Selecting the brightest for post secondary education in Colombia : the impact of equity

THE WORLD BAN. Discussion Paper EDUCATION AND TRAINING SERIES Report No. EDT61 Selecting the Brightest for Post-Secondary Education in Colombia: The Impact of Equity Emmanuel Jimenez Jee-Peng Tan February 1987 Education and Training Department Operations Policy Staff The views presented here are those of t-he author(s), and they should not be interpreted as reflecting those of the World Bank.  Discussion Paper Education and Training Series Report No. EDT61 SELECTING THE BRIGHTEST FOR POST-SECONDARY EDUCATION IN COLOMBIA: THE IMPACT ON EQUITY Emmanuel Jimenez and Jee Peng Tan Research Division Education and Training Department February 1987 The World Bank does not accept responsibility for the views expressed herein, which are those of the author(s) and should not be attributed to the World Bank or its affiliated organizations. The findings, interpretations, and conclusions are the results of research or analysis supported by the Bank; they do not necessarily represent official policy of the Bank. Copyright C)1987 The International Bank for Reconstruction and Development/ The World Bank  ABSTRACT Selection in post-secondary education is highly competitive in many developing countries. Ability is often an important criterion, but economic reasons may account for a bias against people from poor families. This paper uses Colombian data to examine the impact on equity if selection were based solely on ability. The resulting change indicates a significant increase in the share of lower income students at the expense of students from wealthier backgrounds. For such a shift to materialize, however, strengthening the availability of financial aid would be necessary to relieve economic constraints barring the access of bright students from poor families.  The comments of George Psacharopoulos and Martin Carnoy on an earlier draft are gratefully acknowledged. Kyuee Ha Phak provided able research assistance. Partial funding was obtained from the World Bank Research Project 673-96.  In many developing countries, access to post-secondary education is highly selective. The demand for this level of education is frequently very strong because of heavy public subsidization. However, not all qualified applicants can be accommodated because budgetary constraints often limit the government's capacity to expand the supply of places. The result is that in many of these countries, admission into post secondary education is rationed according to some selection criteria, such as examination results. For various reasons, most systems of selection tend to discriminate against qualified students from poor families. In examination-based systems, for example, such students are at a disadvantage because their home environment is seldom conducive to study, and they may not be able to afford additional out-of-school inputs, such as private tutoring, to boost their scores. For most of them, repetition is not an affordable option, so that even the most intelligent candidates are at a disadvantage because they are competing against others who have had the benefit of at least one more year of intellectual growth (Somerset 1974). Self-selection is another important reason why some high ability candidates do not enter higher education. The long queue for admission, and the fact that chances of being admitted are slim might discourage potential entrants from applying. The incidence of non-application is likely to be higher among people from poor families, because of the lack of financial resources to support further studies after secondary school. Even when tuition is free, the student still has to incur the cost of  -2- forgone financial resources to support further studies after secondary school. Even when tuition is free, the student still has to incur the cost of forgone income. In the absence of educational credit, these students cannot finance their studies against future earnings. For this reason, even if some candidates from poor families do apply and are selected, they may not be able to accept the offer. If secondary-school graduates who do not proceed into higher education are more qualified than some of the entrants, and if ability and education complemented each other to determine a person's productivity, this result would be inefficient. The misallocation of resources across individuals arises because the returns to additional education would have been greater if selection were influenced primarily by a person's aptitude rather than by his/her socio-economic background. This point has been demonstrated in a study by Pinera and Selowsky (1981), where the magnitude of the efficiency gains is presented through the use of simulation analysis. The possible under-representation of qualified candidates in the poor families also has obvious adverse effects on equity. The purpose of this paper is to document the equity impact of reforming selection outcomes in Colombian higher education. The paper is organized as follows. The next section briefly reviews the analytical framework for assessing the extent to which a selection system discriminates against high ability candidates. The results of applying this framework to data from Colombia are reported in the succeeding two sections.  