• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

A Mixed Method Study Testing Data-Model Fit of a Retention Model for Latino/a Students at Urban Universities

Posted on: Tuesday, 30 May 2006, 06:00 CDT

By Torres, Vasti

This study presents the conceptualization and subsequent model fit analysis of a retention model for Latino/a students at urban commuter universities. The three institutions involved in the study represent different environments for Latino/a students. Two are Hispanic Serving Institutions (HSI) and the third represents a predominantly White environment. The model emerged from a concurrent nested mixed method design and reflects the influence of social cognitive theoretical foundations. Using structural equation modeling, the model that emerged from the qualitative data was found to have a good fit and explained 33% of the variance on the "intent to persist in college" variable.

Over the past several decades the desire to retain students in college has generated a large body of research within higher education (Braxton, Hirschy, & McClendon, 2004). Although the majority of administrators are likely to recognize retention models that focus on full-time, traditional age, and residential students (Reason, 2003; Tinto, 1993), they are less likely to recognize how these models may not be appropriate for students who attend institutions that serve commuter and non-traditional students. The diversity of institutions is further complicated with the increased diversity of students. Although institutions of higher education recognize the changing demographics of college students, the research has offered only a few retention models that are inclusive of diverse students. In addition, few retention models are inclusive of non-traditional commuter institutional environments (Keller, 2001; Reason; Rendon, Jalomo, & Nora, 2000). Although the recognition of diversity among college students is an important first step, there continues to be an assumption that "issues related to retention of minority students [are] similar, if not identical, to those of majority students" (Rendon et al., p. 130). The use of models and theories created for majority students is well intentioned, yet it may not be appropriate to apply those theories to certain populations. Tanaka (2002) noted that administrators should acknowledge the potential misuse of theories and research that do not pertain to minority populations.

There are two concerns that deserve attention in the higher education literature: (a) the unique characteristics of diverse populations, such as Latino/a students, and the effect culturally oriented characteristics have on retention; and (b) the need to further investigate non-traditional institutional environments, such as commuter urban universities.

Retention research is one of the areas that should be carefully reviewed to ascertain specific population effects. The inability of early retention models to incorporate non-majority students' values and cultural identities is a major criticism within the current research on retention (Attinasi, 1989; Rendon et al., 2000; Tierney 1992). Tinto's (1987) student departure model is criticized for promoting assimilation into the dominant culture (Rendon et al.) and for focusing only on the individualistic level rather than on the collective level that is important to many students of color (Tierney, 1992). In his second edition, Tinto (1993) acknowledged these critiques and focused on membership rather than integration as a way to clarify issues raised about conformity and assimilation (Hurtado & Carter, 1997). These criticisms and revisions of retention models advance the understanding of the process of retaining students, however; additional research is still needed (Gloria, Robinson Kurpius, Hamilton, & Wilson, 1999; Hurtado, & Carter; Rendon et al.).

The second concern focuses on the non-traditional commuter environment found at urban universities. Students who live in their own home or live with their parents view the college environment very differently than traditional college students at residential colleges (Bean & Metzner, 1985; Cabrera, Nora, & Castaneda, 1993). Commuter students are more likely to experience "conflicts among their obligations to family, work, and college" (Braxton et al., 2004, p. 35). The desire to raise retention rates in commuter institutions has prompted administrators to focus on changing the composition of the student body, rather than recognizing the conflicts inherent among commuter students' obligations.

The focus in this study is specifically on urban universities because this type of institution is more likely to serve first generation college students (Elliot, 1994) and provide educational opportunities to students who prefer to stay in their community or are required to live at home. This study considered the influences on the choice to stay in college for Latino/a students at urban commuter universities using a mixed method research design. The study is unique in several aspects. First, it is a customized retention study that looks at a particular population of interest (Latino/a students). Second, all of the students attend urban commuter universities. Because 94.6% of the students in this study live with their parents or in their own home, the college environment is viewed very differently than traditional college students (Bean & Metzner, 1985; Cabrera et al., 1993). These distinctions are important because few studies are conducted on only Latino/a students or within this type of institution. And finally, the context of urban universities allows for the exploration of experiences among first generation college students who are more likely to attend an institution close to their home (Elliot). Research on Latino/a students within this context is critical to understanding the choices made by these students about their college experience.

THE RESEARCH PROCESS

The use of a mixed method design requires the researcher to provide information about process as well as methodological issues. The research process in mixed method studies creates unique circumstances, and therefore, research decisions must be explained in very specific detail. For this reason, the series of decisions that guided this study will be presented first. Next, an overview of methodology is presented, followed by qualitative analysis and discussion of the theoretical framework that emerged as influencing the researcher to conceptualize the proposed model. The review of literature is presented in the order that it was seen as an influence on the research process; for this reason there are separate literature reviews presented in this article. The remaining two phases of the study use structural equation modeling to test the proposed model for overall data-model fit. The results of the quantitative analysis are then presented along with a discussion of the retention literature.

Research Decisions

This study is part of a longitudinal research project focused on the choices Latino/a students make to stay in college. As part of this larger project the decision was made to use both quantitative and qualitative data to investigate the choices of Latino/a college students. The interviews used a semi-structured set of questions and probes influenced by literature on the college environment, cultural orientation, and family influences.

The scales used in the survey portion were from a study conducted by Nora, Kraemer, and Itzen (1997) with Latino/a community college students. The scales were selected because the items were culturally sensitive and they were previously used in a study with Latino/a participants. A natural first analysis on the survey data was to test the model used by Nora et al. on this sample of urban university students. This analysis found that the model constructed for community college students was not a good fit for this sample population. This initial negative model fit analysis opened the possibility to alternative interpretations about the influences that impact students' intent to return to school.

