA* 12 | No | 175 | 0.81 | 0.03 | -3.49 | <0.01 |
Yes | 242 | 0.83 | 0.03 |
GPA* = Grade Point Average (coefficient of student performance).
Scholarship | N | Average score | SD | T | P | |
---|---|---|---|---|---|---|
MCQ | No | 157 | 7.29 | 0.94 | -0.50 | 0.61 |
Yes | 225 | 7.34 | 0.92 | |||
Essays | No | 157 | 6.04 | 2.10 | 0.72 | 0.47 |
Yes | 225 | 5.88 | 2.23 | |||
CSA | No | 157 | 6.06 | 3.00 | -0.31 | 0.75 |
Yes | 225 | 6.16 | 3.03 | |||
Interview | No | 157 | 7.30 | 3.59 | -0.85 | 0.39 |
Yes | 225 | 7.62 | 3.69 | |||
Total MR | No | 157 | 7.38 | 0.65 | -0.59 | 0.55 |
Yes | 225 | 7.43 | 0.68 |
MCQ = multiple choice questions; CSA = Clinical Skills Assessments; Total MR = final score of medical residency.
DISCUSSION
We observed an absolute increase in students receiving scholarship over the years. Those being granted scholarships were from families with lower per capita income and had a better performance during medical school.
In Brazil, research scholarship has been progressively more valued as a strategy for student-centered education and engagement. From 1986 to 1995, scientific scholarships by the national research agency increased from 1,000 to 19,000, reaching almost 28,000 between 2004 and 2010, after a decrease in 1997, 1998 and 2003. Even though almost 60% of students were awarded in this sample, the incentive is not universal and does not fit the parallel increase in positionsfor higher education in recent decades, and the new student profile10.
One issue of relying on scientific research scholarship can be observed analyzing data from the official National Council on Research (CNPq) website. The shares of scholarships are passed on to the professors who meet the terms of the calls. Thus, the quotas for scientific initiation scholarship are proportional to the number of registered researchers, and number and size of graduate programs. Since this study was conducted at a strong research university, with more than 90% of its faculty showing a doctoral degree, the number of scholarships granted increased even more than the increase in the number of students.
Our study population differs from Pereira and Pazello11regarding parent’s schooling. These authors found lower household income and lower parental level of education among higher education students receiving scholarships for social support, with no difference for gender, according to the answers provided by participants at the National Examination of Performance for Higher Education Students (Enade), with national coverage. In a large Brazilian study, including 787,470 students from public and private schools, a positive association between income and performance, regardless of scholarship, was observed3. According to these studies, if, on the one hand the social capital of higher income students helps them perform better, on the other hand better financial support for those in need should help them overcome differences, giving full attention to their educational program.
It is worth noting that differences in scores may have been influenced by the large sample size, which tends to overvalue small differences, leading to statistical significance12.
For these students at Unicamp, the main reasons for applying for a research scholarship were the intention of improving their curriculum and deepen their knowledge in the area, having a closer, direct support from the professor, and, to a lesser degree, due to the financial aid13. Several studies have observed that students receiving research grants (regardless of socioeconomic status) develop better academic performance and personal growth14-20.
In addition, published research shows that students from underrepresented minority groups that received academic and mentoring support displayed better retention and performance21-23. Three other studies also showed marked improvement of performance associated with financial aid24-26. However, whether the better performance is due to greater financial support, the presence of a supervisor close to the student, or the scientific reasoning associated with research activities themselves (including self-management, bibliographic search, critical thinking) cannot be answered25,26.
Our findings corroborate with Boatman and Long6 who also found a better performance score (GPA) in the third year of the medical course among the talented undergraduate students of color with financial need, who received a renewable scholarship for talented students, as part of a Program named Gates Millennium Scholars (GMS). Nevertheless, after correction for race, the performance was equivalent. In our study, we observed no difference between scholarship recipients in relation to race, age, type of primary or high school, and parents’ education. One question that arises is whether the scholarship recipients were actually the best performers, or whether the best students and those with the highest incomes did not participate in the scholarship programs during this period, which can be modified over time by a higher valuation given to a curriculum full of publications, especially for those who seek to subsequently pursue an academic career. However, this does not change the fact that we find a pattern contrary to other publications, where higher income would be related to better performance.
Matsumoto et al.7, evaluating the performance of students participating in the Japanese government’s Quotas and Scholarship program, focused on public schools, also observed that their academic performance was superior to that of their peers.
One major limitation is that it was not possible to adequately separate students who were awarded social assistance scholarships from those who received them for research activity. We know which students received a scholarship for socioeconomic assistance (6.2%) in the four years studied, but we cannot identify how many received the scientific scholarship that would also need economic assistance. However, we showed that per capita income among scholarship students was significantly lower than the remaining group, corroborating the good performance of the less favored students who received scholarships. This minority that was associated with the scholarships earmarked for socioeconomic aid, did not change the result of the analysis when excluded from the statistical model for income analysis. In a qualitative and quantitative study, in which 27 students from the Federal University of Santa Catarina were interviewed, 18 students affirmed receiving a work scholarship (at the university), internship scholarship, fellowship monitoring or scientific initiation as a way of maintaining subsistence in higher education27. We believe that the same may have happened among our students. It must also be noted that research scholarships for students can be directed by the faculty, who tend to select a student with outstanding performance, but can consider other factors as social disadvantage.
One point that deserves clarification is that the performance of students who were candidates for receiving scientific initiation scholarships was only taken into account for the purposes of tie-breaking, not as a primary selection criterion, and was not understood as a relevant confounding factor.
Another common discussion that arises from analysis of this research is the option of analyzing in a single group all students receiving scholarships in undergraduate courses, not stratified by the types of scholarships. This choice was made because the group was mostly homogenous, since 217 (97.31%) received a scientific initiation fellowship, of which 130 (58.3%) were from the Institutional Scholarship Program (Pibic), 87 (39%) from Foundation for Research Support of the State of São Paulo (Fapesp) and 4 (1.8%) others, such as the tutorial education program –PET; 20 (8.2%) of the scholarships were for social support alone. Given that statistically, comparative groups with less than 25 members lose reliability, we opted to analyze the students in a single group, which due to the large sample size, would make the outliers unable to significantly influence the result; however, further studies with the division between groups should be undertaken.
There is evidence to indicate that students who enter and remain in higher education through inclusion policies are as able as other professionals28. Thus, students at a social disadvantage do not become less skilled professionals. Scholarships could therefore help these students. Additional studies are important to elucidate to what extent each component of an institutional support program can influence the performance and retention of the university student in different fields of knowledge.
CONCLUSION
Better academic performance was reported among medical students who received scholarships, despite presenting a lower household income. This study reinforces that disadvantaged household income is not necessarily reflected in academic disadvantage, other forms of student support during undergraduate training may contribute to better performance.