Part 2 (1/2)

While students choose specific fields of study, previous research also suggests that faculty engage in distinct practices across fields. Faculty members in different disciplines value specific domains of knowledge and forms of interaction. Consequently, they structure courses, interact with students, and emphasize and reward distinct interests, abilities, and competencies.46 They also differentially encourage specific educational practices such as faculty-student contact or engagement in active learning, and they are more or less likely to communicate high expectations to students.47 In other words, faculty in different fields create distinct socializing environments which foster development of specific skills, att.i.tudes, and values.48 Perhaps as a product of these socializing experiences (and individual selection into fields), students exhibit different acultures of engagement.a Students in social science and humanities demonstrate engagement by talking to professors outside of cla.s.s, contributing to cla.s.s discussion, and asking questions in lectures. Students in science and engineering fields, on the other hand, place higher value on quant.i.tative skills and cla.s.ses that help to solve problems, and they engage much more with peers (by studying together and helping each other understand course material).49 Given these findings, it may not be surprising that students in some fields perform higher on the CLA than others. The CLA measures a specific set of skillsa”namely critical thinking, complex reasoning, and writinga” that is far from the totality of learning or the full repertoire of skills acquired in higher education. As students both select and are socialized into specific ways of knowing and thinking, they will perform well on the CLA to the extent that their disciplines emphasize the skills a.s.sessed. Moreover, some fields may be focusing more on oral than written communication, and thus while students may be acquiring critical thinking skills, they may not demonstrate them as readily in the written format. And even when students in certain fields do not perform as well on the CLA, this does not mean that they are not gaining valuable skills in other areas. In part, this is why there are certification requirements in fields such as education and health that require students to demonstrate knowledge in an occupationally specific domain.

It would be easy to conclude that students in different fields focus on different sets of skills, only some of which are captured by the CLA, thus leading to the observed differences in CLA performance. However, faculty members across subjects overwhelmingly agree that critical thinking, complex reasoning, and writing are key skills to be taught in higher education. One could hardly argue that we would not want teachers who are educating our children, or business majors who might be responsible for approving home mortgage loans, to develop the capacity to think critically or reason a.n.a.lytically. Moreover, health majors perform significantly higher on the CLA than do business majors, although both are applied fields. Differences across fields may become more p.r.o.nounced as students immerse themselves more deeply into their chosen majors in the second two years of college. However, presented findings at least raise the question about the extent to which, despite the importance and value of specific skills for different fields, general skills such as critical thinking, complex reasoning, and writing could and should be developed across the undergraduate curriculum.

Financing College Education and Learning.

Financing higher education is a persistent worry for students, parents, and policymakers. Given the high and rising costs of college, the decreasing availability of adequate grant support, and an increasing reliance on loans, not to mention the current economic crisis, the question of how to pay for college is a constant source of discussion and concern. The research community has partic.i.p.ated in this dialogue by aiming to understand whether certain sources of funding are related to student outcomes, namely persistence and degree attainment. Much of the debate has focused on the role of financial aid, including merit versus need-based aid, in facilitating degree completion. Although many articles have been published on the topic, it is difficult to ascertain the true effects of financial aid because students who receive different types of aid often differ on a number of important but hard-to-measure characteristics. A large-scale experimental study currently conducted by Sara Goldrick-Rab and Douglas Harris at the University of Wisconsina”Madison is poised to provide some more definitive insights into the consequences of financial aid for student outcomes in the near future.

Notwithstanding the debates about causality, which we cannot engage given the observational nature of our study, we explored the relations.h.i.+p between sources of funding (namely grants / scholars.h.i.+ps and loans) and studentsa performance on the CLA. We asked students to indicate the percentages of their college costs that were covered by grants, scholars.h.i.+ps, and loans. The results indicate that grants / scholars.h.i.+ps have a positive a.s.sociation with learning while loans have no relations.h.i.+p. Figure 4.6 reports the predicted 2007 CLA scores for hypothetical students covering between 0 and 100 percent of their college costs with grants / scholars.h.i.+ps or loans. These predictions are adjusted for 2005 CLA scores as well as a range of individual characteristics and inst.i.tutions attended. Students with higher proportions of college costs covered with grants / scholars.h.i.+ps have higher predicted 2007 CLA scores. Comparing students at the extreme, a student who covered all of his or her college costs with grants / scholars.h.i.+ps would score 45 points higher on the 2007 CLA than a student who received no grants or scholars.h.i.+ps. There is no relations.h.i.+p between the proportion of college costs covered with loans and CLA performance. While the gray bars in figure 4.6 slope slightly downward, this trend is relatively weak (approximately half the size of the trend for grants / scholars.h.i.+ps) and not statistically significant.