The unique research design of this project provided some options not always available to researchers. Because this study used mixed methods, decisions were made about implementation (data collection), the priority (which method is given precedence), and integration (process of analysis and incorporation of findings) (Creswell, 2003). The implementation decision was already made, and data were already collected concurrently. At this point, the researcher had to make a decision about priority; which method would take precedence in exploring the issues that influenced students to remain in college. With only one retention model that considers Latino/a students in urban settings (Nora et al.), the decision was made to give priority to the qualitative data and explore the possibility of a new model. Though the survey data collected limited the exploration of different constructs, the researcher was confident that the scales used by Nora et al. addressed many of the culturally unique issues facing Latino/a students in the urban context. The resulting design emerged as a result of these research decisions.

Research Design

In conceptualizing a retention model for Latino/a students at urban universities, a concurrent nested strategy was used by collecting qualitative and quantitative data concurrently but having two distinct phases for data analysis(Creswell, 2003). The decision to use the qualitative data to explore the issues that impact Latino/ a students' decision to stay in college guided the order of the phases. The use of a concurrent nested design allows for a broader approach in understanding the phenomenon of interest. The mixing of the methods at the analyses phase of the study allows each method to be considered separately while also allowing them to influence each other. This design required that that methodology for data analysis be carefully considered as part of the design (Creswell).

Because the qualitative data are given priority, it is important to select strategies that are commensurate with the chosen paradigm and theoretical perspective (Greene & Caracelli, 2003; Guba & Lincoln, 1994). The unique research methodology used in the development of this model requires that the presentation of the proposed model be done in the sequence it emerged in the research process. Therefore the methodology section includes the rationale for epistemological, methodological, and method decisions.

Methodology

This study used a constructivist epistemology, which recognizes that the interaction between researcher and participants as necessary in order to understand the meaning of experiences shared during the research process (Lincoln & Guba, 1986; Charmaz, 2000). The use of both qualitative and quantitative methods in this study required that the methodology considered be inclusive of the use of multiple methods. Grounded theory methodology was selected for two important reasons. First, the goal of the research was to understand the emerging theory in order to provide a plausible explanation of the phenomenon by grounding the theory in the data (Strauss & Corbin, 1998). Through constructivist grounded theory the "'discovered' reality arises from the interactive process" between participant and researcher (Charmaz, p. 524). The second reason is that both qualitative and quantitative methods "have roles to play in theorizing" (Strauss & Corbin, p. 34) using the grounded theory research process. The key to using mixed methods is to have a circular interplay in which the qualitative data directs the quantitative data and then feeds back to inform the qualitative data. This circular process is also an evolving process in which each method is contributing to the theory in ways that are unique to the method considered. The use of grounded theory methodology requires concepts to emerge from the data; therefore, it was essential for the qualitative method to have priority. This research decision to give one method priority supports the concurrent nested strategy design used in this study (Creswell, 2003; Strauss & Corbin). In the first phase, analysis for the interview data was done using grounded theory techniques (Strauss & Corbin). First, I used open coding by conducting a line-by-line review of the data along with two inquiry auditors. Once this stage of the analysis was completed, the research team moved into axial coding, thus reconnecting the pieces of the data back together into themes that could illustrate an understanding of the environmental issues being considered (Strauss & Corbin).

The second phase of analysis focused on the quantitative data. From the themes that emerged in the interview data, a model was conceptualized that reconfigures previously used scales and provides a plausible model to consider the intent to persist in college. This re-conceptualized model was tested using latent variable structural equation modeling (SEM) to assess the data-model fit and the direct and indirect effects of various causal influences on students' intent to persist in college. This type of analysis was believed to be consistent with the research methodology (grounded theory) because it does not involve exploratory analysis of the quantitative data; rather it tests the emerging themes within a theoretically constructed model.

Sample. Qualitative and quantitative data were obtained from three urban institutions that agreed to be part of the longitudinal project. Two of the institutions are Hispanic Serving Institutions (HSI), one with over 90% Latino/a student enrollment and the other with approximately 28% Latino/a representation; the third institution represents a predominantly White environment (with Latino/a students representing approximately 4% of the overall undergraduate population). All self-identified Latino/a freshmen at the participating urban universities were surveyed in the spring of 2003 (N = 1474). The response rate was 36.1% (n = 541), and approximately 6.3% (34) of the students who responded to the survey were interviewed within their own university environment.

The sample of students who responded to the survey included more females (64%) and a mean age of 20.73 (SD = 5.8). The majority of the students (59%) were born in the U.S. of parents who are immigrants (labeled as second generation U.S. residents); only 18.4% (100) were foreign born. The remaining students were third or more generation U.S. residents. Approximately 77% of the students were first generation college students, and the majority claimed Mexico as their country of origin, followed by Puerto Rico, Cuba, El Salvador, and other countries. The majority of students (74.4%) lived with their parents, and an additional 19.7% lived in their own home. It is difficult to determine if this sample is representative of the overall sample at these institutions because institutions only collect data on ethnicity and not other demographic characteristics. This sample does reflect characteristics of the broader Latino/a population in higher education. In the overall population the largest percentages of students are of Mexican origin (Guzman, 2000; Therrien & Ramirez, 2000), and the largest increase in college attendance is for Latinas (females of Latino origin; American Council on Education, 2002).

An open sampling technique was used to invite participants for the interviews because the technique would not prematurely close off the sample and the students were volunteers from the survey respondents. The interviews took between 20 and 45 minutes, and all interviews were taped and transcribed. The content of the interviews concentrated on questions about the college environment, cultural orientation, and family influences. Additional follow-up probes were done to further prompt students to explain influences or decision- making processes about the choice to stay in school.

The interview sample consisted of 22 women and 12 men. Twenty- five of the participants were Mexican American, four were Puerto Rican, and three were from other countries (Cuba, El Salvador, and Costa Rica). Like the survey sample, the majority (19) of the interview participants were second generation U.S. residents, 8 were third or more generation U.S. residents, and 7 were born outside the U.S. Each student was asked to create a pseudonym for the purpose of this study.

Trustworthiness. Trustworthiness was achieved using a variety of strategies. These strategies provide insight into the goodness (validity) of the qualitative research process (Arminio & Hultgren, 2002; Lincoln & Guba, 1985). First, I maintained a researcher's journal for reflection and to chronicle research decisions. Second, inquiry auditors were used to help the researcher interpret the data and to assist identifying potential bias or misinterpretation. And finally, because this is part of a longitudinal study, member checks were done with the participants the following year in order to ascertain if the emerging themes were appropriate assessments of their experiences.