An Analytical Framework One way to evaluate the present selection criteria into Colombian post-secondary education is to answer the following questions: If se:Lection were based solely on ability, how would the present enrollment patterns change? In particular, how many secondary school graduates who presently do not participate would then enter higher education? How many poist-secondary students would lose their place to these new entrants? The impact of a selection reform can be illustrated diagrammatically. In figure 1 (pre-reform), the distribution of ability scores for all secondary school graduates is marked as f. Only some in this pool of potential candidates can afford the private cost of post-secondary schooling. The distribution of academic ability of persons in this latter subset is marked as h in figure 1. Since we assume that incomes are uncorrelated with ability scores, distributions f and h are centered around the same mean ability score - point A. Not all persons able and willing to pay for post-secondary education may in fact enter. In particular, if there is excess demand, selection would be restricted to those meeting a minimum ability criterion, such as A1 in figure 1. Those selected are represented by the shaded area in figure 1. Their mean academic ability or aptitude would be A2 (>A).  Frequency Al A A., Ability Figure 1 The likelihood that a high ability student does not enter post-secondary education depends on the extent to which factors other than ability matter in the choice for further education. Referring to figure 1, this probability is greater the greater the area between f and h. If selection were based entirely on academic aptitude, some of the entrants to post secondary education would be replaced by more qualified persons. The outcome of this new selection criterion is depicted in figure 2 (post-reform). The displaced entrants are represented by area II. They are replaced by students represented by area III. Those in  -5- sysem.Since r eyhosen under the prereforra selectiOn syste of places in post-secondary educi s post-secora a are equal. The imean acadeemic aptitudeoftudents in poNturallyar eduation e after the selection reform is indicatdb dentu l A e~eeducat2n the pre-refo level Of aptitude along secondary education. IC I u 2bilit Figure 2  - 6 - Data The data for this analysis were originally collected under the World Bank's study to evaluate diversified secondary education (DiSCus project) in Colombia. The first round of data were raised in 1981 in a survey of final year secondary school students. Baseline information was gathered on each student's socioeconomic background, as well as his or her score on a battery of tests. Two types of tests were administered, one to measure mathematical and verbal aptitude, and the other to measure achievement in subject areas, such as commerce, agriculture and technical knowledge. The aptitude tests "measure the student's innate mental ability, and hence, predict how well an individual may learn" (Psacharopoulos and Loxley 1986). The students were followed up one year later, when information became available about their work or study status after secondary school. For the analysis here, we define two dichotomous populations: individuals who study full-time or part-time in 1982 are categorized as students; the other group consists of those who work full-time after graduating from secondary school. The broad definition of students is used here because it would highlight the importance of factors other than ability in determining selection outcomes. The bias in selection would be more significant if a high ability person did not study at all, than when he or she studied part time.  Reforming Selection in the Overall System Ability is assumed to be measured by the average score on the verbal and mathematics tests. The mean ability score of students was 51.74 percent, compared to 51.62 in the overall sample (table 1); these figures correspond to A2 and A respectively in figure 2 above. It is interesting to note that the range of the ability scores in both populations is the same, suggesting that ability is not central to the present selection system. Consider the effect of the following hypothetical reform is then examined: Reform A: All students and non-students in the sample are ranked according to their average ability score. Keeping the total number of places in the system constant, a student is replaced if he/she scored less than the highest ranked person among the available pool of non-students. How many present post-secondary students would be replaced by secondary school graduates who are presently in the labor force? In the overall population of 1479 students, 841 or 56.9 percent would be replaced by new students of higher academic aptitude. An interesting point is that of the replacement students, nearly half (404 students) did not apply to enter any form of post-secondary education. This implies that altering selection criteria used by higher level universities and other institutions would not be sufficient to draw all high scoring individuals. Demand would also have to be stimulated to convince them to apply.  