PLACE IN RETURN BOX to remove this checkout from your record. TO AVOID FINES return on or before date due. MAY BE RECALLED with earlier due date if requested. DATE DUE DATE DUE DATE DUE 5I08 KIProj/AodPras/CIRClDaIaDuaindd STUDENTS’ RETURN TO COMMUNITY COLLEGE: A MIXED METHODS EXPLORATION OF POSTSECONDARY EDUCATIONAL PATHS THAT INCLUDE AN INTERRUPTION IN ATTENDANCE By Carolyn K. Ozaki A DISSERTATION Submitted to Michigan State University In partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Higher, Adult, and Lifelong Education 2009 ABSTRACT STUDENTS’ RETURN TO COMMUNITY COLLEGE: A MIXED METHODS EXPLORATION OF POSTSECONDARY EDUCATIONAL PATHS THAT INCLUDE AN INTERRUPTION IN ATTENDANCE By Carolyn K. Ozaki This mixed methods study was designed to contribute to exploration of the reasons students returned to college, stopout, and the factors that influenced that decision. In the study’s first part I conducted a quantitative analysis of the relationship between factors associated with a high risk for departure and community college students who returned to school after departure. The second part of the study was a qualitative inquiry into the educational paths of community college students who stopped out, but were currently enrolled. The third article was an exploration of the intersection between the quantitative and qualitative results. The research questions addressed across all three articles were: 3) Why do community college students return after a period of nonenrollment?; b) Which factors influence the decision to return?; c) What is the influence of these factors on the decision to return to return? In the first part, I used correlations and logistic regression to examine data from the Beginning Postsecondary Survey 96/01 , focusing on community college students who departed college between 1996 and 2001 (n=779). The final model had four significant variables associated with who stays out and who returns to school. Students with mixed enrollment were more likely to return to college, while students who were divorced/widowed/separated marital status, job skill development as a reason for initial enrollment, and minimal student disadvantage were less likely to return. ..t ' I..lbr y... .J’ The purpose of the second part of the study was to explore the external and internal reasons and factors involved in students’ decisions to return to college after an extended absence. This study specifically sought to explore the role of students’ concepts of who they might be (or want to avoid becoming) in the college and career domains of their lives (possible selves). Analysis of the interviews revealed two different educational paths that were partially shaped by students’ college possible selves as they initially entered college. The different educational paths and influence of possible selves and other important factors came to light as students discussed critical decision-making points along their journeys. Changes in students’ possible selves also resulted in difi'erent attitudes and approaches toward school, resulting in more academic success and persistence. The purpose of the third article was to present the results of the quantitative and qualitative studies and analyze the intersection of the data. While the results reported for the quantitative portion of the study remain unchanged, the data from the qualitative study were reanalyzed and coded to better understand and expand on the ways that the significant risk factors in the quantitative model appeared in and influenced decision- making about college departure and reentry among the participants in the qualitative study. In general the qualitative data supported or partially mirrored the quantitative results and expanded upon the results to demonstrate how these risk factors manifest and influence students personal and academic lives in relationship to their educational persistence decisions. Copyright by CAROLYN K. OZAKI 2009 ACKNOWLEDGMENTS While I may be the sole author of this dissertation, I certainly didn’t write it alone. There are scores of faculty, friends, and family that I must acknowledge and thank for their support and help in this process. First, I am very grateful to my dissertation committee, Kristen Renn, Marilyn Amey, Jim Fairweather, and Matt Diemer, for all of their guidance, feedback, and support throughout the dissertation process. I especially want to thank Kris Renn for taking an active role and interest in my development as a researcher, teacher, and scholar. Your friendship and mentoring have enabled me to grow and achieve my academic and career goals. I owe Marilyn credit for introducing me to the area of community college scholarship, which shaped the direction of my own research interests. Thank you for ushering me into such a supportive and encouraging group of scholars. Thank you to Jim and Matt for being my “go to” guys for statistics help and guidance. You responded to a myriad of paniced emails with patient guidance and direction. Finally, I am gratefirl to my entire committee for the thoughtful and specific feedback they provided from proposal through dissertation. You helped me write better, convey my thoughts more clearly, and prepare a dissertation that can help me transition into my new role as a faculty member. I also owe a number of faculty and students within MSU’s College of Education for assisting me with the qualitative part of my dissertation. First, I am very thankful to Dr. Barbara Schneider and Jeff Keesler, for without their help there would have only been a qualitative study. . .no mix in “mixed methods.” When I took Barbara’s Sociology of Education class she emphasized graduate student use of large databases. Thank you, 'a \\.. A ‘1'... a. - AIL u .1 ~ . U N. “I. I .n, b 1 I; Barbara, for following through on those words and adding me to your license and requesting the BPS database on my behalf. I also owe Jeff a giant “thank you” for making all of that happen. Finally, thank you to Yisu Zhou and Yun-Jia Lo for being my statistics ambassadors. You helped me to shape my qualitative study. . .and understand it! Thank you to the institution that allowed me to come into your halls and recruit and interview students for hours on end throughout October. Specifically, I am very grateful to Gail Ives for doing much of the footwork for me on campus. The extra effort you put forth for my dissertation was beyond any expectations. Also, thank you to the 52 students who took time out of their schedules to share a piece of their stories with me. Their candor and openness allowed me to learn more about student academic progress. My family and fiiends have been amazingly supportive, especially since I’m sure I was not always the easiest person to be around. Melissa, Sheryl, and Debbie, thank you for being a listening ear and people to vent to. You kept me emotionally sane. Debbie, thank you for reading ALL those drafts. Remember, I’m here to return to favor. Gretchen, Jim, and Kacy, I am thankful that I had such good fiiends and cohort-mates to navigate our program with, even when it felt unbearable you made it fun. Dan, I am thankful that you let me take advantage of our friendship in order to get an “outsider’s” perspective of my drafts. Your continued fiiendship is so important to me and I hope I can be a source of support to you as you finish up you dissertation. Gary, your loving patience and encouragement throughout this last year sustained me, thank you. Finally, Mom and Dad, your support allowed this experience to happen. The love and encouragement you raised me with was foundational to the person I’ve developed into. You provided a safe environment that let me know I would always have a loving vi kl...“ I p, ' h§§hx family to return to, yet you armed me with the courage to pursue goals and experiences that were new and intimidating. You have always encouraged, supported, and prioritized my education, thank you for passing on those values. Thank you for loving me and raising a child confident and motivated enough to pursue her goals. Moving to Michigan State University, away from my family and fiiends in California, was a scary endeavor that promised the best education and preparation for my future career. Now that I am on the far end of that decision I can see that MSU and the individuals I met here made good on that promise (or hope). As my father has said, “This is where you were supposed to go. This was the best place for you.” For all of those who endured this past year of dissertation writing coupled with job searching with me, my acknowledgments and gratitude are no measure of the significant you played in my ability to complete this dissertation and degree. vii l ‘ 5 ll 14;. TABLE OF CONTENTS LIST OF TABLES ........................................................................ x LIST OF FIGURES ........................................................................ xi CHAPTER 1: INTRODUCTION Introduction ........................................................................ 1 Statement of Problem ............................................................. 3 Research Questions ............................................................... 8 Research Design ................................................................... 9 Significance of Study ............................................................. 11 Structure of the Dissertation ...................................................... 12 Summary ........................................................................... 13 CHAPTER 2: METHODOLOGY IntrOduction ........................................................................ 