Reliability and Validity of Quantitative Data. The items used in the survey portion of the data collection were from previously validated scales by Nora et al. (1997) and others. These items were previously used with a commuter Latino/a student population at a community college and were seen as culturally sensitive to the issues of Latino/a commuter students. These scales were found to have strong reliability estimates: Family Responsibilities (α = .82), Encouragement (α = .78), Cultural Affinity (α = .77), Satisfaction with Faculty (α = .72), Academic Difficulty (α = .68), Academic Integration (α = .69), and Institutional Commitment (α = .89). A Family Status factor was created using both parents' educational level, and the Intent to Persist variable was measured by a single item asking about the student's intention to return to the institution.

QUALITATIVE DATA RESULTS

During the interview, students were asked what aspects of the college experience were helpful and what they would like to see changed. Theoretical sensitivity (Glaser & Strauss, 1999) was considered when creating the question because the focus was on understanding the influence of the college environment on their choice to stay in college. Through the analysis it emerged that these students have a different perception of the college environment than what many research studies on traditional aged students describe. Three themes emerged as prevalent influences on students' college experience. The first was designated as "come to class and leave" to describe the hectic nature of students lives at urban commuter universities. The second was designated as "unsure I could make it" to illustrate the previous expectations towards academics and how this attitude influenced students' college experiences. And the third theme was designated "show me the way" to illustrate the ways students were able to create cognitive maps to help them maneuver the college environment.

Come to Class and Leave

Students at commuter institutions often have multiple priorities and see being in college as one of many commitments they maintain. When Oscar was asked what he thought was helpful and what things he wished were different about the college environment at his university, h\e responded by saying: "I usually just come for classes, and just leave. So I don't think I can really answer that." Although he noticed the various opportunities available to him, he also noted that he did not have the time to participate. Oscar worked off-campus and felt he had to maintain his work schedule in order to meet his financial obligations. Jennifer also struggled to think about what was helpful during her college experience and stated that: "I really don't understand when they say college experience because I just go to school. I go to class, and then I am out of here."

Other responses to the question about college experiences centered on simple descriptions like, "it is nice" or "calm" on campus. Although these students wanted to better themselves through education, traditional views of college life were seldom applicable and in some cases not desired. Because these students did not define the college environment in traditional ways, more exploration was done about specific aspects of their experiences in order to understand potential influences. Probes during the interviews allowed for more specific issues to emerge that influenced the students and thus provided insight into the students' perception of college life. These explorations highlighted prior academic experiences that illustrated the messages that they would not succeed academically and thus prepared them to have low expectations of themselves and the academic environment.

Unsure I Could Make It

Although not applicable to all students, several of the students who had not experienced academic success in high school were doing much better in college. This theme focuses on the institutional stereotyping that can happen in society and the resulting low level of expectations set for students (Steele & Aronson, 1995). Susie is a student who had a difficult time in high school, but changed her attitude and behavior in college. As she put it:

. . . high school was more for me a social scene than going to learn. . . . In high school you . . . are willing to cut, we usually didn't get attention. The only attention we got was from security . . . they take our IDs away. . . . All they do is single us out from the rest of the class.

The low expectations and lack of encouragement experienced by Susie did not promote behavior that would lead to academic success. When asked what made a difference for her, Susie responded by saying that she was tired of all that stuff and that "being in college, I finally realized education is important." She recognized she was not academically prepared for college level courses as a result of the low expectations in her high school.

Susie found support through an academic assistance program designed specifically for Latino/a students at her urban university. Susie talked about her advisor in the program helping her find tutoring and connecting her to people "that are going to be beneficial to my career." The mentoring relationship she was able to establish helped her to see herself and her future in a different light.

Another student was Colleen, who in high school felt like most teachers expected her to cut class even when she had a doctor's note to legitimize her absence. In her high school Colleen assumed she was not being seen as a legitimate student. Because she had a 2.1 high school GPA and is economically disadvantaged, she did not seriously think about college. The change began when one of her teachers reached out to her and let her know that her financial difficulties should not stand in her way of going to college. She describes the interaction by saying:

In my junior year, my earth science teacher, he was the one that kind of pushed me to go to college. He saw that I was doing good in earth science and he wanted me to become an earth science major. And he went here and started talking to me about here [university] and at the same time he was telling me about some type of tuition merit waiver.

At the time of our interview, Colleen was preparing to apply to the honors program and had maintained a 3.5 GPA in her first year of college. She had continued to stay in touch with her mentor and relied on him for information and advice about maneuvering the system within the university.

When asked what made a difference for students who had not performed well in high school, the students repeatedly referenced a mentor or faculty member who believed in them and encouraged them to attend college. This interaction prompted them to change their expectations about education, and in turn their behavior became more consistent with what is needed to be a good student. Even students with good grades referenced adult figures who helped them figure out the system. Although they did not experience the same academic difficulties as Susie or Colleen, they talked about not knowing how to do things in college and their feelings of not knowing if they would do well in the new environment. These comments illustrated the insecure feelings many first generation college students expetience, yet the feelings are further exacerbated when there is less cultural congruency in the environment. These comments led to the third emergent theme in the qualitative data.

Show Me the Way

The students at the three urban institutions repeatedly referenced mentors, special academic assistance programs, and faculty as making a difference in their college experience. The helpful aspects of these programs included having students like themselves around, oneon-one attention, and the ability to get specialized academic help. Danneal, though not academically required to enter a special program (GPA entrance criteria), chose to go through the same program as Susie because she "felt like maybe being in a little bit more around my people [Latino/as], and more like one- on-one counseling, it would help me a great deal." She went on to describe that the personal attention made a difference for her and allowed her to feel more comfortable with being a college student. Rachel entered a different type of an academic support program by being part of a cohort program in one of the schools within the university. She found that the friendships she made within her cohort helped her through her academic difficulties and made her feel part of a community.