The replacement of low ability students with high ability non-students in post-secondary institutions will affect average scores and the socio-economic profile of the "new" student population. The results appear in table 1. After the hypothetical reform, the ability score of the student population increases to 58.97 (A3 in figure 2), a rise of 14 percent. The reform does not change significantly the proportion of entrants from public secondary schools. The impact of the selection reform on the distribution of students by income group appears in the last six rows of table 1. Relatively few graduates of secondary schools come from the lower income groups. But the present system of selection into post-secondary education worsens this bias. The proportion of students from the lowest two quintiles of the income distribution profile is 22.7 (5.3 + 17.4) percent for the overall sample, compared to 20.7 (4.9 + 15.8) percent for the subsample of students. On the other hand, the proportion from the highest two income groups is 53.8 for the overall sample and 57.5 percent for the student subsample. Had selection been based only on the ability score alone, the share of students from the lowest two income groups would have increased to 23.2 percent. This increase comes at the expense of a reduction in the share of those in the highest income quintile, whose share would decline from 24.4 to 21.7 percent. The percentage share of the middle 40-80 percent of the income distribution profile remains virtually unchanged.  - 9 - Table 1: Characteristics of the Student Population Before and After Selection Reform A-Colombia 1981 Secondary School Cohort Students and Students Only Non-Students Before Reform After Reform A Number in Sample 3494 1479 1479 Ability Score Mean 51.62 51.74 58.97 Standard Deviation 7.95 7.70 4.59 Minimum 26.00 26.00 53.00 Maximum 77.00 77.00 77.00 Percent From Public Secondary Schools 76.7 76.9 75.6 Distribution of Students by Household Income Quintile (%)a/ 1 5.3 4.9 5.1 2 17.4 15.8 18.1 3 23.4 21.8 22.0 4 32.9 33.1 33.1 5 20.9 24.4 21.7 Total 100.0 100.0 100.0 Percent Female 47.4 48.5 41.0 a/ The income groups are the same as those in the 1978 Colombian urban income survey (Mohan, forthcoming). If one accepts the hypothesis that innate ability is uncorrelated with income, then the distribution of students across different income strata should be close to that of the overall sample (students plus non-students). This can be done by comparing the student profiles with the first column of the income distribution profiles of table 1. While the match is not perfect, the post-reform (relative to the pre-reform) profile is much closer to the overall sample profile. The reform would also reduce the proportion of females in the student population -- from 49% to 41%. In selecting purely on the basis of ability, there  - 10 - seems to be a trade-off between two equity goals: there is a more equitable share of student-places for the lowest 40% in the income distribution but a proportionately smaller share of places for females. How can this be, if innate ability is equally distributed among males and females? One possible explanation for this result is the following. Among those who are less able, the private rate of return to schooling may be greater for females than for males. Opportunity costs may be greater for males who are expected to be semi-independent of the family shortly after secondary school. However, the private rate of return differential between males and females may lessen with ability. This could occur because for those with greater ability, the private cost differential between males and females is not as important in the choice calculus. Opportunity cost effects would tend to be offset by the gain in expected earnings in determining the private rate of return, due to the complementarity between innate ability and education. If so, lower-ability females would be overrepresented in the pre-reform subsample. The reform would then lead to a decline in the proportion of females in the student subsamples. Reforming Selection by Type of Institution The impact within each type of institution of selecting by ability alone can be assessed by imposing the following reform: Reform B: Students in each type of institution are ranked according to their ability score. Keeping the number of places in each institution constant, students are replaced by higher-scoring non-students. We examine below the impact of Reform B in the universities and technical institutes, and in public and private higher education.  - 13 - Selection in Universities and Technical Institutes: There are several types of higher education in Colombia, three of whiLch are distinguished in the present data base: university, intermediate technical institutes, and other institutions (pre-university or in-service training centers such as SENA). The profile of the student body before reform is shown in the first two columns of table 2. There is great similarity between the two types of institutions: in the average ability score, in the proportion of students coming from public secondary schools, in the proportion of females, and in the distribution by income group. Following Reform B, about three-quarters of the entrants to universities, and over 90 percent of the entrants to technical institutes would be replaced. In both cases, about half the replacement students did not apply to enter post-secondary education. As before, the high rate of non-application indicates that self-selection for economic reasons might have been important in determining who eventually enters higher education. The representation of graduates of public secondary schools, and that of females in both types of institutions would decline as a result of the reform. Not surprisingly, the average ability score of the new student population would rise dramatically: by 19.5 percent in universities, and by 28.1 percent in technical institutes. The increase in mean ability score suggests that both universities and technical institutes have difficulty in attracting the students with the best academic potential. But the problem is more severe in the technical institutes.  - 12 - Table 2 also indicates that Reform B would change the composition of the student body by income groups. In both universities and technical institutes, it would shift enrollments in favor of the lower income groups. Correspondingly, fewer students would come from the highest income quintile after the selection reform. These results show that there would be more equity in the access to post secondary education if entry depended only on a student's aptitude score. Table 2: Selection in Universities vs. Technical Institutes Before Reform After Reform B Universities Tech.Inst. Univ. Tech.Inst. Number of Students 641 203 641 203 Ability Score Mean 51.83 51.59 61.94 66.11 Standard Deviation 7.60 7.51 3.87 3.26 Minimum 26.00 34.50 57.00 62.50 Maximum 77.00 72.00 77.00 76.50 Percent From Public Secondary Schools 77.7 77.3 75.0 70.0 Percent Female 49.9 50.7 37.6 32.5 Replacement Students Number - - 482 186 % of Total in Each Type of Institution - - 78.2 91.6 % Non-applicant - 49.8 55.4 Distribution of Students by Household Income Quintile 1 3.4 4.9 4.7 5.4 2 15.0 13.8 20.4 21.2 3 20.3 16.8 20.8 21.2 4 33.5 38.4 36.4 36.5 5 27.8 26.1 17.8 15.8 All Income Groups 100.0 100.0 100.0 100.0 -: not applicable  - 13 - Selection in Private and Public Higher Education: In Colombia, public higher education is highly subsidized. Fees charged to students are very low, constituting, on average, only 5% of overall income. In contrast, private higher education institutions must recover most (85% on average) of their costs through fee revenue (Reyes-Posada 1986). Because of the differences in private costs, we should expect differences in the socio-economic profiles of students in the two types of institutions. The characteristics of the student populations before and after Reform B appear in table 3. Before the reform, the average ability scores of those who enter public institutions of higher education are about the same as those who enter private institutions. There is also little difference in the percentage of these two groups who are female or who come from public secondary schools. As expected, the lowest income quintiles of the population are under-represented in the group entering private institutions. only 13.2% come from the bottom 40% of the population income distribution. Their share in enrollments is higher in public institutions, although the poorest income group is still under-represented: some 22% of all those going into public institutions come from the bottom 40%. Following Reform B, 78% of students in public institutions and 87% of students in private institutions would be replaced by higher ability non-students. The average ability score would increase by 20% in public institutions and by 24% in private institutions. These results indicate that there would be fewer changes in public institutions in Colombia than  - 14 - private institutions as a result of the selection