1 5 Purpose of the Study ............................................................... 15 Conceptual and Theoretical Framework ........................................ 16 Research Approach ........ 18 Research Design ................................................................... 20 Limitations .......................................................................... 42 Summary ............................................................................ 43 CHAPTER 3: ARTICLE l—QUANTITATIVE STUDY Introduction .......................................................................... 44 Background ......................................................................... 46 Method .............................................................................. 51 Results ............................................................................... 59 Discussion ........................................................................... 63 Implications and Future Research ................................................ 72 Summary ............................................................................ 75 CHAPTER 4: ARTICLE 2—- QUALITATIVE STUDY Introduction ......................................................................... 76 Background ......................................................................... 77 Methods .............................................................................. 85 Findings .............................................................................. 90 Summary ............................................................................ 111 Implications for Theory, Practice, and Future Research ...................... 113 CHAPTER 5: ARTICLE 3—MIXED METHODS Introduction ......................................................................... 1 19 Background ......................................................................... 122 viii Research Design ...................................................................... 126 Part A: Quantitative .................................................................. 127 Methods ....................................................................... 127 Data and sample ............................................................. 128 Analysis ....................................................................... 129 Dependent variable .......................................................... 130 Independent variables ...................................................... 13 1 - Results ........................................................................ 132 Part B: Qualitative ................................................................... 136 Method ........................................................................ 136 Data source .................................................................... 136 Analysis ....................................................................... 138 Findings ................................................................................ 139 Discussion .............................................................................. 150 Summary ............................................................................... 160 Limitations ............................................................................. 161 Implications and Further Research ............................................ . . ...162 Conclusion ............................................................................. 165 CHAPTER 6: CONCLUSIONS Introduction ........................................................................... 166 Article 1 Summary ................................................................... 167 Article 2 Summary ......... _ .......................................................... 170 Article 3 Summary .................................................................. 174 ThemesAcross Articles ............................................................. 178 Implications and Suggestions for Practice ........................................ 180 Future Research ...................................................................... 184 , Conclusion ........................................................................... 185 APPENDICES Appendix A: Full Description of Variables for the Quantitative Study ...... 186 Appendix B: Glossary and Description of Variables ........................... 187 Appendix C: Frequency and Chi-Square Table ................................. 192 Appendix D:_Recruitment Poster .................................................. 196 Appendix E: Recruitment Email to Students ..................................... 197 Appendix F: Research Participant Information and Consent Form ........... 198 Appendix G: Interview Protocol and Demographic Questionnaire ........... 201 REFERENCES ................................................................................. 203 ix w. ‘nen a I «n L. l J y A y .—l\ he. urn . ...y in .4 T13 T i v . LIST OF TABLES Table 1 Research Questions and Accompanying Data and Analysis Plans ........... 21 Table 2 Categories of variables ............................................................. 26-27 Table 3 Logistic Regression Predicting Likelihood of Returning to College After N onenrollment ...................................................................... 30 Table 4 Themes and Subthemes for Qualitative Part of Study ......................... 39-40 Table 5 Additional Themes and Subthemes for Qualitative Part of Study ............ 42 Table 6 Characteristics of the Sample ..................................................... 53-54 Table 7 Dependent Variable ................................................................ 57 Table 8 Variables in Code .................................................................. 60 Table 9 Logistic Regression Predicting Likelihood of Returning to College After Nonenrollment ..................................................................... 63 Table 10 The Percentage of Students Single, Married, and Divorced/Separated/Widowed Students Who are Low, Minimal, and Moderate/High Risk of Attrition/Nontraditional ............................. 67 Table 1 l, Dependent Variable ............................................................... 131 Table 12 Logistic Regression Predicting Likelihood of Returning to College Afier Nonenrollment .............................................................. 13 5 Table 13 Descriptions of Variable and Codes ............................................ 186 Table 14 Frequency Table for Variables .................................................. 192—194 Table 15 Chi-Square Tests for Independence: Dependent variable (Stay_Stop2) and Independent Variables ...................................................... 195 ‘r‘t.. l.-.- 5 ii ‘1' LIST OF FIGURES Figure 1 Stopout Cycle ..................................................................... 92 Figure 2 Student Paths and Educational Events ......................................... 111 {/1 ‘A‘ v! «V . 9;~ “ -s\ r .. P L” pa? M“? ‘4 .VII V 3;. Y CHAPTER 1: INTRODUCTION Introduction “A test of leadership: Charting the future of US. Higher Education” (2006), a report commissioned by the US. Secretary of Education to review the status of higher education, provided recommendations for the future. It identified four major issues facing U.S. post-secondary education, critical findings associated with these issues, and recommendations to improve access to and success in postsecondary education. The first recommendation called for every student in the nation (to) have the opportunity to pursue post-secondary education. We recommend, therefore, that the US. commit to an unprecedented effort to expand higher education access and success by improving student preparation and persistence, addressing non academic barriers and providing significant increases in aid to low-income students (US. Department of Education, 2006, p. 16). Yet, today there are a significant number of students who do not achieve and accomplish their educational goals. Low-income students are less likely to graduate from college than more affluent students (Adelman, 1997; Renner, 2003). One example of this gap is that the college persistence and goal attainment rate in 1998 among low-income students was 59% in comparison to the 71% of higher income students (Choy, 2000). This gap was also evident between racial/ethnic minorities (i.e., African-Americans, Native Americans, and Latinos) and White students (Adelman, 1997; Bradbum, Berger, Li, Peter, & Rooney, 2003; Hagadorn, Siadat, Fogel, Nora, & Pascarella, 1999; Jacobson, Olsen, King Rice, Sweetland, & Ralph, 2001; Renner, 2003). For example, the attainment of bachelors degrees by race in 2005 were as follows: White—20%; Black—12.5%; Hispanic—8.5%; Asian—32%; American Indian—10% (KewalRamani, Gilbertson, Fox, & Provasnik, 2007). Students at higher risk of not graduating were also more likely to attend community colleges, where uninterrupted attendance and persistence to educational goal achievement are less likely than for those students at four-year institutions (Adelman, 1999; Horn, 1996). As institutions that serve a larger proportion of hi gher-risk students, community colleges are charged with supporting and assisting their students to achieve a broad spectrum of educational goals (Berkner, Horn, & Clune, 2000). Furthermore, many of these students face numerous personal, academic, and institutional challenges to their educational goals (Hoachlander, Sikora, & Horn, 1998; Horn, 1996). One common occurrence is the choice of students to leave school for a period of time and then return to continue their education at a later date (Horn, 1998). This phenomenon, stopout, is associated with multiple factors—poor academic performance, financial burden, and family responsibilities are among the most common reasons (Horn, 1998). What is troubling to institutions about this issue is that students who experience an interruption in enrollment are more likely to experience subsequent stopouts and are less likely to graduate (DesJardins, Ahlburg, & McCall, 2006). Furthermore, research on stopout is focused on the why students leave, how stopout affects student outcomes, and what differentiates stopout from dropout behavior. Yet, little is known about why students return after a period of stopout and how they make the decision to return. This infrequently studied issue has potential implications for the 30% of students who leave college after their first year (Horn, 1998) and the ability of institutions to assist ..1 3.9L.» hr, :21 LIV-Edi] ?