Several students also talked about individual faculty being very helpful and willing to assist them outside the classroom. The availability of faculty on these urban campuses emerged as a major positive influence for these students. Isis described the interactions with the faculty by saying: "The teachers are more . . . into the students." At another university, Roberto described it by saying that college "is more challenging than high school. . . . So far, all my teachers have been a lot of help." For Roberto, a peer mentor program was very helpful because it allowed him to talk to older students who could provide advising about academics and other concerns. On a different campus, Panfilo mentioned that he had to search for tutorial help, but once he found the supplemental instruction sections he felt like they provided the assistance he needed. This supplemental instruction is based on the learning community model and focuses on gatekeeper courses where students have traditionally had difficulty being successful.

Other personal connections with faculty and mentors were made through either their academic advisor or on-campus job. Isis had a work-study job in an academic department, and because she often worked alone, she was able to reflect about issues connected to her studies. In addition, she found the faculty in the department to be helpful: "They give me advice. . . . It helps, because you keep on thinking about what it is that is going on." Even non-academic jobs on campus seem to help. Manuela worked a few hours a week at the university bookstore and found that knowing what courses are offered and the books being used "helps me understand a lot."

Many of the students had strong ethnic support systems away from the college setting. These systems were maintained because they continued to live at home, rather than leaving to go away to college. Students also described their social support as being outside the college environment. Most saw their family as their central social support system and saw their social relationships as part of their communities rather than the college environment. Almost all of the students felt their families were supportive but did not necessarily understand their experiences in college. This conflict between support and understanding created cultural issues for students. For example, Maria explained how cultural expectations conflict with expectations of college life. For Maria, this type of conflict emerged when she had to attend an out-of-class group meeting:

They are so strict. I have to sometimes go to meetings and projects and they are like, "Well no," and I am like, "Well this is different." It is not high school, it is not grammar school. It is college, I need to go out.

In addition, the majority of the students continued friendships they already had in high school because they stayed together in college. The continued friendships were not attributed to the college environment; rather the students saw them as part of their neighborhood or community.

Based on the themes that emerged, the traditional predictors of retention such as previous academic performance or measures of academic achievement (Nora et al., 1997; Tinto, 1993) did not seem to be the most important factors influencing student success in this context. Instead, campus experiences emerged as being more important in helping these students stay in college.

Discussion of Qualitative Results

During the analysis phase the emerging themes focused on how students saw themselves within the college environment and the level of support they were able to gain from positive contacts they madewith adult figures who could help them figure out the system. The majority of the students in the qualitative portion of this study seldom mentioned their deficiencies and tended to see their previous academic performance in light of the stereotype imposed on them as low achievers and as a result low expectations placed on them by high school teachers. These interpretations are consistent with other research on minority students (Immenvahr, 2003; Tierney, 2000). Other research has found that the campus environment was a stronger influence on institutional commitment than student entry characteristics (Strauss & Volkwein, 2004). In their propositions about student departure at commuter institutions, Braxton et al. (2004) assert that support from significant people would decrease the likelihood of departure from the institution. Although other propositions made by Braxton et al. may apply to these students, it is important to note that "racial or ethnic minority students often feel pressured to spend more time with family or to oversee family matters, which decreases the amount of time available to engage the academic and social aspects of the institution" (p. 49). This caveat indicates that some of the propositions posited by Braxton et al. make Latino/as deficient in their ability to be successful in college. This deficiency lens is not in congruence with the research question and does not explain why some students succeed with these limitations. What is critical for minority students is "how they negotiate these conflicts and how much support students receive from significant others for college attendance" (Braxton et al., p. 50). These caveats in the literature warrant the use of a different lens.

A Different Theoretical Lens

As part of the on-going research and reflection that occurs in qualitative analysis, I began to look at the concept of self- efficacy. Although many researchers consider self-efficacy, few recognize that this is only one portion of a larger theoretical lens proposed by Bandura (1986). The theoretical lens that includes self- efficacy is social cognitive theory, which advocates that people function (make choices) as a result of the interaction between behavior, cognitive and personal factors, and environmental influences (Bandura). Social cognitive theory explains human functioning (choices) in terms of the "triadic reciprocality in which behavior, cognitive and other personal factors, and environmental events all operate as interacting determinants of each other" (Bandura, p. 18).

Social cognitive theory views the nature of people as defined by their capacity to: (a) symbolize, (b) have forethought, (c) learn vicariously, (d) self-regulate behavior, and (e) self-reflect on their experiences (Bandura, 1986). The development of these capacities can permit the person to achieve their goals. Although many of these students did not have these capacities when they began college, it was the ability to develop these capacities that made a difference for them to stay in college. The majority of the students, approximately 77%, were first generation college students. This factor alone could have limited their ability to create positive symbols about how to maneuver the college maze and how to stay motivated in order to persist through the college years. In addition, the low expectations in high school may not promote the level of cognitive development needed to create symbols and have forethought that would transcend their present experiences and allow them to envision a different reality or life goal. The inability to create positive symbols about alternative life paths impacts the ability to have forethought about what to expect in the college environment (Bandura). Forethought can also influence the college choice process because more selective colleges require students to apply early on and will not consider students after certain deadlines. For some of these students the possibility of college attendance was not considered until after their high school graduation, and their choice was also dependant on their family responsibilities.

Within education there is much emphasis placed on experiential learning, yet few of these students were provided the opportunity to learn about college experientially. Instead, they had to depend on vicarious learning, which occurs through observing others' behavior and assessing the consequences. Many of the students spoke about their parents' desire for them to have a better life. By observing the difficulties their parents experience without an education, they recognized the need for education, but there were few role models for them to observe that illustrate an alternative life path that includes education. Often they entered the college environment with few tangible expectations and did not know what to ask for if they needed assistance. This lack of mentors, role models, or advisors may influence their self-regulatory and self-reflective capabilities. These capacities impact a student's self-efficacy towards being able to make it in college. Oftentimes these students have erroneous beliefs and information. In order to maintain some congruence they conform their behavior to those erroneous thoughts (Bandura, 1986). The ability to help students create positive images and capacities is critical to their ability to reconstruct their self-imagine and change their behavior to assure success (Torres & Baxter Magolda, 2004). The students who were able to develop these capacities in the college environment were better able to describe the behaviors needed to succeed.

