BY MAI MIKSIC | A student’s college choice plays a large role in determining his or her future success. An important consideration is whether or not colleges offer sufficient academic rigor to challenge and inspire. High achieving students who attend colleges that underchallenge them are more likely to do poorly, and many eventually drop out (Bowen, Ghingos, & McPherson, 2009; Goodman, Hurwitz, & Smith, 2014). This phenomenon is particularly prevalent among high achieving students from disadvantaged backgrounds (Smith, Pender, & Howell, 2013). It is not surprising, then, that there has been a great deal of research investigating this phenomenon, also known as “undermatching.” A new study by Hill, Bregman, and Andrade (2014) sheds light on what the mechanisms behind what a good match might be and what might alleviate the problem of undermatching.
High achieving students from disadvantaged backgrounds undermatch for many reasons, such as financial constraints. But there are two other factors that contribute to undermatching, both of which are more amenable to intervention: a profound lack of information about the college application process and lack of guidance throughout the application process (Dillon & Smith, 2013). Here, the construct of social capital plays a large role in explaining the undermatching phenomenon. Social capital can play an instrumental role in students’ college choice process, transition to college, and eventual success in college (Gonzalez, Stoner, & Jovel, 2003; McDonough, 1997; Plank & Jordan, 2001).
Scholars do not agree upon a universal definition of social capital, but it can fairly be thought of as having three components: (1) information and resources, that are (2) embedded in social networks, and (3) that individuals can use to achieve specific goals (Bourdieu, 1985; Lin, 1999; Portes, 1998). Hill et al. (2014) used this definition in their new study. In the college choice framework, students’ social capital includes the information and resources they can gather through their social networks (which can include parents, teachers, peers, guidance counselors, etc.) in order to decide which school to attend.
Racial, ethnic, and economic groups have differing amounts of social capital and use and deploy their social capital in different ways (Stanton-Salazar & Dornbusch, 1995). If students from low-income and otherwise disadvantaged backgrounds possess less social capital, it follows that they are more likely to undermatch – which compounds achievement gaps and puts disadvantaged students even further behind in the long term.
Students’ Social Networks and Selective College Enrollment
There is relatively little research on students’ social network composition and how that affects their college choice. We know that the quality of information about the college process matters to students’ college choice, but, does who provides students with information matter?
Hill et al. (2014) investigated these questions and sought to understand how the college choice of students from two urban high schools was influenced by their social network composition. As previously stated, social capital is defined as having three main components, (1) the resources and information (2) embedded in a social network (3) used to achieve a goal. Therefore the authors are only investigating one part of social capital, which is the social network itself.
In order to study this topic, the authors needed a special dataset that would include high achieving urban students of color who were likely to go to college. The authors examined the data from 1998-1999 of two urban magnet schools (previously converted from public schools) with a predominantly Black student population. These schools were chosen because their college enrollment rates were exceedingly high (90% and 70%, respectively), and it would have been more difficult to find a suitable sample size in the traditional public schools. Thus, though the sample for this study fit the needs of the research question, the results may not be applicable to all types of schools.
Despite the schools’ high college enrollment rates, many of the students were from disadvantaged backgrounds; for example, many of the students’ parents had not attended a four-year college. The sample for this study was particularly unique in that the students came from disadvantaged backgrounds and were likely to attend college. A total of 311 high school juniors and seniors were recruited to be a part of the study. Of the sample, 74% were Black or Latino, 20% were White, and 6% identified themselves as being another racial group. These students completed surveys that collected information about their demographic backgrounds, high school experiences, postsecondary plans, and social networks.
The main outcome variable of interest in this study was the selectivity of students’ first college choice. The authors made the assumption, based on previous research, that the students would mostly likely enroll in their first school of choice if admitted. Because of the high level of college enrollment rates at each high school, the authors further assumed a strong likelihood of first-choice acceptance. However, the authors were not able to ascertain whether or not the students actually enrolled in their first choice.
The authors were interested in the selectivity of the students’ first college choice because they assumed that the more selective a college was, the better the match would be for the students. This assumption is based on the fact, again, that these students were relatively high achieving, and that attending a selective 4-year college would increase the social and economic mobility of disadvantaged students, particularly Black and Latino students (Alon & Tienda, 2005; Bowen & Bok, 1998; Brand & Xie, 2010; Dale & Krueger, 2002, 2011; Espanshade & Radford, 2009; Jencks, 1979; Kane & Rouse, 1995; Long & Kurlaender, 2009). Thus, it is important to remember that these authors consider choosing a highly selective college indicative of a “good match.”
