Despite the tremendous growth in U.S. Native American and Latino/a undergraduate student enrollment (a 33% and 66% increase from 1994 – 2004, respectively), the graduation rates for these students are low in comparison to the enrollment rates (ACE, 2007). For example, the six-year graduation rate for Native American and Latino/a students at “removed for blind review” was 19% and 47%, respectively, for students entering fall 2002 (SU, 2009). Hernandez and Lopez (2004) refer to the gap between enrollment rates and graduation rates as the “leaking pipeline.”

Indeed, Native American students, when compared with students from other racial and ethnic groups, are the least successful in regard to college persistence (Benjamin, Chambers, & Reiterman, 1993). One of the most important factors in the success of Native American students is relationships with faculty and staff (Brown & Kurpius, 1997). Jackson, Smith, and Hill (2003) conducted a qualitative study to understand Native American college student persistence. Students’ perception of faculty and staff warmth contributed to their persistence. Jackson et al. (2003) also documented that dealing with racism and adapting to a bicultural identity were important to persistence. Overall, a strong combination of structured academic and social support contributed to the academic success of Native American students (Brown & Kurpius, 1997; Jackson, Smith, & Hill, 2003; Steward, 1993).

Educators are concerned about graduation rates for Latino/a students as well. Although Latino/a student enrollment, from 1979 to 2004, experienced a 372% growth from 353,000 to 1,667,000, Latino/a students still struggle with success and persistence in college as evidenced by the modest 94,644 bachelor’s degrees conferred in 2003-2004 (Hernandez & Lopez, 2004; KewalRamani, Gilbertson, Fox, & Provasnik, 2007). Arbona and Nora (2007) presented data indicating that the number of Latinos/as ages 25-29 holding bachelor’s degrees (10%) is lower than it is for African Americans (18%) or Whites (34%).

Factors influencing academic success for Latino/a and Native American students are similar. Social and academic support are particularly important for Latino/a students (Arbona & Nora, 2007; Torres, 2006). Hernandez (2000) reported that learning and retention were positively impacted when Latino/a students were involved in a Latino/a community on a predominantly White campus. Dennis, Phinney, and Chuateco (2005) investigated factors that affected academic success in first-generation Latino/a students; they concluded that peer support was a stronger predictor than family support of college adjustment and grades. Similarly, Latino/a students who reported that if at least some of their friends planned to attend college, they were 40% more likely to attain an undergraduate degree than those students who reported that none or few of their peers planned to attend college (Arbona & Nora, 2007).

Similar to Native American and Latino/a students, retention and graduation rates for first-generation students lag behind their second-generation peers (McCarron & Inkelas, 2006). Ishitani (2003) examined factors that affect student attrition and reported that the likelihood of first-generation students dropping out of college within the first year was 71% higher than it was for their peers with college-educated parents. In a separate study, Ishitani (2006) supported his previous findings and reported that first-generation students were 1.3 times more likely to drop out of college than their peers with college-educated parents. In addition, first-generation students with parents who had some college education were 99% more likely to leave their initial college than their counterparts whose parents had graduated from college. First-generation students were 8.5 times more likely to drop out in the second year of college than their peers who had parents who graduated from college. Similarly, of the 1,692 first-generation students surveyed by McCarron and Inkelas (2006) who aspired to finish a bachelor’s degree when in high school, only 29.5% had attained an undergraduate degree when surveyed again, 10 years later.

In a study by Pascarella, Pierson, Wolniak, and Terenzini (2004), first-generation students were followed through their second and third years of college using the National Study of Student Learning data. First-generation students tended to be less likely to become involved in structured extracurricular activities with peers, although, paradoxically, this group benefited more from these involvements than second-generation students. Pascarella and colleagues concluded that, just as with all students, first-generation students learn most when they are socially and academically involved on campus. However, they were less likely to be involved, and most likely to live and work off campus, and to study part time. Nonetheless, the good news is that for first-generation students who persisted, their learning outcomes were similar to those of second-generation students (Pascarella et al., 2004).