Although personal determinants, such as entry skills, emotional ties, and personal standards, are difficult to gauge and control, institutions can impact social or environmental determinants that can positively impact students. Contact with mentors, advisors, and faculty should be seen as a mechanism for fortuitous encounters that can positively impact students' social cognitive capabilities (Bandura, 1986).

While reflecting on the findings of the qualitative analysis and how those themes impact a new retention model, I began to see the relationship between what the students were saying and the concepts of social cognitive theory. The themes that emerged from the interviews had several commonalities with the concepts in Bandura's (1986) theory. The decision was made to consider using social cognitive theory as a framework to consider how to configure a model that considers the influences on the Intent to Persist for these students. Using the qualitative data as a guide, the elements of the model were reconfigured. The following section will connect the emergent themes to aspects of social cognitive theory. This study does not test social cognitive theory; it only uses this theory to enlighten the issues that can possibly influence the choice to stay in college. By looking at the environmental factors through a social cognitive perspective, the data can be put together conceptually to represent a series of choices that illustrate a model that could help explain intent to stay in school.

Environmental Influences. The themes of "come to class and leave" and "unsure I could make it" illustrate the environmental influences, or capacities, that students have to develop in order to stay in college. Because the interviews occurred in the second semester of the freshmen year, it is assumed that these students have worked on these capacities in order to stay in college. The theme, "come to class and leave," illustrates the influence of family responsibility and family status. The theme, "not expected to succeed," illustrates some of the aspects of academic difficulty and satisfaction with faculty and interplays with the comfort level students feel within their environment (cultural affinity).

These environmental influences were considered the exogenous variables that influenced the latent variables: (a) family status- representing parents' level of education, (b) family responsibilities-which considers obligations that may interfere with school work, (c) cultural affinity-which represents the presence of other Latino/as in the college environment, (d) satisfaction with faculty-which looks at students' impression of the faculty as well as potential adult mentors, and (e) academic difficulties-which considers students' self-perception of their cognitive abilities.

Behaviors, Cognitive, and Personal Influences. The theme, "show me the way," most clearly illustrates that the experiences students had while in the college environment centered on relational issues. There were influential mentors, advisors, faculty, friends, and/or family that encouraged them to be in college. These older peers or adult figures helped them create symbols that allowed them to envision success and thus create cognitive maps to maneuver the college system. These connections assisted their ability to figure out the system and create positive symbols for themselves. For this reason the items previously used by Nora et al. (1997) for the Encouragement scale were used as a proxy for the latent variable cultural and social symbolism. These items focus on family, friends, and teachers/advisors support for students in college. Through this encouragement from family, peers, and mentors, students are better able to create the symbols necessary to envision their success. The latent variable academic behavior was operationally defined by the items Nora et al. named the Academic Integration scale. These items focus on the use of tutors and meeting with faculty and academic advisors as well as other academic behaviors that would lead to success in college. Students' ability to reflect on their environment and appreciate the importance of their educational experiences was seen as an internal commitment to the college experience. This capacity was operationally defined as a latent variable using the institutional c\ommitment items from Nora et al. that focus on the level of importance a student placed on a college education and the sense of belonging he/she senses.

The functioning required to persist in college was operationally defined using the Intent to Persist variable. Together these emerging issues serve as the basis for the reconfiguration of a different retention model for Latino/a students at urban universities that focuses on students' ability to make meaning of their experiences while in college rather than focusing on pre- college attributes.

The reconfigured model conceptualizes the influences on the intent to stay in college as focusing on academic behaviors (defined by Academic Integration scale items in Nora et al., 1997), cultural/ social symbolism (defined by Encouragement scale items in Nora et al.) and the influence these factors have on students' reflection of the environment (defined by Institutional Commitment scale items in Nora et al.). These variables are considered together to look at the Intent to Persist for Latino/a students at urban universities (see Figure 1).

FIGURE 1. Social Cognitive Retention Model for Latino Students at Urban Universities (Standardized Solution)

Rather than focusing on academic deficiencies, this model focuses on the adaptations students are able make in the college environment and the kinds of variables that can influence those adaptations and lead to academic success. This type of retention model seeks to identify ways Latino/a students can modify their relationships rather than presume a disassociation from their culture (Rendon et al., 2000). The model reconstructs how institutions can view their services to better serve Latino/a students.

QUANTITATIVE DATA

Review of the Literature on Retention Models for College Students

The two most widely used retention models focus either on the interaction effect between integration of academic and social characteristics with the student's motivation (Tinto, 1993) or on organizational, personal, and environmental variables that shape the attitude and intent of the student (Bean, 1983). The commonalities between these models include: a view of retention as a complex set of interactions, the use of pre-college characteristics to view influence, and the advocacy of a successful match between the student and the institution (Hossler, 1984).

Initially validation studies for these models were conducted with traditional residential college students. Since this time, they have been tested on commuter students (Bean & Metzner, 1985), and additional research has illustrated that academic and social integration can be applied to ethnically diverse campuses (Murguia, Padilla, & Pavel, 1991; Nora, 1987). Yet this type of research has also brought up additional questions prompting others to argue that integration is the expectation that students will adapt to the culture of the campus (Tanaka, 2002). In addition, these models are seen as expecting minority students to enter and adapt to the dominant cultural frame (Tierney, 1992). Although there is substantial research on retention, there is still much more information needed to understand the college experiences of nonmajority commuter students.

In a recent monograph, Braxton et al. (2004) posited 16 propositions about the retention of commuter students in higher education. These propositions emerged from a review of the literature on the retention of commuter students. Although the propositions provide an interesting synopsis of issues, they do not provide insight into what helps students stay in college; they mainly focus on the characteristics that can influence the departure of a commuter student.