Hill et al. (2014) ranked each named college based on its selectivity by using Barron’s College Selectivity Index (2000). Using the Index as a framework, the authors created four main categories in which a college could fall, “least selective,” “selective,” “very selective,” and “most selective,” based on a combination of entering students’ average grades, class rank, SAT or ACT score, and the school’s admission rate. Additionally, the authors added a fifth category in which the student could identify that they had no college plans.
The main independent variable used to predict students’ college choice was the students’ social network composition. In the original data that was collected, students were asked to list up to six people who they relied on for information and guidance on the college admissions process. Students were then asked (1) who was the most influential person in shaping their postsecondary plans; (2) who was the most influential person in shaping their high school choice; (3) the person they relied on the most for choosing their courses; and (4) the person who they relied on the most for the help with the college application process. The current authors then took this data and categorized each person, listed by the student, into four categories; parent, teacher/counselor, peer, and self.
Hill et al. (2014) used a regression technique to predict the selectivity of the student’s first college choice. The regression can tell us whether or not the social network composition of the students is associated (or correlated) with the selectivity of the colleges, but cannot tell us anything about the causal mechanisms. The regressions are an improvement over basic correlations, in that they allow for the use of control variables, which are factors that can be held constant. In these regressions, the authors used race, students’ grade point averages, parent education, student and parental education expectations, the high school the student attended, and the family’s economic resources. Controlling for these variables rules out their influence in order to isolate the effect of social network composition on the students’ college selectivity choice.
The main finding was that the selectivity of students’ college choice was primarily influenced by their peers, and not parents, teachers, or counselors. However, before we discuss these results, we need to examine the descriptive statistics.
First, it may be possible that the composition of a student’s social network may vary based on race or ethnicity. The authors investigated this possibility and found that the composition of the social networks did not vary across race or ethnicity, which meant that all students regardless of race or ethnicity had relatively the same number of parents, peers, and teachers and counselors in their social network.
If we look at just the averages, without controlling for any other factors, students primarily relied on their parents for information (64%), which is surprising considering the fact most of the parents had never attended college. Almost a quarter (22%) of the students reported that they relied on themselves or their peers for information. Finally, 14% of students reported getting information from teachers or counselors.
Crosstabulation (purely descriptive statistics) of the selectivity of students’ first choice school and students’ social network composition showed that students who wanted to go to “the most selective colleges” relied heavily on their parents (83%), 11% of students relied on teachers and counselors, and 6% relied on themselves. Results for students who wanted to go to “very selective colleges” were similar; 69% relied on information from their parents, 7% relied on their peers, 10% relied on their teachers and counselors, and 14% relied on themselves. For students who wanted to go to “selective colleges”, 58% depended on their parents, 15% depended on their peers, 14% on their teachers and counselors, and 14% on themselves. Then, for students who wanted to go to the “least selective colleges”, 67% relied on their parents, 5% on their peers, 12% on their teachers and counselors, and 17% on themselves. Finally, for students who reported that they had no college plans after high school, 56% relied on their parents to make this decision, 9% relied on their peers, 19% on their teachers and counselors, and 15% on themselves. Overall, there were no overwhelming differences between the selectivity of the students’ college choice and whom they went to for information about colleges.
These results, however, are purely descriptive and cannot account for the individual personal differences between students that may influence their school choices and who they go to for information. The regression results are more compelling, because they can hold personal characteristics constant, such as students’ grade point averages and family economic resources, and then predict the relationship between the selectivity of college choice and social network composition.
The main results show that, controlling for a number of factors previously mentioned, there is indeed a relationship between the selectivity of students’ college choice and social network composition, but only for peers. There was no relationship between college selectivity and parents being the dominant source of social capital. This is an important finding, since, as previously mentioned, the descriptive statistics reported that students overwhelmingly relied on their parents for information about which college admissions. Yet, when factors were controlled for, parents did not actually predict whether or not students chose a selective college. Additionally, getting information from teachers and counselors also had no influence over students’ college selectivity choice.
