Introduction

Intercollegiate athletics at colleges and universities have been referred to as “American higher education’s ‘peculiar institution.’ Their presence is pervasive, yet their proper balance with academics remains puzzling” (Thelin, 1994, p. 1). For a host of reasons, scrutiny of this “peculiar institution”—intercollegiate athletics—has intensified in recent years. Such articles as “Jock Majors” (Suggs, 2003a) and “Grades and Money” (Suggs, 2003b) lament that academics and athletics are out of balance. The recent scandal at the University of Colorado, Boulder (Jacobson, 2004) highlights the almost weekly reports of problems in athletic departments across the country. Calls for reforms (e.g., Bowen and Levin, 2003) are coming from inside and outside the academy as well as from the National Collegiate Athletics Association (NCAA) national office. Even federal legislation has been contemplated to bring intercollegiate athletics back into proper perspective.

Some of the most scathing and influential critiques of college athletics are by Shulman and Bowen (2001) and Bowen and Levin (2003). Their research suggests that student-athletes routinely receive preferential treatment in the admissions process and are more likely to be academically under-prepared than their peers. As a result, student-athletes earn lower grades in college. Additionally, they argue that institutions allow student-athletes to create their own subculture and that it flourishes, isolated and insulated from the larger campus culture.

Previous Research on the College Experience of Student-Athletes

The findings from studies other than Shulman and Bowen (2001) and Bowen and Levin (2003) are somewhat less pejorative in terms of the effects of participating in intercollegiate athletics on the quality of the undergraduate experience. For example, competing in intercollegiate sports appears to have little influence on such college outcomes as learning for self-understanding, higher-order cognitive activities, and motivation to succeed academically (Wolniak, Pierson, and Pascarella, 2001). Other studies reveal no differences between student-athletes and non-athletes with regard to cognitive development (Pascarella, Bohr, Nora, and Terenzini, 1995; Terenzini, Pascarella, and Blimling, 1996), grades in college (Hood, Craig, and Ferguson, 1992), or time devoted to studying or attending class (Richards and Aries, 1999). For example, Richards and Aries found no significant difference in GPA between athletes and non-athletes despite the fact that athletes entered college with significantly lower SAT scores. But other studies, like the work of Shulman and Bowen (2001) and Bowen and Levine (2003), report that student-athletes competing in Division III athletics at Ivy League institutions perform at lower levels academically than non-athletes. Such differences are less evident for female student-athletes and student-athletes in non-revenue generating sports (Pascarella et al., 1999) as compared to male student-athletes and athletes playing revenue-generating sports, such as football and men’s basketball.

The NCAA suggests that colleges and universities are to maintain an environment in which the athlete and the athletics program play an essential role in the student body, thereby creating social bonds between and among student-athlete and non-athlete (National Collegiate Athletic Association [NCAA], n.d.). Yet, concerns remain that participating in intercollegiate athletics may lead to social isolation (Riemer, Beal, and Schroeder, 2000; Wolf-Wendel, Toma, and Morphew, 2001). For example, spending time with teammates may strengthen bonds between athletes, but limit interaction with non-athletes (Wolf-Wendel et al., 2001). Despite this possible isolation, numerous researchers have reported that student-athletes were often more satisfied and involved than their non-athlete peers (Astin, 1993; Pascarella and Smart, 1991; Ryan, 1989). In addition, the findings are mixed in terms of whether participating in athletics is negatively associated with self-understanding and openness to diversity (Wolniak et al., 2001), or whether student-athletes interact effectively with people from diverse backgrounds because “athletes compete with and against people from socioeconomic, racial and ethnic, and religious backgrounds other than theirs” (Wolf-Wendel et al., 2001, p. 385).

Given their demanding training and practice routines, it’s not surprising that student-athletes devote significantly more time to extracurricular activities than members of other groups and have acquired valuable time management skills (Richards and Aries, 1999). Even so, in his study of NCAA Division III male basketball players at a small, private, liberal arts college, Schroeder (2000) concluded that athletes were highly engaged in their academics, spending an average of 15 hours per week studying with the majority earning grade point averages exceeding 3.0.