This study sought to investigate the retention of Latino/a students by listening to their voices through the data gathered during interviews and then conceptualizing the results into a retention model that could reflect a culturally sensitive reality. For example, students in the qualitative phase seldom mentioned prior academic performance as critical to their success. This is because many of the students at the urban universities had not received appropriate academic preparation in high school and seemed to understand they needed to work harder. Another difference is that institutional fit did not seem as critical because the choice to be at an urban university was based on family expectations to live at home and on the relatively low cost of attending an urban university (as opposed to other universities in the area or within the state) rather than students looking at social and academic fit variables. As an example, one student thought about a particular major, but when he found out it was not offered at his urban university he settled on another major. Students were not as concerned with whether they fit with the environment as much as whether they understood how to navigate the system. Although these issues can be intertwined, researchers seem to interpret fit as choosing from many options. Yet for these students, cultural or family reasons dictated a limited range of options.

TABLE 1.

Confirmatory Factor Analysis of Items

Quantitative Results

The second phase of analysis in this study involved using SEM on the survey data to test the initially conceptualized model (Figure 1) for overall data-model fit. As recommended by Byrne (1998), a two- step analysis process was conducted using LISREL (Jreskog & Srbom, 2001). First a confirmatory factor analysis (CFA) was performed to assure that each latent variable was measured appropriately (see Table 1). Results from the CFA indicated that several items were not contributing to the hypothesized latent variables and, after theoretical consideration, were removed from the model. The items deleted were part of the Academic Integration scale and dealt with use of the library and use of a computer outside of class. Because these were mainly commuter students with many other obligations, the use of the library was seen as linked to spending more time on campus. The use of computers outside of class was seen as possibly linked to the digital divide and the fact that for some of these students a computer may be a luxury item. Several students spoke of using computer labs on campus, which they may have associated as within the classroom environment.

The second step consisted of assessing the overall data-model fit of the model with the modified measurement portion from the first step in place (see Figure 1). Overall fit results suggest an acceptable fit between the collected data and the hypothesized model: The standardized root mean square residual (SRMR) of .062, root mean square error of approximation (RMSEA) of .046 (confidence interval .041; .050), and CFI of .97 indicate good overall data- model fit; whereas the adjusted goodness of fit index (AGFI) of .89 suggests some degree of data-model misfit. Hu and Rentier (1999) suggested that target values for acceptable fit are SRMR ≤ .08, RMSEA ≤ .05, CFI and AGFI ≥ .95. However Byrne (1998) suggested that the RMSEA is perhaps the most informative criteria to consider. For this hypothesized model the RMSEA is .046 indicating a good fit. This would indicate that the hypothesized model using a social cognitive theory lens is a plausible explanation of the influences on the Intent to Persist in college for Latino/a urban university students. Again, this model does not test social cognitive theory; it uses theory only to inform how the variables are placed in the model.

In this initial model, Family Status, Family Responsibility, Satisfaction With Faculty, Cultural Affinity, and Academic Difficulty accounted for 20% (R^sup 2^ = .20) of the variance in Academic Integration (Academic Behaviors) and 42% (R^sup 2^ = .42) of the variance in Encouragement (Cultural and Social Symbolism). Furthermore, Academic Integration (Academic Behaviors) and Encouragement (Cultural and Social Symbolism) accounted for 25% (R^sup 2^ = .25) of the variability in Institutional Commitment (Reflection). Finally, the latter three latent variables explained 33% (R^sup 2^ = .33) of the variability in Intent to Persist.

In considering the direct and indirect effects (see Table 2), I first focused on the effects of the latent variables Academic Integration, Institutional Commitment, and Encouragement on the variable Intent to Persist: Only Institutional Commitment had a strong and significant (.54) direct effect on Intent to Persist; the effect of Academic Integration (.02) and Encouragement (.05) on Intent to Persist were not significant influences. The indirect effects from these two variables on Intent to Persist were stronger (due to a solid direct effect of Institutional Commitment on Intent to Persist): The indirect effect of Academic Integration through Institutional Commitment was .11 whereas the indirect effect of Encouragement through Institutional Commitment was .21. Interpreted within the context of social cognitive theory, the behaviors and symbols are not as important to intent until the appropriate internal reflection has occurred to consider college as an important aspect of a student's life (Institutional Commitment).

TABLE 2.

Direct, Indirect, and Total Effects on Endogenous Variables (Standardized Path Coefficients)

Finally, in considering the effects of the exogenous variables on the endogenous variables, there are two variables that yielded interesting results. The largest influence (direct effect) on encouragement is Cultural Affinity (.45) followed by Satisfaction With Faculty (.24). The same two exogenous variables had the largest influence on Academic Integration with Cultural Affinity having a .25 direct effect and Satisfaction With Faculty having a .25 direct effect. The largest total effect (the sum of direct and indirect effect components) was .715 and occurred between Cultural Affinity and Intent to Persist, through both Encouragement and Institutio\nal Commitment.

DISCUSSION

Models that are dependent on looking at the institutional fit approach have been critiqued for use with students of color as to whether this approach explains the real situations faced by students who are economically or culturally limited to certain college choices (Rendon et al., 2000; Tierney, 1992). In addition, the deficiency approach that focuses on the poor academic preparation some students receive prior to attending college has also been critiqued for not recognizing the potential in students (Tierney, 2000). With these critiques in mind the discussion of the reconfigured model using social cognitive theory as the theoretical foundation will look at three areas: the need to help students create cognitive maps that show them how to maneuver the college environment, the role of mentoring to help students create positive symbols, and the need to consider the inclusion of social and cultural values on campus because of their importance for students when they are away from campus.

Because the majority of students in commuter urban universities are likely to be first generation college students, there is a critical need to assist students in creating cognitive maps. In order for them to learn the capacity to have forethought about how to succeed, these students must be taught how to create a cognitive map that includes positive symbols, self-reflection, self- regulations, and forethought that allows them to understand and maneuver the college system. This is critical for Latino/a students, who often have additional pressures to not only understand the college experience, but also attempt to explain it to their parents (Torres, 2004). The ability to create a cognitive map that allows them to envision how they can succeed in college is an important aspect of transitioning to college and having a sense of belonging (Hurtado & Carter, 1997).