However, students who relied heavily on their peers were most likely to choose more selective schools. This means, according to the assumptions and interpretation of the authors, that students are more likely to make an appropriate college match when they rely on their peers for information and resources. This is a remarkable finding, and a departure from previous research results. But why would this be? The next section discusses the possible explanations for these results.
Unfortunately, the authors did not spend much time trying to explain their results. Absent such a discussion, we are left wondering what could be made of their discoveries.
A remarkable amount of the parent involvement research has been dedicated to the study of how parents can influence their children’s college choices. Hill et al.’s (2014) regression analysis, however, showed that parents had no effect on the selectivity of their children’s college choices. Why is this? One possible explanation is that while parents can influence whether their children go to college at all, they do not influence the selectivity of those colleges.
What about the influence of teachers and counselors? The finding that these other adults had limited influence over the selectivity of students’ college choice is interesting, considering the wealth of knowledge they have about the college admissions process. As with parents, it is possible that teachers and counselors only influence whether, but not where, students go to college. A more compelling answer lies in redefining social capital to take into account the nature of the relationships within a network, not merely the information the network provides.
As previously mentioned, there are contending definitions of social capital. The social capital definition used in Hill et al.’s (2014) work focuses on information conveyed via personal networks. However, Coleman (1988) and other theorists, such as Robert Putnam, define social capital in terms of the relational qualities within a social network such as trust, reciprocity, expectations, and shared values. If we were to relax both definitions and consider them concurrently, we could reach an interesting conclusion: perhaps the relationships these students have with their teachers and counselors lacks the necessary relational qualities (trust, reciprocity, shared values) to influence students’ college choices – despite the information embedded in them.
If this is the case, how do we explain the influence of peers on these students’ decisions to attend selective colleges? Using our expanded definition of social capital, a compelling explanation emerges: it is not perhaps the actual resources and information that peers provide, it is the quality of the relationship between the student and his or her peers. These results are consistent with previous findings on the important role peers play during adolescence, specifically the influence of peers on academic achievement (Ryan, 2000).
Assumptions and Shortcomings
While the results of this study do contribute new information to the research literature, strong and causal conclusions cannot be drawn from the findings. This is because the study relies on too many assumptions, such as the idea that the students would automatically be accepted into and enroll in their first choice of colleges. But there are two central assumptions that call into question the extent to which we can extrapolate findings from this study.
First, the authors assume that all students in the sample are “high achieving,” merely because they attend high schools with strong college enrollment rates. This assumes too much. The study sample did not measure grades, GPA, advanced placement courses, or test scores. The authors noted vaguely that students in these schools had received basic preparation for 4-year college admission, but that does not necessarily mean much when it comes to selective colleges. The authors jumped to another assumption, that the optimal matches for this cohort were the maximally selective colleges. This, too, assumes too much. It is quite possible that students in the study were not necessarily high achieving by specific terms, and that highly selective colleges may not have been the best academic match for them.
Second, it seemed as if this study assumed that all the types of people in the social network provide the same amount and quality of social capital. In real life, parents, teachers and counselors, peers, and students all have access to different amounts and quality of information. The researchers seemed to have been aware of this issue, which is why they separated the social network into subgroups. Yet, they did not make it explicit whether or not they thought the different groups provided the same amount and quality of social capital.
The authors of this study claim that their results have important implications for the undermatching and social capital research literature. However, their inability to isolate an actual “high achieving” student sample means the authors are not actually measuring what they think they are measuring. Furthermore, while their analysis of the influence of social networks is valuable, it is incomplete; they do not interrogate the weight assigned to different members, nor inquire as to why students value certain participants more than others. While the authors did not exactly fail in their mission to answer the research questions of interest, their results cannot be taken very far.
This study contributes to the undermatching and social capital research literature by elucidating the important role of peers in the college selection process. This finding is new and is sure to give researchers plenty to think about. Yet, Hill et al. (2014) offer scanty advice about creating an optimal match between the high achieving urban disadvantaged students and colleges and no advice about how to help low achieving students find an appropriate college match.
Of course, we must remember that matching students to colleges concerns more than academic fit. Financial resources, geography, and social infrastructure once in college, need to be considered as well. In light of the growing student debt problem this nation faces, the college selection process needs to be evaluated carefully and with an eye towards value. Obama’s proposed new college rating system would provide students with information regarding a college’s access (such as the number of students receiving Pell grants), affordability (average tuition cost, scholarships, and loan debt), and outcomes (such as graduation rates). Such a rating system would indeed help students make more informed choices – provided the right players within their social networks deliver the news.