While the harsh critiques of Shulman and Bowen (2001) and Bowen and Levin (2003) may apply to some athletics programs and institutions, not enough is known about the extent to which intercollegiate athletes devote time and energy to activities that are empirically linked to desired outcomes of college. Most of the previous work on the collegiate experiences of student-athletes focuses on a small segment of higher education drawing from elite Ivy League colleges (Bowen and Levin, 2003; Shulman and Bowen, 2001) or the experiences of athletes at only a small number of institutions (Pascarella et al., 1995, 1999; Wolniak et al., 2001). Thus, it is difficult to get a clear, definitive grasp of the nature and quality of the undergraduate experiences of intercollegiate athletes across a large number of colleges and universities and whether athletes are, as some claim, shortchanged in terms of learning and personal development.

Student Engagement

According to Pascarella and Terenzini (2005), the impact of college is largely determined by the degree to which students engage in various in-class and out-of-class activities. Student engagement is a function of both the individual effort of each student and institutional practices and policies that encourage students to participate in educationally purposeful activities (Astin, 1993; Kuh, 2001; Kuh et al., 2005; Pace, 1984, Pascarella and Terenzini, 1991, 2005). Such activities range from traditional academic pursuits, such as reading and writing, preparing for class, and interacting with instructors about various matters as well as other activities considered important outcomes of college, such as learning how to effectively collaborate with peers on problem solving tasks and working productively together in community service activities (Kuh, 2001). Thus, participating in educationally purposeful activities directly influences the quality of students’ learning and their overall educational experience. In addition, “students who are involved in educationally productive activities in college are developing habits of the mind and heart that enlarge their capacity for continuous learning and personal development” (Kuh, 2003, p. 24). Edgerton and Shulman (2002, p. 4) extend the concept of engagement, proposing it as a desired end in itself as well as an indirect indicator of learning (see also Shulman, 2002):

We need to learn more about the forms and conditions of engagement that relate to student competence and commitment in arenas of practice. There are important questions, too, about engagement not as a means to an end...but as an experience worth having in itself. We go to the symphony, after all, not to improve ourselves but to hear the music, to have the experience. Similarly, there are aspects of the college experience—participating in a seminar, for instance, or a role in student governance—that have a kind of value we have not yet learned to describe in detail or to document.

Purpose of the Study

This study compares the engagement of student-athletes in effective educational practices with that of their non-athlete counterparts by using a national sample of undergraduates. Three research questions, derived from the theories that underpin the construct of student engagement, guide the study:

  1. 1.

    How do the educational experiences of student-athletes compare with those of non-athletes?

  2. 2.

    If the experiences of student-athletes differ from those of non-athletes, do these differences vary significantly by institution attended?

  3. 3.

    What is the relationship between the level of competition (NCAA division, NAIA membership) and engagement in good practices in undergraduate education, perceptions of the campus environment, and self reported gains?

Methods

Data Source and Instrument

We use data from The National Survey of Student Engagement (NSSE) to compare the collegiate experiences of student-athletes with those of their non-athlete peers. Pace’s (1984) concept of quality of student effort, Chickering and Gamson’s (1987) “Seven Principles of Good Practice in Undergraduate Education,” and Astin’s (1984) theory of student involvement serve as the foundation from which the conceptual framework of NSSE was drawn.

The NSSE database is especially well-suited for this study because as previously discussed, research on the college student experience indicates that students learn more when they are engaged at reasonably high levels in a variety of educationally purposeful activities (Astin, 1984; Kuh et al., 1991; Kuh, Douglas, Lund, and Ramin-Gyurnek, 1994; Pascarella and Terenzini, 1991, 2005; Tinto, 1987).

NSSE was specifically designed to assess the extent to which students engage in empirically derived good educational practices and what they gain from their college experience (Kuh, 2001). Although NSSE does not assess student learning outcomes directly, the main content of the NSSE instrument, The College Student Report, represents student behaviors that are highly correlated with many desirable learning and personal development outcomes of college.

NSSE collects information directly from random samples of first-year and senior undergraduates at four-year institutions. Because of the potential bias introduced among student-athletes due to attrition to the senior year, this study focuses only on first-year students. Table 1 displays student-athlete status,1 gender, and athletic division for the sample used for the study. Included in the sample are 57,308 undergraduate students—7,821 student-athletes and 49,407 non-athletes—who completed NSSE in the spring of 2003. Students from 395 four-year colleges and universities were represented. Of these institutions, 107 are NCAA Division I, 93 are NCAA Division II, 145 are NCAA Division III, and 50 are NAIA schools.