Another group on the rise in U.S. universities is college women — but there is a corresponding concern about college men. Astin (1998) investigated the changes in men and women college students between 1966 and 1996. He noted that the Women’s Movement influenced the involvement of women in college. In 1966, only two out of five women surveyed reported being interested in pursuing a graduate degree. By 1996, two thirds of women surveyed aspired to graduate degrees (Astin, 1998). Women showed a significant increase in bachelor’s degree attainment in the past decade. According to ACE (2007), between 1994 and 1995, the proportion of men and women who were awarded bachelor’s degrees was 45.4% and 54.6%, respectively. By 2004-2005, the percentage of women who graduated with bachelor’s degrees had risen to 57.5 and the percentage of men who received bachelor’s degrees had dropped to 42.5. In addition, among 1992 high school graduates who entered postsecondary education by December 1994, women were more likely than men to earn a bachelor’s degree or higher, and men were more likely to earn no more than a high school diploma. These relationships held for those who were in challenging courses in high school. Therefore, even among men and women who were better prepared academically in high school and had entered college, women were more likely than men to attain a bachelor’s degree (ACE, 2007). Similarly, Ishitani (2006) discovered in his study that women were 56% more likely to graduate within four years with a bachelor’s degree than were men.

Researchers have studied gender in college to understand the achievement gap between men and women (Sax, 2008). Researchers of Native American (Olenchak & Hébert, 2002), Black (Strayhorn, 2008), and White (Davis, 2002) men indicated that men experienced a high amount of challenge adjusting to college (Davis, 2002; Olenchak & Hébert, 2002), but were reluctant to seek help (Capraro, 2004; Davies et al., 2000). Men did not seek support, Capraro suggested, because asking for help is associated with femininity. For example, student affairs programming sometimes represents domesticity to men; for men who are negotiating with masculinity, any behavior associated with femininity (such as domesticity) may be threatening (Capraro, 2004; Davis, 2002; Edwards & Jones, 2009). Gender roles for men and women are shifting rapidly and many men experience identity confusion, shame, and guilt (Capraro, 2004). This gender-role shift may create unique challenges for men in the college environment.

One way to assist students in overcoming challenges in higher education is to enhance the college environment. The environment of a college or university campus has a direct effect on the learning and retention of college students (Gurin, Dey, Hurtado, & Gurin, 2002; Hurtado, 1992; Hurtado & Carter, 1997; Pike & Kuh, 2006; Tinto, 2006; Whitt, Edison, Pascarella, Nora, & Terenzini, 1999). Environment has been described as consisting of many components: the physical, social, institutional, and ecological (Conyne & Clack, 1981). A supportive campus environment is evidenced by increased student involvement and quality relationships among faculty, staff, and students (Arbona & Nora, 2007; Creighton, 2007; Kezar & Kinzie, 2006). Involvement is an important signifier of support that is linked to learning and retention (Pascarella & Terenzini, 2005). Filkins and Doyle (2002) stated that students’ perceptions of a supportive academic and social environment were the strongest predictor of students’ gains in academic development, vocational skills, and personal and social development. The more that students perceived their institutional environment as supportive, the better they exhibited learning, including effective writing skills, critical and analytical thinking, and effective self-guided learning (Filkins & Doyle, 2002; Pike & Kuh, 2005). Additional findings indicate that not only was learning affected, but without a supportive environment, attrition from college was far more likely (Hernandez & Lopez, 2004; Murguia, Padilla, & Pavel, 1991). This finding was particularly true for students of color, first-generation, and economically disadvantaged students (Hernandez & Lopez, 2004; Pike & Kuh, 2006).