‘.~ 6 . ~ku“ L g - D LH4 l ‘1 . ~‘ u ‘.~u",.‘ Q» ,‘ I... students who want to return to school. This introduction explores stopout behavior and why it is problematic, the significance of this topic to students, institutions, and higher education, and how this study expanded an understanding of this issue. Statement of Problem In this section I discuss community college students’ persistence and stopout behavior. Additionally, I present a theory of possible selves as a useful lens for exploring students’ decisions to return to college after a period of absence. Persistence for community college students While there are concerns about the academic achievement of students at higher risk of dropping out across the educational pipeline, in the postsecondary sector a primary concern is retaining students through their academic goals and helping them to endure the academic, organizational, and personal factors that challenge their ability to persist. Across postsecondary education, approximately 30% of students will depart from higher education (Horn, 1998). Yet, the relative percentage of students who leave fi'om four-year institutions (16%) is not equal to the percentage of students who leave from two-year colleges (44%) (Horn, 1998). This inequality draws attention to the community college sector for the study of persistence (Horn, 1998). The open access, low-cost characteristics of two-year colleges have broadened access to higher education for students facing challenges such as financial hardship, poor academic preparation, a lack of basic English language, or proficiency skills (Grubb, 1999). As a result, approximately 54% of community college students enter with at least one characteristic that places them at risk for not completing postsecondary education (Hoachlander et. al., 2003). These characteristics are typically linked with nontraditional ..(1 -iau' 3"» ism-N... vat-tu- “I... ‘.A ”f... 5" ‘I- 3 ~..‘ 11 Rat. -—--.. student status such as delaying enrollment between high school and college, part-time attendance at the first institution attended, working full-time while enrolled at the first institution, completing high school by certificate or GED, having a child, being a single parent, and being financially independent (F eldman, 1993; Horn, 1996; Horn & Premo, 1993; Lee, 1996). Furthermore, while personal and financial reasons are among the most often cited reasons for departure, a lack of academic preparation and poor academic performance are also provided as reasons for leaving higher education (Bonharn & Luckie, 1993; Lee, 1996; Ogletree, 1992). In addition, the majority of students with risk factors (e. g., part-time students, single parenthood, poor academic preparation, and poor academic performance) for early departure choose to attend community versus four-year colleges for their first experience with higher education (Adelman, 1999; Horn, 1996). The figures previously mentioned demonstrate that (a) community colleges continue to attract students at a higher risk for departing from college before they achieve their educational goals and (b) community colleges are important and authentic settings to study students at a higher risk of leaving postsecondary education prior to completing their educational goals. One of the challenges that these institutions confront when considering how to improve student retention is the stopout phenomenon. Stopout behavior A significant challenge to institutional retention and overall postsecondary educational persistence is a phenomenon that occurs when students leave college after their first year: stopout. Horn defined a first-year stopout as “a beginning student who interrupted his or her enrollment in the first year with a break of at least four months before reenrolling. . .this includes students who finished their first year, but did not In ‘ .Ir Q 'w’r “Mu.“ ‘1 I‘.’ “‘ .W 1P..-:Li"ll , W’ . I. “ 'v I :‘W‘c‘g. Ref-‘- -. 677.5 C” 4.1m. .lugg‘ll‘} y. « '4 '3”- ""‘sc. \ "er HJCT ed.“ “:4 reenroll for a second year” (1998, p. 4). In contrast, a dropout is a “nonreturnee who had not accomplished an educational goal and stated specifically that he or she no longer plan(s) to work toward that goal” and an optout is “a nonreturnee whose education goal was met; they opt(s) out of firrther study related to that goal” (Bonharn & Luckie, 1993, p. 545). Of concern to institutions is that students who experience an interruption in enrollment are more likely to experience subsequent stopouts and are less likely to graduate (DesJardins et al., 2006). These outcomes bring into focus the potential problems with stopping out for students and raise the issue of how institutions can help students maintain continuous enrollment and provide support for students to return to college. In recent studies of attrition and retention, researchers began to recognize and take into account this stopout phenomenon. Of all students who attended college in 2006- 2007, approximately 65% of students went to a four-year institution and 35% attended a two-year college (Provasnik & Planty, 2008). Approximately 30% of students in all of higher education leave after the first year, yet half (57%) of those who leave will return to higher education within five years (Horn, 1998). Of those who leave from postsecondary education, 63% are from community colleges, 22% are from four-year institutions, and 15% are from other types of schools. Yet, 50% of those who depart from the two-year sector return, while 64% of those who left a four-year institution return within five years. This gap in returns indicates that students who leave after their first year of school and begin at a two-year institution are less likely to return to college than students who begin at four-year institutions, but represent a greater number of students overall. This situation not only provides support for the need to study stopout behavior at community colleges, r" hm» vq -o.- \ u- r... JL .A. I” ’m. is. ‘ J] but also supplies logic for studying this phenomenon at two-year versus four-year institutions. Furthermore, students who stopout primarily return to community colleges with 79% of students who start at a two-year and 24% of students who start at a four-year institution returning to community colleges (this includes students who return to the same institution at where they began and students who transfer from another two- or four-year institution) (Horn, 1998). The previous figures indicate that a majority of students who depart higher education after the first year both leave from and return to community colleges. In light of this evidence, exploring stopout behavior at community colleges is not only reasonable, but preferable as it is these institutions that most ofien encounter stopout behavior and, therefore, are likely to be most vested in responding to it. Furthermore, knowing and understanding not only why students leave but why they return, how to facilitate that return, how they think about that process, and what is most influential in that decision can be helpful to researchers and practitioners to (a) encourage and help students choose not to leave in the first place and (b) identify ways to support students to return to their educational goals. There are multiple important influences and factors that contribute to understanding how and why students decide to return to college. Based on previous research related to aspirations and goal achievement, one of these factors is likely to include the academic and career hopes students have for themselves and the ways that students view their hopes and fears of achieving and becoming these possible selves. Possible selves Research has demonstrated that the academic aspirations students hold for themselves are related to their ability to persist and achieve their goals (Kerpelman, Shoffner, & Ross-Griffin, 2002; Leondari, 2007; Norman & Aron, 2003; Oyserman, Bybee, Terry, 2006; Pizzolato, 2006; Yowell, 1999, 2000). One approach to this concept is the study of possible selves. Possible selves are the cognitive representations and ideas of what individuals believe they might become, what they would like to become, and what they are afraid of becoming—the cognitive components of their hopes, fears, goals, and threats (Markus & Nurius, 1986). Markus and Nurius (1986) claim possible selves are important for two reasons. First, they function as incentives for future behavior and serve as a cognitive bridge between who an individual is now and who he or she can be in the future, between their motives and goals and the behavior/action necessary to achieve those goals (or avoid particular selves) and maintain their motivation. Motives are represented as “dispositions” to strive for particular incentives (goals) or avoid negative incentives (threats). Possible selves cognitively represent these motives and the plans/paths for achieving them. Markus and Nurius (1986) suggest that desire or motivation is not sufficient for motivating behavior, but that the desire must be translated into a vision of the self as healthy, active, and strong and must be accompanied by specific plans and strategies for becoming these possible selves. These possible selves are cognitive representations of the incentives for mastery, and without them there should be little instrumental behavior in the direction of mastery. (1986, p. 961) Furthermore, balance between hoped for and feared possible selves is related positively to acaderrric persistence among undergraduates and Afiican-American middle- .5.» O‘- ...