Table 1. Students by Athlete Status and Athletic Division
Full size table

Data Analysis

We conducted a series of hierarchical linear models (HLM) to explore the effects of being a student-athlete on the collegiate experience. The dependent variables fall into four categories: student engagement, perceptions of campus environment, self-reported gains, and grade point average. Student engagement is measured using three scales: (1) level academic challenge, (2) student–faculty interactions, and (3) active and collaborative learning (see Appendix A for a listing of the constructs used as dependent variables in the models and the items contributing to each measure). The perceived campus environment category includes two measures: a supportive campus environment scale and an overall satisfaction with college scale. Students’ gains in learning and intellectual development are represented by three scales: gains in personal and social development, gains in general education knowledge, and gains in practical competencies. Our final set of models predicts student-reported grades.2

Pascarella et al. (1995, 1999) and Wolniak et al. (2001) have noted that studies collegiate experiences of student athletes should account for both individual background characteristics (e.g., race/ethnicity, social origins) and institutional contexts (e.g., NCAA division). This follows the logic of Astin (1993), Kuh (2001), Kuh et al. (2005), Pace (1984), and Pascarella and Terenzini (1991, 2005), that student engagement is both a function of the individual effort of each student and institutional practices and policies. However, as we have noted, previous research studying student athletes often deals with small numbers of institutions preventing researchers from making any confident claims about the effect that institutions might have on the college experiences of college athletes.

Additionally, much of the previous work may have used inappropriate statistical procedures to estimate college effects. In most studies of organizational or institutional effects, researchers must decide about the appropriate unit of analysis (Hu and Kuh, 2004; Raudenbush and Bryk, 2002). Should they build regression models by aggregating to the institution level, or should they attach institution-level characteristics to individuals? If researchers build models at the institution level, they are prone to the “ecologically fallacy,” whereby individual differences are masked (Hu and Kuh, 2003; King, 1999; Kreft and De Leeuw, 1998). For example, an analysis aggregated to the college level might reveal that athletes at Division III institutions are more engaged in effective educational practices than students at Division I institutions; yet an analysis of Division I colleges might reveal that athletes are as engaged or more engaged than their peers. Because we are especially interested in the differential impact of individual institutions on the experiences of student-athletes, we must model “nested data structures.” According to Raudenbush and Bryk (2002), HLM provides the only accurate way to estimate institutional and individual effects when analyzing nested data.

Studies where institution-level characteristics are attached to an individual also may be flawed (Ethington, 1997). First, such efforts violate the general assumption of ordinary least squares regression (OLS): observations are independent of one another. Second, they assume that individuals within a group are affected identically by group-level characteristics. Finally, the inclusion of group-level variables into an OLS regression equation leads to poorly estimated standard errors and inaccurate numbers of degrees of freedom, increasing the likelihood of committing a Type II error (i.e., two variables are different from one another at a level of statistical significance, when they are not).

Using HLM overcomes the problems associated with unit of analysis by simultaneously modeling both individual and institutional effects. HLM models individual-level and group-level variables simultaneously, resulting in more accurate parameter estimates, making it possible to determine what is an individual-level effect or a group-level effect. Because these effects can be partitioned, each can be modeled with their respective characteristics.

HLM also allows the intercept to vary, thereby partitioning the variance between the institution and the student. In other words, we are able to accurately attribute the variance associated with the student and the variance associated with the institution. Because we hypothesize that what colleges do affects engagement, we allow the intercept to vary and model it using institutional characteristics. Additionally, we also seek to test whether the experiences of student-athletes are different at different college campuses; therefore we can allow the athlete slope to vary by institution. We grand mean center all of our level one and level two independent variables with the exception of the student athlete variable, which we group mean center. By allowing the student-athlete slope to vary and centering it around the group mean, the coefficient for student-athletes then represents the average institutional difference between student-athletes and non-athletes. If the athlete effect varies significantly by institution, we can then model the average athlete differential with institutional characteristics.

For each of the dependent variables, we estimated separate models for men and women. In the first set of models, the within institution models (where we only model student characteristics), we examine whether student-athletes differ, on average, from their non-student-athlete peers on the outcomes of interest. We allow both the intercept and athlete slope to vary but do not include any institutional controls that predict either the student-athlete slope or the intercept. The model controls for many of the student characteristics as covariates (age, race, gender, transfer, grades, Greek, major, part-time status, residing on campus, parents’ education)3 that Pascarella et al. (1995, 1999) and Wolniak et al. (2001) have cited important when study college athletes. Table 2 shows the descriptive statistics of variables included in our models.