The amount of positive interaction among culturally diverse students on a college campus has a positive relationship to student learning. Pike, Kuh, and Gonyea (2007) conducted a study using the National Survey of Student Engagement (NSSE) 2004, The College Student Report, the Integrated Postsecondary Education Data Systems (IPEDS), and Barron’s ratings of institutional selectivity across a nationally representative population of 428 postsecondary institutions to investigate the relationship between interaction among culturally diverse students and learning outcomes. They documented that both the quality and amount of interaction between individuals from different cultural backgrounds were significantly related to gains in understanding diverse groups. Campus diversity was also demonstrated to be significantly related to the amount of interactions among culturally diverse peers and gains in cultural understanding. Pike and Kuh (2006) explored diversity within a student body and informal peer interactions. Pike and Kuh (2006), using data from the 2001 NSSE with 305 institutions of higher education, suggested that the increased multicultural diversity of a student body exposed them to more viewpoints and posed greater challenges through interaction with those culturally diverse peers.

In addition to peer interaction with culturally diverse students, faculty interaction with students has a demonstrated effect on student learning, development, and retention (Golde & Pribbenow, 2000; Tinto, 2006). Establishing a relationship with a faculty member as soon as possible upon beginning a college career helped students to feel a sense of validation, and it reflected upon the faculty members’ interest in helping the students to find a niche in their new academic community (Terenzini et al., 1994). Hurtado and Carter (1997) reported that students who had frequent interaction with faculty members outside of class had a higher sense of belonging than those students who did not engage as often with faculty outside of class.

The context of the current study is “removed for blind review.” The first part of the “removed for blind review” mission statement is to provide an outstanding undergraduate residential education. Unique features of the university affect the achievement of that mission. In general, there are a greater percentage of women, first-generation, Native American, and Latino/a students, compared to its Carnegie peer institutions. The 6% of Native American students allows quantitative analysis, which is often a limitation of quantitative research on Native Americans (Steward, 1993). In pursuit of the undergraduate residential mission, some initiatives are evident. For example, the numbers of living-learning communities have increased from 7 to 20 from academic year 2007 to 2008. In addition, the university offers specific Native American student services; a goal in the strategic plan is to become the nation’s leading university that serves Native Americans (SU). The current study is designed to answer questions about student success in demographic groups at “removed for blind review.” With this design in mind and using freshmen and seniors attending “removed for blind review” during the 2004-2005 academic year, this study’s researchers sought to answer two research questions. First, what student background characteristics and college environments contribute to the undergraduate student learning at SU? Second, what student background characteristics and college environments contribute to undergraduate student retention at SU?

The National Survey of Student Engagement (NSSE), developed in 1998, was funded by a grant from The Pew Charitable Trusts. A successful pilot study of over 75 colleges and universities was conducted in 1999 (NSSE, 2007). The NSSE consists of 100 questions, including subquestions, some of which are the Likert type. According to Pike (2006), 49 of the questions are based on NSSE benchmarks of effective academic practice. The NSSE benchmark scales represent aspects of the college environment, including: (a) “level of academic challenge”; (b) “active and collaborative learning”; (c) “student-faculty interaction”; (d) “enriching educational experiences;” and (e) “supportive campus environment.”

Of the environmental measures, “level of academic challenge” (α = .75) is a measure of the synthesis and judgment required in courses and the length of papers and books assigned; a sample item includes “Worked harder than you thought you could to meet an instructor’s standards or expectations.” “Active and collaborative learning” (α = .67) is a measure of extensive involvement in the learning environment; a sample item includes “Asked questions in class or contributed to class discussions.” “Student-faculty interaction” (α = .75) is a measure of the interaction between students and faculty inside and outside of the classroom; a sample item includes “Discussed ideas from your readings or classes with faculty members outside of class.” “Enriching educational experiences” (α = .66) is a measure of meaningful contact with students of a different race or religion, and participation in internships, study abroad, or capstone academic experiences; a sample item includes “Talking with students with different religious beliefs, political opinions, or values.” “Supportive campus environment” (α = .77) is a measure of social and academic support, and the quality of relationships with faculty, staff, and students; a sample item includes “Campus environment provides support you need to help you succeed academically.” The outcome measure “Gains in general education” (α = .80) is a measure of gains in writing, speaking, quantitative, and general education skills. See Appendix A for a complete list of scale measures.