‘z ..-a ' Anon . . ‘ school students (Oyserman, Gant, & Ager, 1995) and a strong relationship between self- regulation strategies and academic possible selves is related to academic achievement and persistence (Oyserman, Bybee, Terry, & Hart-Johnson, 2004, Oyserman et al., 2006; Oyserman & Fryberg, 2006). In relationship to the issue of stopping out, it is reasonable to believe that in the process of leaving and returning to postsecondary education, who that person believes he or she can be academically and in his or her career could play a part in what that student’s goals are and become, which academic goals he or she pursues (e. g., returning to school, changing a major upon return, not returning to higher education), and, ultimately, whether or not he or she returns to college after an absence. Utilizing a possible selves framework to explore the role these constructs play in the decision to return to college and, conversely, how college and academic possible selves change as a result of stopping out can provide insight into why students choose to return to college, how students operationalize that decision into action, and how students manage their college and academic hopes, fears, and goals in relationship to challenges. Research Questions A significant amount of previous research on achievement and persistence focused on attrition and retention, but less explored why students return to higher education after an absence (Braxton, Hirschy, & McClendon, 2004; Swail, Redd, & Perna, 2003). In this study I addressed important questions about stopout not significantly explored in previous research: 1. Why do students return after a period of non-enrollment? 2. What factors impact this decision? 3. What is the influence of these factors on the return to college? l- ',w~~ . ‘- usl li‘f"? 4~( F- r“ .J."~'_-‘ r .,. ‘ I \ .T‘ . . ’_ I I’f‘ ‘ '0: N u 4. What role do college and career possible selves play in the decision to return to college after an absence? 5. How does stopout impact individual’s academic and career possible selves? Research Design This study was a two-part mixed methods study that investigated the retum-to- college aspect of stopout behavior. Each part focused on a specific aspect of understanding the decision to return to college. Part One: Quantitative. The first part focused on the analysis of a large scale, national dataset (Beginning Postsecondary Study: 1996/2001) for factors that affect students’ decisions to reenroll after leaving college for a period of time and differences in characteristics (e. g., race, income, work level, parental status) between students who stayout and stopout. It addressed the question, “What factors affect the decision to reenroll?” Specifically, I focused on the relationship between characteristics associated with a high-risk of student departure, G.P.A., reasons for enrolling, and the probability of return. Contributing to the design and scaffolding of this study was the literature on persistence in college, specifically for community college students. Prior research on retention and persistence identified many risk factors for departure. Among the most salient were traits associated with non-traditional student status (e. g., delaying enrollment between high school and college, part-time attendance at the first institution attended, working full-time at the first institution, completing high school by certificate or GED, having a child and being a single parent, being financially independent, and being older) (Feldman, 1993; Horn, 1996; Lee, 1996), a lack of academic preparation and poor academic performance (Bonharn & Luckie, 1993; Hoachlander et al., 2003; Lee, 1996; Ogletree, 1992), and having no degree objective (Choy, 2001). These factors served as independent variables for analysis of the decision to return to college and as the basis for understanding how significant factors influence the decision process itself. The binary dependent variable represented dropout and stopout student outcomes. Dropout represented students who left college without return and stopout represented student who left college but returned during the survey’s six years. Correlations and logistic regression were used for analysis. Part Two: Qualitative. The second part explored in more depth the decision- making process to return to college. The guiding research questions for this part of the study were, “What does the decision-making process look like? What role do college and career possible selves play in the decision to return to college after an absence?” I interviewed 48 community college students about their educational experiences and decisions to leave and return to college. To address the first question, I asked students to describe why they left school and what their decision to return to college looked like. In order to understand why students return to school it is important to understand why they left in the first place and what changed, if anything, within those reasons for leaving. Therefore, I used the risk factors included as variables in the quantitative part for specific points of exploration in the qualitative part. In addition to continuing a persistence framework from the first to the second part, I used possible selves theory to further understand how a student’s conceptions of his or her possible academic and career selves influence the decision to return. Balance among and the presence of hoped for and feared possible selves have been shown to 10 influence student outcomes and achievement in secondary and postsecondary students (Oyserman et al., 2004; Oyserman et al., 2006; Oyserman & Fryberg, 2006) and the ability to translate those visions of one’s self into strategies for achievement (Oyserman, 2004; Pizzolato, 2007) are critical in the achievement of goals. Understanding how participants conceive of their possible academic and career selves prior to departure and when returning can provide insight into changes in the perception of one’s possible selves and strategies for the achievement of possible selves that are important in the decision to reenroll. Furthermore, this study can also provide insight into how and what impacts change in the representation of possible selves. Finally, in addition to using possible selves in the construction and analysis of the qualitative data, I employed learner disposition theory to better understand data during analysis. A person’s learner disposition is his or her approach to and perceptions of learning. Learners’ beliefs about the nature of knowledge, the purpose of education, their abilities based on prior learning experiences, and the value placed on particular areas of study and learning experiences are important factors in their disposition (Bloomer & Hodkinson, 1999, 2000). I found that this theoretical approach helped to explain the changes made in the academic and learning domain of students’ lives. Significance of Study This study was significant and has the potential to contribute to research and practice in multiple ways. First, stopout behaviors have not been widely studied in general, and, more specifically, an effort to understand the decision to reenroll is all but ignored in the current literature. Investigation into this aspect of stopout helps researchers and practitioners not only understand why students leave in the first place, but why and 11 2 I A... 5..“ -—.‘ >.J\. . ‘ \ '._--_ Wu. -. t ‘4 ‘M‘ 3%.- "K. what is significant in the decision to return to college. Second, understanding how students manage and resolve the barriers that led them away from college after their first year is critical to creating campuses and supportive policies that will allow and invite students to return to higher education. Practitioners at the institutions, community colleges, that encounter stopout most often can be better equipped to understand why students leave and how they can support and encourage them to return. Finally, the second part of this study provides a deeper and better understanding of how possible selves influence decision-making in general and the decision to persist in the postsecondary sector specifically. Also, on a theoretical level the study contributes to identifying what influences and insti gates change within possible selves. Structure of the Dissertation The mixed-methods approach to this study lent itself well to writing and organizing the findings in an alternative format to a five or six chapter traditional dissertation format. Specifically, I followed a journal article format for the organization of this document. Dissertations written in this format generally include an introduction, a series ofjoumal-style articles (number as agreed upon with the dissertation chair), and a conclusion that ties the articles together, reminding the reader of overarching significance and potential applications for the study (Thomas, Nelson & Silverrnan, 2005). Advantages of the journal article format are that it allows the author to become more familiar with this style of writing, further preparing and training him or her for writing for publication, and it increases the odds for publication of the dissertation as it will require less transformation in the future (Thomas, Nelson, & Silverman, 2005). This dissertation is well-suited to be written in a journal article format because the individual 12 ~-' “I .o rt! b -.L. -. ‘ V? \ .' A) 5 U ‘1 ... ‘\ 5., \ ‘ ‘ " .