Table 2. Descriptive Statistics for Variables Included in Models
Full size table

The second set of models represents average institutional differences (intercept), and in some cases the student-athlete slope, using institutional characteristics. In terms of institutional context, we created dummy-coded variables for the four athletic divisions (Division I, Division II, Division III, and NAIA) and included it at level two to determine whether student experiences differ by division. Division III was designated as the omitted group. Due to constraints presented by multicollinearity between athletic division and other institutional characteristics (e.g., size, selectivity, Carnegie Classification), we include athletic division in the intercept and slope models. Athletic division may be associated with certain correlates of engagement, such as institutional size mentioned earlier. However, because this study focuses on the experiences of student athletes, we propose that athletic division is a more salient institutional characteristic than size and is especially relevant for those responsible who make policies concerning collegiate athletics. Thus, one would expect that students (athletes and non-athletes) smaller colleges, most often represented among Division III and NAIA, would be most engaged.

The continuous independent and dependent measures are standardized in the models, meaning that the unstandardized coefficients in all of the tables represent effect sizes. An effect size is the proportion of a standard deviation change in the dependent variable as a result of a one-unit change in an independent variable. The larger the effect size the more likely the differences between groups represent performance that warrants serious discussion and, perhaps, intervention. Taking the advice of Rosenthal and Rosnow (1991) we consider an effect size of .10 or less to reflect a trivial difference, between .10 and .30 small, between .30 and .50 moderate, and greater than .50 large.

Results

Engagement in Effective Educational Practices

Table 3 presents the effect sizes and significance levels for the models predicting student engagement in effective educational practices. On average, student-athletes are as engaged in most educationally purposeful activities as their peers. Compared with male non-athletes, male student-athletes are as challenged academically, interact with faculty as frequently, and participate as often in active and collaborative learning activities. Female student-athletes are comparable on the academic challenge measure to their non-athlete peers. Although the effect sizes are small, female student-athletes are more likely to interact with faculty and participate in active and collaborative learning activities.

Table 3. Coefficientsa for Level-two Models Predicting Student Engagement in Effective Educational Practices
Full size table

The variance component for the student-athlete slope indicates whether the impact of being a student-athlete differs by institution. If the variance component for a slope is statistically significant, one can conclude that the impact of being a student-athlete is different because of the institution attended. Because none of the variance components for the student-athlete slope differ significantly from zero for student engagement, it appears that the nature and frequency of student-athlete engagement does not differ between institutions. This means we cannot model the student-athlete slope. Note the instances where the coefficient for the student-athlete slope is statistically significant; this is the case for women on both active and collaborative learning and student–faculty interaction. At the same time, the variance component is not statistically significant, which means that the impact of being an athlete on engagement in effective educational practices is similar across institutions.

Because the variance components for all of the model intercepts (institutional averages) are statistically significant, we are able to model the average institutional engagement with institution-level variables such as athletic division. Both men and women students at Division III schools report higher levels of academic challenge. Similarly, students at Division III schools interact with faculty more than students at Division I and Division II schools. Furthermore, men at Division I schools are statistically significantly less likely then men at Division III schools to engage in active and collaborative learning activities. Women at Division I and Division II schools are less likely than women at Division III schools to participate in active and collaborative learning activities.

Some have suggested that for various reasons Division III athletes have a qualitatively different, better-rounded educational experience than their counterparts attending schools that belong to other athletic divisions. The results from this study suggest that, in general, very few differences exist between the engagement of student-athletes and non-athletes on a given campus. However, because students at small residential liberal arts colleges (most of which are Division III schools) generally are more engaged than students at other types of institutions (Kuh, 2003; National Survey of Student Engagement, 2003; Umbach and Kuh, 2006), Division III student-athletes are more likely to be engaged than student-athletes in other divisions.

Perceptions of Campus Environment

Similar patterns of results emerge from the supportive campus environment models (Table 4). Male and female student-athletes report that their campuses provide more academic and social support than do their non-athlete peers. In addition, female student-athletes are more satisfied with the overall college experience than female non-athletes. However, male student-athletes appear to be less satisfied than other men on their campuses.

Table 4. Coefficients for Level-two Models Predicting Student Perceptions of Campus Environmenta
Full size table

After examining the variance components for the student-athlete slope, it appears that the impact of being a student-athlete on support and satisfaction does not differ between institutions. However, because the variance components are significant for the intercepts, we are able to model them using athletic division. On average, both men and women at Division III schools report they receive more support compared to students at Divisions I and II schools. Also, men and women at Division III schools are more satisfied with their overall college experience than their counterparts at Division II schools.