In a recent study of colleges with higher than predicted scores on the National Student Survey of Student Engagement, researchers found that one of the lessons learned is that colleges need to "recognize that new students need affirmation, encouragement, and support as well as information about what to do to be successful" (Whitt, 2005, p. 3). This is an acknowledgement that institutions need to see themselves as co-creators of students' cognitive maps. To be effective in helping students succeed, colleges cannot assume students will know how to succeed in college. The findings of this study highlight this need among Latino/a commuter college students.

The role of institutional commitment was interpreted in this study as the ability to reflect on the environment and to accept the college experience as an important part of the students' life. With this interpretation in mind, it is understandable why institutional commitment is strongly linked to persistence (Cabrera, Nora, Terenzini, Pascarella, & Hagedorn, 1999). As students create their cognitive maps on how to succeed they are better able to adjust their behavior in ways that help them succeed. With this feeling of success comes the reflective aspect necessary to achieve a stronger level of institutional commitment and the self-regulating behaviors that promote academic success. Once students see college as an important aspect of their lives, they are better able to create symbols that help them conceptualize how to better manage their behavior, the environment, and the personal and cognitive factors that impact their ability to succeed in college. This learning process is the foundation of social cognitive theory (Bandura, 1986).

An important influence to enhance cognitive factors are mentors and advisors who can help students create those positive symbols "that transcend their sensory experiences" (Bandura, 1986, p. 18). It is through symbols that students can generate innovative ways to complete their task and learn vicariously from their observations. For many urban students, their sensory experiences (real life) may not support the sacrifices and energy needed to be successful college students. The students repeatedly mentioned that although parents were supportive, they did not understand what college life is about (Torres, 2004). This dissonance between support and understanding makes it critical to have adult mentors and advisors who understand the students' cultural needs as well as the college environment these students navigate. Lessening the dissonance allows the students to gain the capacity to create their own symbols. The use of mentors in the college environment is advocated through other research (Hamrick & Stage, 2004) and is one of the foundational aspects of academic support programs (Torres, 2003). The analysis of the total effects yielded that largest total effect (the sum of direct and indirect effect components) occurred between Cultural Affinity and Intent to Persist, through both encouragement and institutional commitment. This total effect highlights the influence of having Latino/a culture represented within the environment and among faculty, staff, and students as an important and significant role in students' intent to persist.

Because first generation college students may lack the social, cultural, or economic capital needed to be successful in a college environment, they may leave because they are unable to conceptualize how this capital may be enhanced (Berger, 2000). Enhancing students' ability to create positive symbols for success helps students overcome inadequate or erroneous information in order to better interpret their misconceptions and in turn their environment (Bandura, 1986). By providing the accurate information and enhancing their cognitive development, students can better understand how their behavior is received and interacts with their environment.

Finally, it is critical to understand that students do not leave their cultural values at the door when they come to college. It is important to create mechanisms that recognize and assist students in making the modifications to their relationships that will benefit them in the future (Rendon et al., 2000). These modifications are made gradually and are closely tied to the student's ability to conceptualize a different future that includes higher education. The challenge in creating inclusive environments is to "develop ways in which an individual's identity is affirmed, honored, and incorporated into the organization's culture" (Tierney, 2000, p. 219). This model illustrates the strong direct effect Cultural Affinity has on the other variables that impact Intent to Persist. In order to address the challenge of creating inclusive environments, Tierney (2000) recommends that institutions must look at models that: address collaborative relationships dealing with power; establish connections between home, community, and schools; address issues of remediation; and provide the academic support needed for students to succeed. These recommendations would also address the needs expressed in this study. With this type of environment, students are better able to gain the capacities needed to have more self-efficacy and thus, succeed within the college environment. Without these enhancements students at urban universities who are first generation college attendees and Latino/ a are less likely to succeed.

LIMITATIONS

The limitations in the study revolve around the generalization of the qualitative data and the measurement issues inherent in survey research. This study focused on data-model fit; therefore, the fit indices were the focal point. Consistent with the methodology of the study, the goal was to test the model. The promising, yet conflicting data-model fit results of the structural equation modeling analysis suggest that the hypothesized model in Figure 1, which is conceptualized from the qualitative results, might be a tenable explanation for the observed associations among the survey items. Additionally, more in-depth analyses of the quantitative data might lead to a slightly modified model that will be of interest to researchers for future testing and validation. Although these limitations are important to note, the potential issues that emerged are also important to consider in future research. The testing of this model using actual retention is limited by the fact that the retention rate for this sample of urban students was low (55%). Urban universities traditionally have lower retention rates than traditional residential institutions because students stop out for one semester to attend the community colleges or to deal with family issues, making institutional retention as the outcome variable difficult. A more accurate outcome variable would be persistence, which is inclusive of students' decision to transfer to another institution or intent to return.

This study provides a different research design to consider when conducting retention research with underrepresented populations while also testing the emerging themes using established survey scales. Although this model is conceptualized using parts from other models and previously validated scales, the interpretation of the interactions should be carefully considered. By focusing on the capacities that must be enhanced once students are in the college environment, academic administrators can better identify the content of programs that can make a difference for Latino/a students.

REFERENCES

American Council on Education (2002). Nineteenth annual report on minorities in higher education. Washington, DC: Author.

Arminio, J. L., & Hultgren, F. H. (2002). Breaking out from the shadow: The question of criteria in qualitative research. Journal of College Student Development, 43(4), 446-460.

Attanasi, L. C., Jr. (1989). Getting in: Mexican American's perceptions of university attendance and the implications for freshmen year persistence. Journal of Higher Education, 60(3), 247- 277.

Bandura, A. (1986). Social founda\tions of thought and action a social cognitive theory. Englewood Cliffs, NJ: Prentice Hall.

Bean, J. P. (1983). The application of a model of turnover in work organizations to the student attrition process. The Review of Higher Education, 6, 129-148.