One way in which different sources of funding could be related to student outcomes is through their relations.h.i.+p to other activities, particularly work. Financial aid packages are often constructed to include employment components, whether through the federal work-study program or various inst.i.tutional programs. Students who request and are eligible for financial a.s.sistance may thus have specific employment obligations included in their financial aid packages. Moreover, since financial aid rarely meets the full cost of attending college, students may seek to work additional hours to cover the difference. College students may also work in order to avoid borrowing. Our a.n.a.lyses reveal that sources of funding are indeed related to hours worked on and off campus. The higher the proportion of college costs covered through grants / scholars.h.i.+ps, the more time students spend working on campus and the less time they spend working off campus. In contrast, the higher the proportion of college costs covered through loans, the more time students spend working off campus. Financing college education through loans is positively related to working on campus as well, but that relations.h.i.+p is weaker than the relations.h.i.+p between grants / scholars.h.i.+ps and working on campus.50 Relations.h.i.+ps between studentsa estimates of college funding sources and hours worked, however, are relatively weak in our sample. Moreover, different forms of employment and different college funding strategies are related to studentsa social background and academic preparation. When we include on-and off-campus employment in the a.n.a.lysis in addition to other individual-level characteristics, the relations.h.i.+p between grants / scholars.h.i.+ps and CLA growth on the one hand and loans and CLA growth on the other does not change notably (see table A4.4 in methodological appendix). In the final a.n.a.lysis, after statistically adjusting estimates for individual characteristics and inst.i.tutions attended, employment during college does not appear to be related to CLA growth. The percentage of college costs covered through grants / scholars.h.i.+ps, however, continues to have a positive a.s.sociation with studentsa learning. While not definitive, these findings point to an area deserving further investigation. Previous research has focused on examining the relations.h.i.+p between financial aid and persistence / attainment; our a.n.a.lyses suggest that learning is another outcome worthy of examination.

Gaps in CLA Growth between African-American and White Students Having shown that specific student experiences facilitate learning in higher education, we return to the concern regarding differences in CLA growth between African-American and white students. In chapter 2, we carefully examined differences in learning gains across different groups of students, focusing in particular on students from different family backgrounds and racial / ethnic groups. Results showed that differences in the social contexts in which students grew up and their academic preparation explained the gaps in CLA growth between students from more and less educated families. However, those factors were not adequate to account for the gap in learning between African-American and white students. Although African-American students on average were more likely to attend schools that had 70 percent or more minority students, took fewer advanced placement courses, and performed less well on college admission tests, these differences only partially explained their lower rate of progress on skills measured by the CLA during the first two years of college. If factors prior to college entry could not explain the gap, could experiences during college provide some insights into the differential growth rates between these two groups?

The first bar in figure 4.7 reports the gap in 2007 CLA scores between African-American and white students, statistically adjusted for studentsa sociodemographic and high school characteristics, academic preparation, and skills at entry into higher education (i.e., 2005 CLA scores). Even after these adjustments, African-American students scored 47 points lower on the CLA at the end of their soph.o.m.ore year than did white students. Next, we include studentsa college experiences in a.n.a.lysis, which slightly increases the gap in learning between African-American and white students. This pattern emerges through a complex combination of differential experiences. The primary factors driving the increase in the gap between African-American and white students are hours spent in fraternities, percent of college cost covered by grants and scholars.h.i.+ps, and college major. These are areas in which African-American students have more positive educational experiencesa”that is, experiences more conducive to learning. They spent fewer hours on average in fraternities and sororities, had a higher proportion of their college costs covered by grants and scholars.h.i.+ps, and were more likely to major in some fields with higher growth in learning (such as health). These positive experiences are partly countered by exposure to contexts less conducive to learning. For example, African-American students reported lower faculty expectations and demandsa”less than a third of African-American students had taken cla.s.ses that required them to both read more than forty pages a week and write more than twenty pages over the course of a semester. These negative experiences do not entirely offset the positive ones, thus leading to the increase in the CLA gap between African-American and white students after including college experiences in the a.n.a.lysis.