Self-Reported Gains

Table 5 presents the results from the HLM analysis of self-reported gains. In general, both male and female student-athletes report greater gains than non-athletes, especially in the areas of personal/social development and practical competence. Male student-athletes report greater gains in general education than their non-athlete peers.

Table 5. Coefficients for Level-two Models Predicting Student Self-reported Gainsa
Full size table

Once again, the gains of athletes do not differ significantly by institution. When we model the intercepts (average institutional reports of gains), few differences between athletic divisions emerge. Women at Division III schools report greater gains in personal/social development than women at Division I schools. Both men and women at Division III schools report greater gains in general education than students at Division I and II schools.

Grades

The final set of models predicts student self-reported grades (Table 6). Male student-athletes report earning lower grades than their peers, and the effect of being a student-athlete on grades does differ significantly by institution for men. The intercept model indicates that all students (student-athlete and non-athlete) at Division I schools have statistically significantly higher self-reported grades than all students at Division III schools. Female student-athletes report similar grades as female non-athletes.

When we model the average male grade differential, we find statistically significant grade differences between athletic division. On average, men at Division II and NAIA schools report higher grades than men at Division III schools. In fact, the effects, .11 for Division II and .17 for NAIA, erase the negative effect (−.06) of being a student-athlete on grades.

Table 6. Coefficients for Level-two Models Predicting Student Self-reported Grade Point Averagea
Full size table

Limitations

This study is limited in four ways. The first is related to the validity of the self-reported gains used in our models. As Pascarella (2001) and others point out, gain scores may be confounded by students’ entering characteristics. However, Pike (1999) provides some evidence to suggest that gain scores are not significantly related to entering ability. A fair amount of research (Baird, 1976; Berdie, 1971; Pace, 1985; Pike, 1995; Pohlmann, 1974) has shown that self-reports are likely to be valid if (1) the information requested is known to the respondents, (2) the questions are phrased clearly and unambiguously, (3) the questions refer to recent activities, (4) the respondents think the questions merit a serious and thoughtful response, and (5) answering the questions does not threaten, embarrass, or violate the privacy of the respondent or encourage the respondent to respond in socially desirable ways (Kuh, 2001). The NSSE survey was designed to satisfy all of these conditions. Although the concerns about self-reported data are legitimate, the gains measures are only one of several sets of dependent variables used in this study.

A second limitation is the way in which the NSSE survey identifies student-athletes. Students respond to the question, “Are you a student-athlete on a team sponsored by your institution?” It is possible that some students participating in sports not sponsored by their institution (e.g., club sports) responded affirmatively to the question. However, given the size of our data set the impact of the error introduced by incorrect coding of athletes is likely to be minimal.

Third, we cannot determine the sport(s) in which the athlete respondents participated. Thus, we are unable to compare the experiences of athletes competing in revenue-generating or non-revenue generating sports. Some previous research (Pascarella et al., 1995,1999) suggests that there are significant differences in the experiences of athletes in non-revenue and revenue generating sports.

Finally, given that cross-sectional data are used in this study, we are unable to control for self selection. For example, it is possible that athletes and non-athletes who matriculate at Division III colleges are more predisposed to seek out campuses where they are surrounded by people who are highly engaged. Some of the divisional differences we see may be due, in part, to a self-selection bias. While not likely, it is possible and warrants a cautionary note.

Discussion and Implications

Much has been made recently about the Bowen and Levin (2003) report that student-athletes who attend highly selective institutions do not experience campus life in the same qualitatively beneficial ways as do their non-athlete peers. This infers that athletes do not engage in effective educational practices at the same level as other students and, therefore, may are not gain as much from college. Results from this study do not support such a sweeping conclusion. One reason may be that Bowen and Levine’s sample—limited to students attending Ivy League schools—differs from the national sample used for this study.

Our results show that student-athletes are at least as engaged overall, and in some areas are more engaged, compared with their non-athlete peers. In addition, student-athletes report that they perceived their campus environment to be more supportive of their academic and social needs, and they report making greater gains since starting college in several areas. This is consistent with evidence that NCAA student-athletes graduate at slightly higher rates than non-athletes, about 62% for student-athletes compared with 60% for the general student body (NCAA, 2004). The difference in six-year graduation rates and perceptions of the campus environment may be partially explained by the amount of tutoring and other academic support many campuses provide to student-athletes. Nonetheless, the pattern of findings in this study corroborates other recent research into the collegiate experiences and outcomes associated with being a student athlete (Pascarella et al., 1995, 1999; Wolniak et al., 2001).