- . h . .lf'“.t . , . n- I parts (i.e., quantitative and qualitative) are related by topic, but are designed to investigate individual questions and do not rely on one another. Therefore, the parts can “stand alone” and be reported as individual studies. In this dissertation I began with an introduction that presented the research questions, study design, and overall importance of the study. A methodology chapter provides detailed information on data collection and analysis of the quantitative and qualitative parts of the study. The first article presents the logistic regression analysis of data from the Beginning Postsecondary Survey: 96/01. I explored the relationship between factors associated with high-risk departure and the decision to dropout or return to college (stopout). The second article is a result of the qualitative portion of the study and describes the multiple paths that the participants are taking through college and their decision making process to leave and return. The third article is a discussion of the intersection between the quantitative and qualitative parts of the study. I conclude with a section that summarizes the findings fiom each article and focuses the reader on the importance of the findings. Summary The percentage of students departing from higher education is a concern to the postsecondary sector. Understanding why students decide to leave and how to retain more of them is a regular theme and issue addressed within practice and research. One group of students that is often blindly combined within the departure statistic are the nearly half of students that leave after the first year who decide to return to school after an absence. With such a significant number of students who decide to reenroll, it is 13 - W .1“. ‘ l'. ' mi ah. . I Re “In .Lx._xr. r ' V _,.- \In‘bs. important to better understand their paths throughout postsecondary education and what influences their decision to return. In the following sections I include three individual articles in which the study’s findings are presented. In the final section, I discuss how these articles come together to extend the research and literature on stopout. I expand on how the articles relate and build on one another to contribute to a fuller understanding of stopout and the varied paths students take through postsecondary education. 14 d. 1,» Mi CHAPTER 2: METHODOLOGY Introduction The goal of this chapter is to describe the methodology and approach I used for this study. First, I review the purpose of the study and research questions. Second, I discuss the framework and research approach for this mixed methods study. Third, I describe and review the steps for data collection and analysis of each part: a) For the first part I include a description of the participants, data collection procedures and protocol, and data analysis procedures and b) for the second part I discuss the survey, access to and preparation of data, and data analysis procedures. Purpose of Study The purpose of this study was to explore stopout behavior and which factors and processes are potentially related to and influence the decision to reenroll after an absence from postsecondary education. To facilitate this investigation I completed a two-part mixed methods exploratory study that utilized data analysis of the Beginning Postsecondary Student 1996-2001 (BPS: 96/98/01) survey, the longitudinal component of the National Postsecondary Student Aid Study (N PSAS: 96) and interviews at a Michigan community college. This study explored the following research questions: Why do students return after a period of non-enrollment? 1. What factors impact this decision? 2. What is the influence of these factors on the return to college? 3. What role do college and career possible selves play in the decision to return to college after an absence? 15 v l 5“ ‘v V ' 4 4. How does stopout impact individual’s academic and career possible selves? Conceptual and Theoretical Framework The design of the this study called for two parts, quantitative and qualitative, that focused on related but different aspects of students’ decisions to return to college after an extended period of non-enrollment. Each part of the study drew from different aspects of literature to provide a framework for the research design that is grounded in prior research and theory. Part One The retention literature and research regularly assume that once students leave college they depart postsecondary education permanently. Yet, in the last 15 years new research has brought to light the stopout phenomenon and begun to ask how stopout relates to persistence. While the stopout literature is not extensive, it has almost exclusively focused on demographic characteristics of students who stopout (Bonharn & Luckie, 1993; Horn, 1998), the reasons that students leave (Bonham & Luckie, 1993a; Hoyt & Winn, 2004), and how stopout is related to persistence (DesJardin, Ahlburg, & McCall, 1994; DesJardin, Ahlburg, & McCall, 1999; DesJardins, Alhburg, & McCall, 2002; Herzog, 2004; Johnson, 2006; Scott & Kennedy, 2005; Stratton, O’Toole, & Wetzel, 2004). Yet, few studied stopout at community colleges or the question of why students return to school. This portion of the study extends understanding of factors that influence students’ return to school in hopes that that information can help institutions attract lapsed students back to college and assist reenrolled students to maintain continuous enrollment. l6 i‘ i ~ . \.. - ~- . ' .LL‘ I selected the independent variables for analysis from the persistence and stopout literature, including factors that prior studies have demonstrated to be particularly important for non-traditional and community college students, in addition to characteristics of students who chose to stopout. Part T wo The second part of this study focused on the same overarching question: “Why do students return after a period of nonenrollment?” The qualitative approach allowed for more depth and detail into the investigation of this question. While the first part of the study addressed factors that affect reenrollment, the second part addressed the decision- making process to return and how possible selves theory factored into that decision. The majority of retention literature focused on the influence of psychosocial factors on commitment (Beil, Reison, Zea, & Caplan, 1999; Napoli & Wortrnan, 1998, Nora, Attinasi, & Motonak, 1990), social integration (Beil et al., 1999; Borglum & Kubala, 2000; Napoli & Wortrnan, 1996, 1998), self-esteem (Bean & Eaton, 2001/2002; Napoli & Worrnan, 1998), and attitude theory (Bean & Eaton, 2001/2002; Polinski, 2002/2003). Most of these studies are based on elements of Tinto’s or Bean’s models and do not explore psychosocial theories that could further explain student persistence and attrition behavior. One such tool is possible selves theory. Research on possible selves theory demonstrated the influence possible selves has on educational motivation, achievement and persistence of youth and adolescents (Cross & Markus, 1994; Kerpelman, Shoffner, Ross-Griffin, 2002; Lee & Oyserman, 2007; Leondari, 2007; Leondari, Syngollitou, & Kiosseoglou, 1998; Norman & Aron, 2003; Oyserman, Bybee, Terry, & Hart-Johnson, 2004; Oyserman, Bybee, & Terry, 2006; 17 |l‘ Oyserman & Fryberg, 2006; Pizzolato, 2006, 2007; Yowell, 1999, 2000). In the process of leaving and returning to school, students are likely to have to evaluate their educational career goals. Depending on why they left, they are also likely to explore whether those hoped for selves can be maintained or must be adjusted in order to return to postsecondary education. This study explored the important mechanisms and aspects of possible selves in relationship to decision-making and return to college. Included were probes for self—schemas, balance, relationship to motivation, and types and domains of hoped for, feared, and expected selves. This provided a more complete and complementary picture of stopout behavior and insight into important aspects of stopout decision-making in relation to identity, motivation, and persistence. Research Approach This study utilized mixed methods, explanatory design approach. A mixed methods study involves the collection or analysis of both quantitative and/or qualitative data in a single study in which the data are collected concurrently or sequentially, are given a priority, and involve integration of the data at one or more stages in the process of research. (Creswell, Plano Clark, Gutrnann, & Hanson, 2003, pg. 212). Ideally, the use of mixed methods research will result in providing a better understanding of the problem than either method, quantitative or qualitative, would yield alone (Creswell & Plano Clark, 2007). Many researchers on both sides of the quantitative-qualitative methods debate argue that the paradigms are incompatible and those who use both are “doomed to fail” (Teddlie & Tashakkori, 2003). Yet, proponents of mixed methods counter that it allows researchers to simultaneously verify theory, 18 . "If 5.; fi-«. ~sL. ." .F"! ._ . .M F. I W- I ‘\l -\ \ J, ‘1“ - . g. -« v e. .r , through confirmatory questions, and generate theory, with