Considering inst.i.tutions attended reveals a different pattern. As reported in chapter 2, African-American students are less likely than their white peers to attend highly selective or selective inst.i.tutions. Colleges attended by African-American and white students may also differ on other inst.i.tutional characteristics. After controlling (i.e., statistically adjusting estimates) for inst.i.tutions attended, the gap in learning between African-American and white students decreases by 15 percent. While this reduction may not appear substantial, it is remarkable given that our a.n.a.lyses already control for a host of individual characteristics and college experiences. African-American students are thus disadvantaged by attending colleges and universities that are less effective at facilitating studentsa development of critical thinking, complex reasoning, and writing skills. These findings highlight once again the importance of college experiences as well as inst.i.tutions attended, not only for overall learning but also for inequality in learning between African-American and white students.

In the final a.n.a.lysis, after we adjust the CLA performance for a range of individual attributes, college experiences, and inst.i.tutions attended, we could explain almost two-thirds of the gap in learning between African-American and white students. Although this is a notable feat, the gap between the two groups remains sizable and statistically significant. Moreover, although some recent studies have suggested acompensatory effectsa of college experiences, indicating that students who enter college less advantaged gain more from positive experiences, we find no such evidence regarding growth in the CLA for African-American versus white students.51 Persistent gaps in test performance between African-American and white students have been reported at other grade levels as wella”no matter what controls are included in statistical a.n.a.lyses, the gaps persist. This pattern has led some authors to consider the role of more subtle cultural factors in producing the gaps in academic achievement between African-American and white students. 52 A prominent early theory argued that due to the long history of discrimination and inequality, African-Americans have developed an oppositional culture, which defines behaviors and traits appropriate for the group in opposition to the dominant white culture. In the context of schooling, the argument goes, the oppositional culture of African-American adolescents has led them to eschew academic achievement. Strong peer pressures and accusations of aacting whitea are argued to be keeping African-American students from doing well academically.53 The empirical evidence for this theory is weak, and recent decades have presented multiple challenges to the oppositional culture argument.54 However, this does not necessarily mean that culture is inconsequential for educational success. Instead of thinking about it as shaping preferences, we may want to think about culture as a atool kita of habits, skills, and styles.55 Since African-American and white students grow up in different contexts, sociologist Douglas Downey has argued that athe skills, habits, and styles blacks are exposed to are, on average, less useful for academic success.a From this definition of culture, the key to understanding academic performance does not lie with African-American att.i.tudes about schooling but with the social isolation of African-American adolescents.56 An alternative explanation for the differences in academic achievement between African-American and white students that has received increasing support in recent decades is termed astereotype threat.a Emanating from the work of psychologists Claude Steele and Joshua Aronson, this argument proposes that African-American adolescents are well aware of the negative stereotypes regarding their groupsa academic achievement. Whenever they are called on to perform academically, they face astereotype threata or the fear of confirming the negative stereotype. This fear leads to their lower performance.57 Even subtle cues like being asked to identify oneas race before a GRE-like verbal test can lead African-American students to perform less well.58 Recent a.n.a.lyses of college students at selective inst.i.tutions have supported this argument. The researchers found that stereotype threat is related to studentsa grade-point average, and that controlling for it helps to reduce the gap in academic performance between African-American and white students.59 Although we cannot test these propositions given our data, more subtle mechanisms, resting on differences in acultural tool kitsa and / or the threats of confirming negative stereotypes, deserve further study to advance our understanding of the inequality in academic achievement.

Variation Across and Within Inst.i.tutions.