Bean, J. P. & Metzner, B. S. (1985). A conceptual model of nontraditional undergraduate student attrition. Review of Educational Research, 55, 485-540.

Berger, J. B. (2000). Optimizing capital, social reproduction, and undergraduate persistence. In J. M. Braxton (Ed.), Reworking the student departure puzzle (pp. 95-124). Nashville, TN: Vanderbilt University Press.

Braxton, J. M., Hirschy, A. S., & McClendon, S. A. (2004). Understanding and reducing college student departure. San Francisco: Jossey-Bass. (ASHE-ERIC Higher Education Report, 30(3))

Byrne, B. M. (1998). Structural equation modeling with LISREL, PRELIS, and SIMPLIS: Basic concepts, applications, and programmingab:. Mahwah, NJ: Lawrence Erlbaum.

Cabrera, A. F., Nora, A., & Castaneda, M. B. (1993). College persistence: Structural equations modeling test of an integrated model of student retention. Journal of Higher Education, 64(2), 123- 139.

Cabrera, A. F., Nora, A., Terenzini, P. T., Pascarella, E., & Hagedorn, L.S. (1999). Campus racial climate and the adjustment of students to college. Journal of Higher Education, 70(2), 134-160.

Charmaz, K. (2000). Grounded theory objectivist and constructivist methods. In N. K. Denzin & Y. S. Lincoln (Eds.), Handbook of qualitative research (2nd ed., pp. 509-536). Thousand Oaks, CA: Sage.

Creswell, J. W. (2003). Research design qualitative, quantitative, and mixed methods approaches (2nd ed.). Thousand Oaks, CA: Sage.

Elliot, P. G. (1994). The urban campus: Educating the new majority for the new century. Washington, DC: American Council on Education Series on Higher Education.

Glaser, B. G., & Strauss, A .L. (1999). The discovery of grounded theory: Strategies for qualitative research. Hawthorne, NY: Aldine de Gruyter. (Original work published 1967)

Gloria, A. M., Robinson Kurpius, S. E., Hamilton, K. D., & Wilson, M. S. (1999). African American students' persistence at a predominantly white university: Influences of social support, university comfort, and self-beliefs. Journal of College Student Development, 40, 257-268.

Greene, J. C., & Caracelli, V.J. (2003). Making paradigmatic sense of mixed methods practice. In A. Tashakkori & C. Teddlie (Eds.), Handbook of mixed methods in social & behavioral research (pp. 91-110). Thousand Oaks, CA: Sage.

Guba, E. G., & Lincoln, Y. S. (1994). Competing paradigms in qualitative research. In N. K. Denzin & Y.S. Lincoln (Eds.), Handbook of qualitative research (pp. 105-117). Thousand Oaks, CA: Sage.

Guzman, B. (2000). The Hispanic population: Census 2000 brief. Current Population Reports, C2KBR/01-3. Washington, DC: U.S. Census Bureau.

Hamrick, F. A., & Stage, F. K. (2004). College predisposition at high-risk minority enrollment, low-income schools. The Review of Higher Education, 27(2), 151-168.

Hossler, D. (1984). Enrollment management. New York: College Board.

Hu, L., & Bentler, P. M. (1999). Cutoff criteria for fit indexes in covariance structure analysis: Convention criteria versus new alternatives. Structural Equation Modeling: A Multidisciplinary Journal, 6, 1-55.

Hurtado, S., & Carter, D.F. (1997). Effects of college transition and perceptions of campus racial climate on Latino college students' sense of belonging. Sociology of Education, 70(4), 324-345.

Immerwahr, J. (2003). With diploma in hand Hispanic high school seniors talk about their future. San Jose, CA: National Center for Public Policy and Higher Education and Public Agenda.

Jreskog, K. G., & Srbom, D. (2001). LISREL 8.50. Lincolnwood, IL: Scientific Software International.

Keller, G. (2001). The new demographics of higher education. The Review of Higher Education, 24(3), 219-235.

Lincoln, Y. S., & Guba, E. G. (1985). Naturalistic inquiry. Thousand Oaks, CA: Sage.

Murguia, E., Padilla, R.V., & Pavel, M. (1991). Ethnicity and the concept of social integration in Tinto's model of institutional departure. Journal of College Student Development, 32, 433-439.

Nora, A. (1987). Determinants of retention among Chicano college students: A structural model. Research in Higher Education, 26, 31- 59.

Nora, A., Kraemer, B., & Itzen, R. (1997, November). Persistence among non-traditional Hispanic college students: A causal model. Presented at the Association for the Study of Higher Education, Albuquerque, NM. (ED 415 824)

Reason, R. D. (2003) Student variables that predict retention: Recent research and new developments. NASPA Journal, 40(4), 172- 191.

Rendon, L.I., Jalomo, R.E., & Nora, A. (2000). Theoretical considerations in the study of minority student retention in higher education. In J. M. Braxton (Ed.), Reworking the student departure puzzle (pp. 127-156). Nashville, TN: Vanderbilt University Press.

Steele, C. M., & Aronson, J. (1995). Stereotype threat and the intellectual test performance of African Americans. Journal of Personality and Social Psychology, 69(5), 797-811.

Strauss, A., & Corbin, J. (1998). Basics of qualitative research techniques and procedures for developing grounded theory (2nd ed.) Thousand Oaks, CA: Sage.

Strauss, L. C., & Volkwein, J. F. (2004). Predictors of student commitment at two-year and four-year institutions. The Journal of Higher Education, 75(2), 203-227.

Tanaka, G. (2002). Higher education's self-reflexive turn toward an intercultural theory of student development. The Journal of Higher Education, 73, 263-296.

Therrien, M., & Ramirez, R. R. (2000). The Hispanic population in the United States: March 2000. Current Population Reports, P20-535. Washington, DC: U.S. Census Bureau.

Tierney, W. G. (1992). An anthropological analysis of student participation in college. The Journal of Higher Education, 63, 603- 618.

Tierney, W. G. (2000). Power, identity, and the dilemma of col

Source: Journal of College Student Development

More News in this Category