Items in the NSSE originated from the College Student Experiences Questionnaire (CSEQ), the Cooperative Institutional Research Program (CIRP) freshman and follow-up surveys, and surveys administered to the students and alumni of the University of North Carolina (NSSE, 2007). Validity and reliability of the NSSE have been extensively tested. Pike (2006) examined the convergent and discriminant validity of the NSSE benchmarks and of the outcome measure, “gains in general education.” Coefficients of generalizability were reported for the five benchmarks and the outcome measure. Coefficients of generalizability are defined as the ratio of the universal score variance to the expected score variance. Only the generalizability coefficients of the scales that were found to be statistically significant in this study are reported here. The generalizability coefficients of “level of academic challenge,” “enriching educational experiences,” and “supportive campus environment” were .71, .79, and .84, respectively. The outcome measure, “gains in general education,” had a generalizability coefficient of .81 (Pike, 2006).

Kuh (2001) identified the reliability of student engagement scores from one year to another. He used Spearman rho correlations to analyze the obtained 2000, 2001, and 2002 data from 80 institutions. Spearman rho correlations were calculated for the five NSSE benchmarks using the aggregated institutional data. These benchmarks and their rho values ranged from .74 to .93 for the three-year comparisons (Kuh, 2001).

The sample group for this study included freshmen (n = 292) and seniors (n = 248) of the 2004-2005 academic year. All freshmen and senior students received e-mails requesting their participation in the study and data were collected in 2005 using a Web-based survey. The overall response rate was 42%. Sample demographics are included in Table 1. Demographics for race, gender, and generation status were similar to the percentages for the campus overall. Asian American students were not included in the analyses because their numbers were too small for quantitative analysis (n = 17).

In the first design, hierarchical multiple regression (N = 464) was computed to investigate the contributing demographic and environmental factors to the NSSE outcome variable, “gains in general education,” for 2005 freshmen and senior students. The blocks included:

In the second design, logistic regression (N =235) was conducted to assess whether the independent variables significantly predicted graduation for only the senior sample within six years of enrollment (using 2005-2007 institutional graduation data). Institutional data confirming six-year (through Spring 2007) graduation were manually inserted into the NSSE database. The eleven predictor variables were: (a) gender; (b) race (4 variables); (c) generation status; and (d) the 5 NSSE benchmarks/environment factors.

In one-way ANOVA analyses, Native American students were statistically significantly more likely than White students (F[6, 656] = 5.52; p < .001; ω² = .039) to experience “gains in general education,” as were students in living-learning communities than students not in living-learning communities (F[1, 253] = 10.29; p = .002; ω² = .035). See Table 2. In chi-square analysis, first-generation students were statistically significantly more likely to be Native American (χ² = 32.30; df = 1; p < .001) or Latino/a (χ² = 15.51; df = 1; p < .001), and statistically significantly less likely to be White (χ² = 38.47; df = 1; p < .001).

The results for the first block of the hierarchical multiple regression (N = 464) indicated that Native American (β = .12; p = .006) and Latino/a (β = .12; p = .048) freshmen and seniors experienced significant “gains in general education” (F[6, 457] = 2.262; p = .037). The final block of the regression analysis, shown in Table 3, revealed statistically significant findings on the outcome of “gains in general education” (F[14, 449] = 17.63; p < .001; R² = .36). No evidence of multicollinearity was present. Predictive variables in Block 1accounted for approximately 2.9% of the variance in the criterion variable (R²= .029). Subsequently, although the second block (SAT and ACT scores) did not reveal statistical significance, the third block, consisting of the five NSSE benchmarks, showed significance. Thirty-two percent of the variability in learning was explained by the environment constructs (R ²= .320). Two out of the five benchmarks, “academic challenge” (β = .30) and “supportive campus environment” (β = .36), demonstrated significance at p < .001. The ensuing two blocks, learning community (LC) participation and grades, did not display significance.