While students do not learn much on average, this chapter has illuminated how specific activities and experiences during college can either facilitate or thwart learning, creating variation among students and inst.i.tutions. Twenty-nine percent of variance in 2007 CLA scores is found across inst.i.tutions. Even if we focus specifically on growth (estimating 2007 CLA scores while controlling for the 2005 scores), 20 percent of the variance is found across inst.i.tutions. Some of it is a.s.sociated with the sorting of students into inst.i.tutions (i.e., inst.i.tutions enroll students with different characteristics, such as different levels of academic preparation). However, even if we control for a range of background characteristics, including race / ethnicity, socioeconomic background, academic preparation, and 2005 CLA scores, students in some inst.i.tutions experience larger gains on average than others. The same finding has been reported in previous research with respect to other outcomes. For example, while student characteristics and school resources are important predictors of degree attainment, there are notable differences in graduation rates across inst.i.tutions, even after accounting for many of these ainputa characteristics.60 If we select top-performing inst.i.tutionsa”inst.i.tutions that show much larger gains on the CLA than others, net of individual characteristicsa” we find, not surprisingly, that their students report higher incidence of behaviors that are beneficial for learning (figure 4.8).61 Students at these inst.i.tutions report greater course requirements: almost two-thirds (62 percent) of their students reported taking courses that required both reading more than forty pages a week and writing more than twenty pages over the course of a semester. The average for other inst.i.tutions is just over one-third (39 percent). In a finding perhaps related to higher coursework demands, students at high-performing inst.i.tutions also spent more time studying, particularly alonea”almost three more hours of solitary study than students at other inst.i.tutions. Three hours is a remarkable difference, considering that students overall on average spent less than nine hours studying alone. Since we have only 24 inst.i.tutions in the sample, some of which have relatively small sample sizes, we are not able to delve deeply into inst.i.tutional differences. However, even this brief discussion indicates that inst.i.tutions differ in the extent to which they create contexts which facilitate positive behaviors and actions a.s.sociated with learning.

Previous studies of inst.i.tutional characteristics and practices substantiate these findings by showing that inst.i.tutions vary notably in how they structure student experiences. George Kuh and his colleagues, for example, conducted in-depth studies of twenty four-year colleges and universities that had higher than predicted graduation rates and higher than predicted levels of student engagement (based on the National Survey of Student Engagement). Among other characteristics, these inst.i.tutions had an aunshakeable focus on student learning.a Their emphasis on undergraduate learning was manifested in a range of practices, from inst.i.tutional openness to new and experimental instructional techniques to faculty investing more time in students and taking greater responsibility for them, as well as showing greater commitment to both providing and receiving feedback.62 Moreover, although many existing college programs focus exclusively on retention, some have potential to facilitate learning. Learning communitiesa”programs that enroll groups of students in a common set of courses and are frequently linked with residence life experiencesa”have shown positive a.s.sociation with a range of student outcomes including persistence, grades, and self-reported learning.63 Researchers have yet to evaluate the effects of learning communities on standardized objective measures of learninga”this is an important area of future research, as these programs are poised to facilitate persistence as well as learning.

While there is variation in student performance across colleges and universities, it is important to note that there is even more variation within inst.i.tutions. This is the case for most educational outcomes, and it has been extensively doc.u.mented with respect to student engagement in higher education.64 High-as well as low-performing students are found at all inst.i.tutions. If, for example, we consider students in the top 10 percent of the CLA growth distribution, we would find them at each of the inst.i.tutions.65 This is remarkable, given that these students are experiencing more than 1.5 standard deviation of growth between the beginning of their freshman and end of their soph.o.m.ore year, which is more than eight times the average growth. Exploring variation within inst.i.tutions highlights the often ignored and untapped potential for improvement. Even at the highest performing schools there is room for growth, as not all students are performing equally well. And even colleges that are struggling have students who spend time studying and make notable progress in critical thinking, complex reasoning, and writing skills during their first two years. Given our sample-size limitations, we cannot provide a detailed account of what students at each inst.i.tution look like and what inst.i.tutions are doing to facilitate their learning. However, each inst.i.tution can look within, as opposed to only looking across, to learn what works and what does not. High-performing students within inst.i.tutions can serve as guides for thinking about and implementing meaningful change.

Focusing on Learning in Higher Education.

Learning is a complex processa”and thus, not surprisingly, myriad factors shape what and how much students learn in higher education. To make matters more challenging, many of these factors are related, such that studentsa backgrounds and academic preparation are related to the inst.i.tutions they attend and their specific experiences within those inst.i.tutions. We have aimed to untangle these different influences to the extent possible with our observational data, and to provide some insights into which factors may lead to greater growth in critical thinking, complex reasoning, and writing during the first two years in college.

Putting it all together, we present results from the final model (see table A4.5 in methodological appendix), which includes all relevant factors discussed throughout the chapter. The final model includes only college experiences that were deemed influential in preceding a.n.a.lyses. The overall framework representing this final a.n.a.lysis is ill.u.s.trated in figure 4.9. What students bring to college matters; this is particularly the case with respect to their academic preparation. However, our primary focus in this chapter has been on what students do while they are in college, as those a.n.a.lyses help us illuminate the direction for improving higher-education policy and practice in the future. These final a.n.a.lyses confirm the results discussed in the preceding pages, reaffirming the importance of studentsa college experiences and inst.i.tutions attended for their intellectual development.