exploratory questions, in the same study (Teddlie & Tashakkori, 2003; Creswell & Plano Clark, 2003). They also claim that inferences from mixed methods are stronger because they can offset disadvantages associated with each method (Teddlie & Tashakkori, 2003; Creswell et al., 2003; Creswell & Plano Clark, 2007). Finally, in response to the criticism that quantitative and qualitative phases from the same study could have contradictory findings, supporters of mixed methods argue that this is not necessarily a bad outcome because divergent findings can lead to reexamination of conceptual frameworks and assumptions (Erzberger & Prein, 1997; Teddlie & Tashakkori, 2003). I approached this study primarily from a dialectic paradigm. From a dialectic stance, the researcher purposely involves multiple methods and approaches in order to create interactions and possibly contradictions about the assumptions and epistemology underlying each approach in order to expand understanding of human phenomena (Green & Caracelli, 2003). I purposely included and emphasized the qualitative portion of this study because the majority of extant research on stopout has utilized quantitative methods, and I sought to extend understanding of this phenomenon through a qualitative, inductive approach. The quantitative portion of this study explored the relationship between variables identified in research as related to a higher-risk of departure and the decision to return to postsecondary education. I analyzed the findings from the quantitative portion of the study separately and in relation to findings from the qualitative portion of the study, but the theoretical thrust of the study focused on the exploration of the decision-making process to return to school and the role of possible selves in that process. The exploratory l9 ~‘u it; Vs. nature of this portion of the study lent itself to understanding the meaning—making process among participants in relation to stopout and theory extension of possible selves. As a result, this study maintains a decidedly inductive focus, represented as quan—)QUAL (Morse, 2003). Research Design Again, this study utilized a two-part exploratory, mixed methods design. The first part was a quantitative analysis of the Beginning Postsecondary Student 1996-2001 (BPS: 96/98/01) survey, the longitudinal component of the National Postsecondary Student Aid Study (NPSAS: 96). Within this study I limited my analysis to the data from students who began postsecondary enrollment at a two-year, academic focused college left postsecondary education within the survey’s five years. Part two involved semi- structured interviews of Michigan community college students who previously left college for at least a semester term or four months (not including summer) before returning. In this section I provide an overview of the design in relationship to the research questions and discuss in detail the procedural and analytical plans for each part. In Table l I describe the analytical approach and data collection method used for each research questions. 20 AI. in) ,7 Lu.» J»;,\ 0.. I , a» Table 1 Research Questions and Accompanying Data and Analysis Plans Research Question Data Used Analysis Approach Why do students return after a BPS: 96/01 and Logistic Regression period of nonenrollrnent? Interviews Pattern Coding/Thematic Analysis Which factors impact the BPS: 96/01 Chi-square test decision to reenroll? Logistic Regression What is the influence of BPS: 96/01 and Logistic Regression factors on the decision to Interviews Pattern Coding/Thematic return to college? Analysis What role do college and Interviews Pattern Coding/Thematic career selves play in the Analysis decision to return to return to college after an absence? How does stopout affect Interviews Pattern Coding/Thematic individuals’ academic and Analysis career possible selves? Part One: Quantitative Dataset. The Beginning Postsecondary Longitudinal Study 1996-2001 (BPS: 96/01) is the longitudinal study of students who began their postsecondary education in 1995-96. The students took the survey only if they had not taken a postsecondary course prior to enrollment. The sample of students were the first time beginning students (FTB) who attended post—secondary institutions eligible for inclusion in the NPSAS: 96 (National Postsecondary Student Aid Study) and were NPSAS eligible individuals (for a discussion of eligibility see Wine, Heuer, Wheeless, Francis, Franklin, & Dudley, 2002). The NPSAS: 96 was a nationally representative survey designed to determine how 21 students and their families pay for college (Wine et al., 2002). Data were collected at three points in time. The first data collection was part of the National Postsecondary Student Aid Study 1995-1996 (NPSAS: 96). The second collection occurred two years later in 1998 and the third and final collection was collected six years after the beginning of the study in 2001. The NPSAS: 96 utilized a two-stage sampling design in which the researchers identified eligible institutions in the first stage and eligible students in the second stage. The institutional sampling frame was constructed from the 1993-1994 Integrated Postsecondary Education Data System (IPEDS) Institutional Characteristics (IC) file. Probabilities proportional to composite measures of size based on overall sampling rates by institutional size and type of student were used to select sample institutions. Researchers created additional implicit stratification within each institutional strata (level) by sorting each strata for institutional level, region, and institutional measure for size. The additional stratification was employed to ensure proportionate representation by institutional level, region, and size (of smaller institutions) (Wine et al., 2002). Researchers identified students in the second stage students through a list provided by sampling institutions ofNPSAS: 96 eligible students with student level (e. g., undergraduate, graduate, professional) information. Student sampling rates were determined based on four student strata: potential FTBs, other undergraduates, first professional students, and other graduate students. Researchers collected NPSAS: 96 data through records maintained by the institution and computer-assisted telephone interviews (CATI). A total of 12,410 cases and over 900 postsecondary institutions were eligible for BPS: 96. Of the BPS: 96 cohort, 10,300 completed the second interview (BPS:96/98). 22 The final interview, BPS: 96/2001, consisted of the respondents to BPS: 96/98, plus approximately 1800 additional FTB NPSAS: 96 students, resulting in a nationally representative sample of institutions and students (Wine et al., 2002). BPS: 96/01 relied on telephone and in-person interviewing for data collection. Topics covered in all interviews were background and demographic information, parental and student financial aid and loan status, enrollment history since the previous interview, undergraduate enrollment, postbaccalaureate enrollment, and postenrollment employment. This study utilized data across all three surveys. In addition to questions for demographic, financial, and FTB-determination information, data were collected that focused on issues of attainment and persistence during the first two collections. Important to this study, data distinguished dropouts, stopouts, transfers, and completers in addition to reasons for the enrollment decisions. Items to identify entering goals (e.g., vocational versus bachelor’s attainment), nontraditional status, and institution type were administered. Data collection in BPS: 96/2001 resulted in an 88.3% unweighted response rate, approximately the same as the previous interview (Wine et al., 2002). Access and preparation. The National Center for Education Statistics (N CES) has a web-based software application, Data Analysis System (DAS), that allows free, public access to NCES surveys. Originally, I planned to use the DAS for analysis, but it did not allow for the level of data manipulation and preparation that I required. Therefore, I needed to acquire the raw data from NCES. This requires a user license and following detailed guidelines for data storage. A faculty member from Michigan State University’s College of Education, Barbara Schneider, who already had a license and stores multiple 23 NCES databases for use, agreed to add me to her license and request the BPS: 1996/2001 data on my behalf. IRB was informed and permission was granted. To supplement my knowledge of secondary analysis and logistic regression, I worked with two tutors from the College of Education who are/were students in the MQM doctoral program. The students assisted by helping me to recognize the steps required throughout the quantitative study and to understand what the implications were for the decisions I made. Study sample. Once received, the data for selected variables were downloaded from the electronic codebook into a SPSS datafile. I will discuss the variables in the next section. I constructed filters to select only students who began at a public 2-year institution and had stopped-out or left postsecondary education without return during the six year study (1996-2001). While these students began at a two-year institution they may have returned to any type of college——technical, four-year, or two-year. To see a list of variables used to construct the filters, see Appendix A. With filters applied the sample total was n=779. A base weight was constructed for the BPS: 96/01 sample to adjust for the sub- sarnple of non-respondents from BPS: 96/98, constructing probabilities that take into account the stratification and probability proportional to size used in the selection of the sub-sample. A cross-sectional weight further adjusted for non-response bias. I developed the longitudinal weights by applying additional non-response adjustment to the final cross-sectional weight. The longitudinal weight for students who responded to all three rounds of the survey (B01 LWTl) was selected to adjust for subsampling and non- response of students who did not respond to NPSAS: 96 or BPS: 96/98 (Wine et al., 24 A.