Alon, S., & Tienda, M. (2005). Assessing the “mismatch” hypothesis: Differences in college graduation rates by institutional selectivity. Sociology of Education, 78, 294-315.
Bourdieu, P. (1985). The forms of capital. In J. Richardson (Ed.), Handbook of Theory and Research for the Sociology of Education. New York: Greenwood.
Bowen, W. G., & Bok, D. (1998). The shape of the river: Long-term consequences of considering race in college and university admissions. Princeton, NJ: Princeton University Press.
Bowen, W. G., Chingos, M. M., & McPherson, M.S. (2009). Crossing the finish line: Completing college at America’s public universities. Princeton, NJ: Princeton University Press.
Brand, J. E., & Xie, Y. (2010). Who benefits the most from college? Evidence for negative selection in heterogeneous economic returns to higher education. American Sociological Review, 75, 273-302.
Coleman, J. S. (1988). Social capital in the creation of human capital. American Journal of Sociology, 94(Supplement: Organizations and Institutions: Sociological and Economic Approaches to the Analysis of Social Structure), S95-S120.
Dale, S. B., & Krueger, A. B. (2002). Estimating the payoff to attending a more selective college: An application of selection on observables and unobservables. Quarterly Journal of Economics, 117, 1491-1527.
Dale, S. B., & Krueger, A. B. (2011, June). Estimating the return to college selectivity over the career using administrative earnings data (Working Paper No. 17159). Cambridge, MA: National Bureau of Economic Research.
Dillon, E. W., & Smith, J. A. (2013). The determinants of mismatch between students and colleges (Working Paper 19286). Cambridge, MA: National Bureau of Economic Research. Retrieved from http://www-personal.umich.edu/~econjeff/Papers/Dillon%20and%20Smith%20(2013)%20Determinants%20of%20Mismatch%20Between%20Students%20and%20Colleges%20NBER%20WP%2019286.pdf
Espanshade, T. J., & Radford, A. W. (2009). No longer separate, not yet equal: Race and class in elite college admissions and college life. Princeton, NJ: Princeton University Press.
Gonzalez, K. P., Stoner, C., & Jovel, J. E. (2003). Examining the role of social capital in access to college for Latinas: Toward a college opportunity framework. Journal of Hispanic Higher Education, 2, 146-170.
Goodman, J., Hurwitz, M., & Smith, J. (2014). Initial college choice and degree completion rates: Using admissions test score thresholds to estimate undermatch penalties. Retrieved from http://www.aefpweb.org/sites/default/files/webform/39th/JGP%2003-07-14.pdf
Hill, L. D., Bregman, A., & Andrade, F. (2014). Social capital for college: Network composition and access to selective institutions among urban high school students. Urban Education. Advanced online publication. doi: 10.1177/0042085913514590
Jencks, C. (1979). Who gets ahead? The determinants of economic success in America. New York, NY: Basic Books.
Kane, T. J., & Rouse, C. E. (1995). Labor-market returns to two- and four-year college. The American Economic Review, 85, 600-614.
Lin, N. (1999). Building a network theory of social capital. Connections, 22(1), 28-51.
Long, B. T., & Kurlaender, M. (2009). Do community colleges provide a viable pathway to a baccalaureate degree? Educational Evaluation and Policy Analysis, 31, 30-53.
McDonough, P. M. (1997). Choosing colleges: How social class and schools structure opportunity. Albany, NY: SUNY Press.
Plank, S. B., & Jordan, W. J. (2001). Effects of information, guidance, and actions on postsecondary destinations: A study of talent loss. American Educational Research Journal, 38, 947-979.
Portes, A. (1998). Social capital: Its origins and applications in modern sociology. Annual Review of Sociology, 24, 1-24.
Ryan, A. M. (2000). Peer groups as a context for the socialization of adolescents’ motivation, engagement, and achievement in school. Educational Psychologist, 35(2), 101-111.
Smith, J., Pender, M., & Howell, J. (2013). The full extent of student-college academic undermatch. Economics of Education, 32, 247-261.
Stanton-Salazar, R. D., & Dornbusch, S. M. (1995). Social capital and the reproduction of inequality: Information networks among Mexican-origin high school students. Sociology of Education, 68, 116-135.