What students do in higher education matters. But what faculty members do matters too. Faculty are most directly involved in shaping student experiences, although the support and incentives advocated by their deans, provosts, and presidents will influence whether and how they engage in activities that facilitate student learning. There are some clear examples of how faculty members may shape student actions and, by extension, their learning. For example, college GPA is positively related to the 2007 CLA scores.66 This indicates that faculty members indeed reward critical thinking, complex reasoning, and writing skills in the cla.s.sroom. The relations.h.i.+p is not perfect but that is to be expected, as not all cla.s.ses are likely to focus on the skills captured by the CLA. Nevertheless, this positive a.s.sociation indicates a potential for considering how those skills, which on surveys faculty report should be crucial components of undergraduate education, can be taught and rewarded in college cla.s.srooms across the nation.

Moreover, as we have seen, when faculty have high expectations and expect students to read and write reasonable amounts, students learn more. In addition, when students report that they have taken a cla.s.s in which they had to read more than forty pages a week and write more than twenty pages over the course of a semester, they also report spending more time studying: more than two additional hours per week than students who do not have to meet such requirements.67 Thus, requiring that students attend to their cla.s.s work has the potential to shape their actions in ways that are conducive to their intellectual development.

While we have reported relations.h.i.+ps between specific college experiences and learning, one may still wonder how much those factors really matter. One way to address this question is to evaluate the magnitude of the relations.h.i.+ps, which we have aimed to do by presenting predicted CLA scores across different dimensions of college experiences. Another approach is to consider the proportion of variance in the CLA scores that is explained by a different set of factors. The final a.n.a.lysisa”which includes all background measures, college experiences, and inst.i.tutions attendeda” explains 42 percent of the variation in CLA scores. This is a substantial amount by social science standards, although it does imply that much more research is needed to understand the remaining variance. Within our a.n.a.lyses, college experiences and inst.i.tutions attended explained an additional 6 percent of the variance, after controlling for academic preparation and other individual characteristics.68 While that may appear to be a small contribution, academic preparation, which has received much attention in research and policy circles, explains only an additional 8 percent of the variance beyond studentsa background characteristics.69 These estimates may seem low, but this is because of our a.n.a.lytic strategy: we are focusing on growth and are thus controlling for 2005 CLA scores, which, as would be expected, explain the largest portion of the variance in 2007 CLA performance. Thus, studentsa college experiences and inst.i.tutions attended make almost as much of a difference as prior academic preparation. If the blame for low levels of critical thinking, complex reasoning, and writing skills of college students is to be placed on academic preparation, then almost an equal amount of responsibility rests with what happens after students enter higher education.

The a.n.a.lyses presented in this chapter illuminate the multiple actors contributing to the current state of limited learning on college campuses. Faculty members are perhaps the easiest to blame, as in some inst.i.tutional settings they are often tempted to focus greater attention on research and other professional demands than on teaching; this reality presents a concrete set of practices that can be critiqued. And many higher-education inst.i.tutions indeed deserve criticism for failing to focus adequately on the core mission of higher education: educating the next generation. Beyond faculty offices and tenure review procedures, however, there are students, who spend far more time socializing than studying.70 Given the little time they spend studying, it is no surprise that they are not learning much on average. This is partly a consequence of lax demands and expectations, but it is wishful thinking to imagine that simply increasing faculty demands will produce greater learning. aCurrent cultural norms among U.S. undergraduates support a conception of schooling as an important, but part-time activity. Other parts of life, notably, social and leisure activities, are at least as important,a sociologist Steven Brint recently observed.71 Judging from studentsa use of time, we find that these social and leisure activities appear much more important than academic pursuits. The college experience is perceived by many students to be, at its core, a social experience.72 The collegiate culture emphasizes sociability and encourages students to have funa”to do all the things they have not had a chance to do before, or may not have a chance to do after they enter athe real worlda of the labor market. Faculty, administrators, policy makers, and parents are all implicated to a certain extent in accepting or at least partly acquiescing to contemporary collegiate culture.

5.