“ a... *- \ I». 2002). I selected this weight because the outcome variable and a significant number of independent variables were constructed from data across all three rounds. The weighted sample resulted in over-estimated significance levels due to the larger population size; as a result a relative weight was constructed to preserve the accuracy of the weighted estimates by reflecting the sample size (Thomas & Heck, 2001). To avoid removal of cases with normalized weight values of zero in analysis, a constant of .001 was added to the weights. This allowed for the cases to remain in the sample for analysis with a negligible affect (de Vaus, 2004). The use of clustering in the latter stages of large-scale surveys contributes to the nonindependence of observations and nonhomogeneity, violating the assumptions of simple random sampling. The typical statistical software treats data as if constructed using simple random sampling and not adjusting for effects from multistage clustering, resulting in variance estimation and standard error bias (Diemer, 2008; Thomas & Heck, 2001). In recognition of potential estimate bias resulting from a simple random sampling assumption, a more conservative alpha level was used to evaluate the model and predictor variables (p<.01) (Thomas & Heck, 2001). The percentage of missing data for the final regression model was 15.3% and within acceptable range (Allison, 2001). A dummy variable for Missing/Non—missing was constructed for each variable and run with the outcome variable to test for independence and potential relationships between stopout/dropout and missing-mess among the independent variables (de Vaus, 2004). There was no significance between any of the groups. 25 ’J .w- J. Variables. I drew on the community college and stopout literature to identify the variables I used in my analysis. I included variables previously identified as having a relationship with or impact on persistence and enrollment. The sample consisted of students who left college prior to June 2001. The dependent binary variable represented students who left and returned to college (stopout) at least once during he survey period and students who left college without return prior (dropout) by June 2001. Students who left college without return during the study’s timeframe, but stopped-out at least once prior to leaving were counted as dropouts in analysis. The independent variables were categorized as followed in Table 2: Table 2 Categories of variables Category Variable Desorption Dependent Variable STOP_STAY2 Filter Filter 8 Weight NORM_LWT2 Independent Variables Background GNDERCOD RACECODE DISADVAN Students who returned to college after a period of unenrollment (at least four months) & who left college without return prior to June 2001. Students who began at a 2-year college; stopped-out at least once; left college without return prior to June 2001. Normalized longitudinal weight for respondents to NPSAS: 96, BPS: 96/98, and BPS: 96/2001 Gender Recode Race/ethnic & permanent resident recode Socioeconomic diversity index 26 ‘Lv‘1 {-1 031; (fl: ~ 1 "u‘, I.‘l Table 2 (con’d) Risk Factors (Nontraditional) RISKINDX Risk Index 95-96 Recode ENIPTTZB Intensity pattern through 2001 ENDELAY Delay-delayed enrollment after HS HSDIPLOZ Received diploma or passed GED Recode DEPENDEN Dependency Status Recode MARRYREC Marital Status Recode KIDSRCODE Number of dependents 95-96 Recode SINGPAR Single Parent Recode 95-96 Risk Factors (Other) REMEDIAL Remedial Courses Taken Recode ENROLRSN Primary Reason for Enrolling 95-96 GPARECOD GPA Recode AHLOANYI Received student loan 1995-96 Factors Examined (Not in Model) ACADINT Academic Integration Index CMSOCREC Have social contact w/ Faculty 95-9 recode CMTALKRC Talk w/ Faculty Outside Class 95-96 recode CMSTUGPR Attend study grps out of class 95-96 recode CMMEETRC Meet w/ Advisor 95-96 Recode A majority of variables were recoded and dummy variables were constructed into a binary format for use with logistic regression. For further explanation of the construction of the variables see Appendix A (Table 13) and B. Parental education was used to construct SES Index, therefore it was not included separately to avoid multicollinearity. In addition, variables associated with social integration was not included because (a) factors associated with academic integration has been shown to be more relevant and predictive of persistence than social involvement on two-year 27 ~ I Li‘- I H mm. 5 campuses (Borglum & Kubala, 2000; Napoli & Wortman, 1996) and (b) the variables that were used to construct the social integration index (i.e., participation in fine arts, clubs, varsity sports, and intramural sports and spending time with friends from school) are not as relevant to two-year institutions. , Analysis. Frequencies and cross-tabulations were used to identify significant differences in stopout/dropout in the 95-96 cohort of students in regard to the independent variables, while direct logistic regression was used to determine the relationship of the latter variables to student return to college. Logistic regression was an appropriate statistical tool because of the dichotomous nature of stOpout (i.e., return to college or dropout), the dependent variable (Hosmer & Lemeshow, 2000). With a dichotomous variable the assigned values are generally 0 and 1, therefore the predicted values for regression take the form of mean proportions or probabilities. This is problematic, as the maximum and minimum values are limited to l and 0, yet the values in linear regression can range above a 1 or below a 0. Therefore, linear regression methods are inappropriate and have little predictive use with designs that have a dichotomous outcome variable. Logistic regression models linearize the dependent variable’s non-linear relationship with independent variables. (Pampel, 2000). I ran frequencies for all variables to evaluate the missing data within each variable, verify variables were constructed appropriately, and describe the sample (See Appendix C, Table 14). As stated previously, chi—square tests for independence were run between the outcome and independent variables to test for significant differences (See Appendix C, Table 15). This information helped inform which variables should be included in the logistic regression model. 28 This study reports the results for a number of tests that indicate how well the model. performs. I used SPSS statistical software, version 17, to analyze the data. The Omnibus Tests of Model Coefficients is a chi-square statistic as a goodness-of-fit test for how well the model performs in comparison to the model without any of the predictors. Another test for goodness-of-fit is the Hosmer and Lemeshow Test. The Cox and Snell R-Square and Nagelkerke R-Square are psudo-R square indications of the amount of pseudo-variation in the dependent variable explained by the model. The Classification Table indicates how well the model is able to predict the correct category. Finally, the significance of the each independent variable, odds ratios, and coefficient are also reported. (Pallant, 2007) All independent variables (with the exception of indexes) were initially entered into the model. I removed variables based on their level of significance and effect on the goodness-of-fit and significance of the overall model, until the significant and necessary variables remained. Gender, race, and disadvantage were held constant. Odds ratios were repOrted in the output and used to compute probabilities (Table 3). 29 Table 3 Logistic Regression Predicting Likelihood of Returning to College After Nonenrollment Variables in the Equation 95% C.I.for EXP(B) B SE. Wald df Sig. Exp(B) Lower Upper Step 1‘ GNDERCOD .054 .178 .092 1 .76 1.055 .745 1.495 BLACKDUM —.558 .287 3.773 1 .05* .572 .326 1.005 HISPADUM -.354 .286 1.537 1 .21 .702 .401 1.228 ASIANDUM -.117 .396 .087 1 .76 .889 .409 1.934 MDISADVAN -.501 .197 6.454 1 .01** .606 .412 .892 HDISADVAN -.295 .293 1.013 1 .31 .744 .419 1.322 DEPENDEN .601 .272 4.880 1 .02* 1.824 1.070 3.108 HSDIPL02 -.364 .274 1.764 1 .18 .695 .406 1.189 MARRYDUM -.676 .312 4.674 1 .03* .509 .276 .939 DIVORDUM -1.469 .506 8.423 1 .004** .230 .085 .621 PARTDUM -.453 .271 2.800 1 .09 .636 .374 1.081 MIXDUM .565 .216 6.866 1 .009** 1.759 1.153 2.683 JOBDUMMY -.690 .281 6.046 1 .01** .502 .290 .869 AACERTDM -.312 .275 1.292 1 .25 .732 .427 1.254 TRNSFRDM -.012 .248 .002 1 .96 .988 .607 1.608 GPARECOD .174 .088 3.917 1 .04* 1.190 1.002 1.413 Constant -.537 .350 2.347 1 .12 .585 Note: Variable(s) entered on step 1: GNDERCOD, BLACKDUM, HISPADUM, ASIANDUM, MDISADVAN, HDISADVAN, DEPENDEN, HSDIPL02, MARRYDUM, DIVORDUM, PARTDUM, MIXDUM, JOBDUMMY, AACERTDM, TRNSFRDM, GPARECOD. *p<.05 **p<.01 Limitations. There were a number of limitations to this study. First, the study is limited to and specifically focused on variables related to non-traditional student status. Of the seven characteristics, two (having dependents and single parenthood) were 30 eventually removed because they did not significantly contribute, and I preferred not to impute them due to the potential reduction to variation. Since there were no significant relationships between missing data within