***p< .001.Graduation

The second multivariate analysis was a logistic regression to predict seniors who graduated within six years of undergraduate enrollment. Of the seniors, 90% graduated in the fourth, fifth, or sixth year of college. When all 11 predictor variables were considered together, they significantly predicted whether a student graduated (χ² [11, N = 235] = 37.00. p < .001). The regression equation correctly predicted 99.1% of those students who graduated and 12.5% of those students who did not graduate (see Table 4). Only the variables gender (p = .004) and generation status (p = .038) were statistically significant (α = .05). Women and second-generation students were more likely to graduate. Although “enriching educational experiences” was not statistically significant at a .05 level, it achieved a marginally significant level of .053.

Note. Only significant findings are presented; EEE = Enriching Educational Experiences.

p< .05.

In logistic regression, odds ratio is interpretable as effect size — the further the odds ratio is from one, the greater the effect size (Wright, 1995). Generation status and gender had odds ratio values of 3.17 and 4.89, respectively. Although the odds ratio for “enriching educational experiences” (1.04) was not statistically significant at a .05 level, it was slightly higher than 1. An odds ratio of 1 signifies that “the predicted odds are the same for any value of the predictor” (Wright, p. 223, 1995). Researchers can infer, with 95% confidence, the odds of graduating within six years are between 1.06 and 9.43 times greater for a second-generation student than for a first-generation student. In regard to gender, with 95% confidence, the odds of graduating were between 1.66 and 14.45 times greater for a woman than for a man.

One important limitation in this study is the restriction to a single campus, so the results may not be generalizable to other campuses. Another limitation is that retention was measured for seniors only. However, the results shed light on environmental factors for student success at a public university in the Southwest, and may be instructive for campuses with similar student demographics. Further, the data are cross sectional and self reported. Finally, although the study sheds light on the significance of first generation, men, Native American, and Latino/a college student experiences, the study does not examine students’ college experiences with respect to the intersections of these identities.

Given the population in the U.S., administrators should consider implications for campus environments designed for Latino/a and Native American student success. For Latino/a students, Torres (2006) noted that it is important for them to be able to imagine the path to academic success, because many do not have cognitive road maps (due to a lack of role models) to lead them there. Structured faculty mentoring programs are one way to enhance Latino/a student success (Torres, 2006). In this study, students were more likely to graduate when they had faculty mentoring. Native American students similarly succeed at higher rates when they have supportive campus relationships. Jackson, Smith, and Hill (2003) concluded that Native American students perform better when they have opportunities to experience community, through faculty, staff, and peer mentoring.

For first-generation students, living-learning communities may offer an effective structure for college success. Inkelas, Daver, Vogt, and Leonard (2007) contended that first-generation students in learning communities have a better social and academic transition to college than do first-generation non-living-learning community students. In addition, unlike second-generation students, informal peer interaction does not predict a better transition to college for first-generation students. Instead, more formal programming such as faculty interaction and using residence hall resources leads to a better transition. It is possible that first-generation students do not possess a shared knowledge about college strategies for success, which informal interactions draw upon. Instead, a more explicit orientation to college life may better serve first-generation students than reliance upon informal friendship networks in leading them to success in college (Inkelas et al., 2007.).

Implications for improved serving of men are to work overtly with them and their related masculinity roles through structured programming (Capraro, 2004; Davis, 2002). Masculinity is sometimes associated with shame, confusion, and anger over the roles and rules of manhood (Capraro, 2004; Edwards & Jones, 2009). College men may benefit from programs that problematize popular masculinity norms, especially when the programs are delivered in a way that is nonthreatening to their masculinity. Older peer role models may represent an effective avenue to open up new possibilities for defining manhood. In order for men to stay engaged in the college experience, a challenging academic curriculum (Olenchak & Hébert, 2002), in combination with structured, supportive relationships (Capraro, 2004; Strayhorn, 2008), may be an effective strategy.

Overall, campuses may take from this study the idea that the success of Native American, Latino/a, men, and first-generation students is encouraged through structured academic and social programming that is tailored to specific needs. Two suggestions for further research are presented here. The first suggestion is to follow up these quantitative results with in-depth qualitative work, forthcoming from this study. The second suggestion is to conduct longitudinal work to follow specific students throughout their undergraduate careers and beyond.