reform. However, the replacement rates and the increase in average ability are still significant in public institutions. As a result of the reform, the proportion of public university students who belong to the poorest 40% of the population would rise slightly from 22.3% to 26.3%. The proportion from the richest 20% would remain virtually unchanged. The picture is very different in private institutions. The proportion of private university students from the poorest 40% of the income distribution would almost double -- from 13.1% to 24.6%. At the same time, the proportion from the richest quintile would fall from 34.4% to 18%. These figures indicate that public universities in Colombia are more successful than private ones in attracting those with the highest ability, regardless of income. On the other hand, private universities cater to the richer income groups, even though they are not as able. This is not necessarily an undesirable outcome, as long as private universities are unsubsidized.  - 15 - Table 3: Selection in Public vs. Private Higher Level Institutions Before Reform After Reform B Public Private Public Private Number of Students 570 395 570 395 Ability Score Mean 51.90 51.24 62.42 63.74 Standard Deviation 7.64 7.73 3.80 3.66 Minimum 26.00 31.00 57.50 59.50 Maximum 73.50 77.00 76.50 77.00 Percent From Public Secondary Schools 77.7 76.7 74.4 72.4 Plercent Female 50.4 50.6 37.5 36.0 Replacement Students Number - - 446 343 % of Total in sector - - 78.3 86.8 % Non-Applicant - - 49.3 50.1 Distribution of Students by Household Income Quintile 1 5.8 2.0 5.1 4.6 2 16.5 11.1 21.1 20.0 3 24.7 16.7 22.1 21.8 4 35.3 35.7 35.8 35.7 5 17.7 34.4 15.8 18.0 All Income Groups 100.0 100.0 100.0 100.0 -: means not applicable.  - 16 - Conclusions Previous work has shown that there could be substantial efficiency gains if ability, rather than willingness to pay private costs, determined which secondary school graduates enter post-secondary education. This paper uses data from Colombia to measure the impact on equity if ability were the sole criterion that determined selection. The main conclusion is that individuals from lower socio-economic strata would gain from such a reform of the selection mechanism. In particular, the proportion of students from the 20-40%ile would rise, and that from the 80-100%ile would decline. The proportions from other income groups would remain roughly the same. This gain in equity is accompanied, however, with a decline in the proportion of females in the student subsample.  - 17 - REFERENCES Bray, M. "High School Selection in Less Developed Countries and the Quest for Equity: Conflicting Objectives and Opposing Pressures," Comparative Education Review, 29 (2): 216-31, 1985. Heyneman, S. and Loxley, W., "The Effect of Primary-School Quality on across 29 High- and Low-Income Countries," American Journal of Sociology, 88 (6): 1162-94, 1983. Mohan, Rakesh, Poverty, Income and Employment in a Developing Metropolis: Consequences of Growth in Bogota and Cali, Colombia, Oxford University Press, forthcoming. Pinera, S. and Selowsky, M., "The Optimal Ability-Education Mix and the Misallocation of Resources within Education," Journal of Development Economics, 8: 111-31, 1981. Psacharopoulos, G. and Loxley, W., Diversified Secondary Education and Development, Baltimore: Johns Hopkins University Press, 1985. Reyes-Posada, Alvaro, "Ayudas Estudiantiles para Financiacion de la Educacion en Colombia: Perspectivas para el Futuro," Ministerio de Educacion, Bogota, 1986. Somerset, H.C.A., "Who Goes to Secondary School? Relevance, Reliability and Equity in Secondary School Selection," in Education, Society and Development, D. Court and D. Ghai (eds.), Nairobi: Oxford University Press, 1984.  - 18 - FOOTNOTES I/ Aside from Examinations, selection is sometimes determined by setting quotas according to such characteristics as gender, geographic residence, and ethnic origin. Such a system might select students with poor academic aptitude while rejecting others with higher aptitude (Bray 1985). 2J Such a hypothetical reform of the system has been suggested by Pierre and Selowsky (1981). However, their investigation focused on efficiency issues and did not use actual data. 3/ For simplicity, the distribution in Figures 1 and 2 are depicted as normal around a mean. The normality assumption is not essential to the results below. 4/ See Psacharopoulos and Loxley (1986) for a more detailed description of the survey and data. 5/ These tests, based on Instituto Colombiano para el Fomento de la Educacion Superior (ICFES) aptitude tests administered to students in secondary school ion 1981, were intended as measures of a person's innate mental proficiency. They are used to predict how well students may learn in whatever field they pursue. Thus, they differ from achievement tests that measure how much a student has learned in a given subject.