Of special interest are the results that illuminate the impact of being an athlete at different types of institutions. Where differences exist between athletes and non-athletes, the impact of being an athlete, on average, is the same across all institutions. In only one instance—self-reported grades of male athletes and non-athletes—did student athletes differ from other students at a statistically significant level.

Interpreting institutional effects and athlete status is more complicated when we consider average institutional engagement, campus support, and gains. The experiences of athletes appear to differ only slightly from their non-athlete counterparts, especially across athletic divisions. Because all students at Division III schools are—on average—more engaged, feel more supported, and report greater gains than their peers at other types of schools, it stands to reason that athletes at Division III institutions would be more engaged than students (both athletes and non athletes) at other types of institutions.

Given the great variation in most aspects of student life, it is almost certain that some student-athletes on a given campus are short changed in non-trivial ways in terms of what they put into and get out of college (Umbach and Kuh, 2004). This is more likely to be the case for male student-athletes at larger institutions where arguably athletics requires a greater commitment of time, both in and out-of-season. Unfortunately, the NSSE database in 2003 did not make it possible to identify the sport in which the student-athletes participated. Perhaps in subsequent years we will find systematic differences between student-athletes who play high and low profile sports, such as football and fencing, respectively.

Implications

The findings from this study tell a different story of student-athletes than the one typically featured in the national media. Most of the recent discussions have emphasized the problems that athletes create or from which they suffer. Many of the deleterious effects are associated with Division I revenue-generating sports. Granted, our results indicate that male athletes may earn slightly lower grades than their peers; at the same time, they appear to have similar or perhaps better quality educational experiences than their non-athlete counterparts in other ways. This same pattern of neutral or positive findings hold for women as well, in as much as female athletes are more engaged, report greater gains, feel more supported, and earn grades similar to non-athlete women. This is not to say that abuses do not occur or that athletes in certain sports or competing for certain institutions are not shortchanged. Certainly, unethical behavior exists within intercollegiate athletic programs, just as it does in research laboratories, fraternity and sorority houses, and classrooms including the widespread reports of cyber-plagiarism (National Survey of Student Engagement, 2003). At the same time, it does not seem fair to tar all athletes and institutions that host athletic programs with the brush of ignominy.

We hope the results of this study encourage productive dialogue about the benefits of participating in intercollegiate athletics as well as identifying areas in which institutions should aim for improvement. One of these concerns are the lower grades reported by male athletes. Even after controlling for pre-college achievement (SAT), male athletes earn lower grades. This gap between male athletes and male non-athletes is greatest at Division III and NAIA schools and warrants further study.

Given the mounting evidence of the importance of student engagement in effective educational practices, the debate about the proper role of athletics and student success should include more than grades and class rank. As important as grades are, it is noteworthy that student-athletes generally are as engaged, and is some instances more engaged, than their non-athlete peers in a variety of educationally purposeful activities. Perhaps this information will motivate those working directory with intercollegiate athletes and athletic programs to seek ways to increase this favorable margin of performance.

In addition to adding to the literature on intercollegiate athletics, our findings may offer some comfort to high school athletes. It appears that whatever college they choose, the odds are that their experiences probably will not differ greatly from other students on their campus.

Finally, it is incumbent on colleges and universities to learn more about the experiences of their student-athletes and determine whether they are taking part in educationally sound activities and benefiting in desired ways from college at levels commensurate with their non-athlete peers. After all, we know a good deal about how student-athletes perform on the playing field. We should also keep score as to the quality of their educational activities elsewhere on campus (Umbach and Kuh, 2004).

Conclusion

For such a popular topic, it is surprising that there is so little evidence at the national level about what student-athletes do during college and how their behavior compares to other students. For example, until recently we knew almost nothing about how athletes spend their time when not on the playing fields and courts. Contrary to many reports in the popular media, the findings from this study indicate that, on balance, student-athletes across a large number of colleges and universities do not differ greatly from their peers in terms of their participation in effective educational practices. In most instances, when differences do exist, they favor athletes. That is a very different picture than what is routinely presented in the popular press.

End Notes

  1. 1.

    Student-athletes are defined as anyone who responded yes to the following question: “Are you a student-athlete on a team sponsored by your athletics department?”

  2. 2.

    Student responses to the following question: What have most of your grades been up to now at this institution? A, A−, B+, B, B−/C+, C/C−/or lower.

  3. 3.

    Because of missing data, we include SAT as a control only for models self-reported grades.