the independent variables and the dependent variable, I chose to exclude those variables. One variable was never included (work intensity) due to multicollinearity. Second, in an effort to remain focused on the relationship between student risk factors more common among community college students and stopout, I did not include variables often associated with more traditional students--academic integration, social integration, and loans/financial aid (Beil, Reisen, Zea, & Caplan, 1999; Braxton, Sullivan, & Johnson, 1997; Borglum & Kubula, 2000; Cabrera, Nora, & Castaneda, 1993; Dowd, 2004; Napoli & Wortman, 1996, 1998). While this exclusion allows for a better understanding of a specific set of factors, it also excludes other factors that could explain the outcome more deeply. Third, I did not calculate the intracluster correlation coefficient (ICC) to measure the internal homogeneity of clusters within the design, a measure that could provide an indication of the degree of bias in estimating variances in data collected through cluster samples. (Thomas & Heck, 2001). Also, I did not compute the delta-p for the continuous variables in the model. This measure would indicate the change in probability per one- unit change in the independent variable (Peng et al., 2002). Additionally, while using a more conservative alpha-value and relative weight to alter the sample size are acceptable adjustments for calculating more accurate standard errors, using specialized software is preferable (Thomas & Heck, 2001). 31 IL ind AU 6 In M- p . GPN if? Finally, due to financial and resource constraints it was not possible for BPS researchers to survey the entire population of students and institutions. It therefore was necessary to use a sampling approach for estimation of trends and patterns across the data. To ensure appropriate representation of the population researchers oversampled selected subgroups of students and institutions. This results in a distortion of the overall representation that required adjustment through the application of a weight. In addition, the clustering design approach contributes to nonindependence of observations and nonhomogeneity, violating the assumptions of simple random sampling. A more conservative alpha level and normalized weight were applied to compensate. Part Two: Qualitative Interviews The qualitative portion of the study was approached as a grounded theory that focuses on the stopout experiences of community college students in relation to their decision to return (Creswell, 1998; Miles & Huberrnan, 1994). Grounded theory was ideal for exploring a similar experience among a number of individuals and developing an analytical schema of that phenomenon (Creswell, 1998). Although it was primarily a grounded theory approach there were also elements of phenomenology as the inquiry intended to collect in depth information about the meaning of stopout experience, why students return to college, and their decision making process. This section will provide information about the qualitative research design, participants and sampling, research site, and analysis. Pilot interviews. After the dissertation committee approved my protocol, I interviewed two community-college students who previously stopped-out, a man and a woman. I used the pilot to further construct the interview protocol. The interviews lasted 32 30 to 60 minutes each. After the interviews I asked students to discuss their experience during the interview and provide suggestions for additional questions. In general, the protocol elicited the information desired, but I made adjustments to the wording of several questions to make it clearer to the participants. Recruitment and sampling. 1 selected the research site through a combination of purposeful sampling and existing contacts. I reviewed this state’s community colleges for the institution’s proportion or percentage of certain characteristics in relation to state and national average. I concentrated on student population race and ethnicity, gender, attendance intensity (e.g., full- or part-time student), and student goals (e. g., transfer, Associate’s degree, certificate). I sought schools that roughly matched the state’s breakdown in race and ethnicity, parity in gender or slight over representation in women, 50% or more of part-time students, and a significant number of students pursuing transfer, an Associate’s degree, or a certificate. This information was found through institutional websites and reports. In conjunction with the dissertation committee, I generated a list of institutions with departmental contacts or connections. Two schools were identified that met the selection criteria and had contacts at them who were former students in connections in my academic department. At the first institution, my contact put me in touch with the Dean of Student and Academic Support and Strategic Enrollment Management. I met with the Dean and got permission to pursue the study on her campus. We agreed to recruit in two ways: 1) an email was sent out to academic advisors asking them to nominate students who fit the recruitment criteria and 2) through the identification of students who left and returned to this institution (gone at least four months) in institutional records. The Dean sent the 33 email to staff and I received one response. I conducted that interview. In subsequent efforts to contact the Dean and reconsider recruitment methods, I was unable to contact her. Therefore, I made the decision to drop this institution as a research site and use the interview as a pilot interview. Through a contact at the second institution, I spoke with and received permission to conduct the study from the Vice President for Academic Affairs. I worked with an institutional researcher from the college to recruit the students. This person also facilitated the space for interviewing. Students were recruited using three methods: 1) Fliers advertising the study were posted across the campus by institutional staff (see Appendix D) I recruited in two classrooms in the electrician program; and 3) an email was sent to the entire student body by the college on my behalf (see Appendix E). In all three recruitment methods I conveyed to the students the criteria for the study and to either contact the researcher through email or provide contact information on a sign—up list. The criteria for participation was that students must have: ' Been enrolled in college previously (any college—4-year or 2-year) and taken a leave of absence that lasted at least one semester (not including summer) to six years; ° They must be currently enrolled at ACC (Archer Community College); 0 Have academic goals of either transfer to a 4-year institution, an associate’s degree, or a certificate. 34 j}! ‘1 As a result of the study’s criteria, all of the students in this part of the study returned to ACC to continue their education, but may have begun their education at another institution (technical, two-year, or four-year). A majority of the participants were recruited through the email. Approximately 115 students expressed interest and 48 interviewed (see Appendices D and E for recruitment materials). Location. This study was conducted at an urban, mid-westem community college (using pseudonym of Archer Community College) that has 10,456 students. Note that I changed the name for all references to the site school and city, including citations in the Reference section of this paper. Approximately 60% are female, 25% are students of color, and 64% are part-time students. The average age of students is approximately 28- 29. I selected this college because of its relatively large population among community colleges in the state and demographic reflection of the state in hopes of recruiting a racially diverse participant group. Demographically the city is approximately 53% African American, 41% White, 3% Hispanic/Latino, 1.1% of more than one race, and all other racial and ethnic groups are at less than 1%. This demographic division of race is roughly reflected in the sample of study participants. It is the fourth largest city in the state, yet the population has continued to decline in the last decade with a reported population of 131,000 in 1998 and 114,662 in 2007. The median income for the city is $28,010 with 26.4% of the population living below the poverty line; 37.4% of those under 18 below the poverty line (City of Archer, 2008). Since the 19203 the city’s economy has been dominated by the auto industry as home to a major American auto company and the United Auto Workers 35 twink F1134 2. a union. As the American auto industry experienced decline and deindustrialization there was (and continues to be) a loss of industry and jobs in the area and, as a result, an increase of significant economic challenges for the city’s residents (Dandaneau, 1996). The study’s findings demonstrate the influence that the local economy had on many students’ decision to return to college. Participants. I interviewed 48 students about their decision and experience returning to a community college campus. Approximately a quarter of the participants are students of color (primarily Afiican American). The sample ranged from 20 to 50 years old, with a mean age of 32.5 and 52% of participants falling between 23 and 33 years old. Fifty-two percent of the participants were women and the rest were men, with one female-to-male individual. All the participants are pursuing transfer, a certificate, or an associate’s degree and are currently enrolled. A majority of students (66%) have only stopped out once, but that stopout ranged from one full semester (not including a summer semester) to 15 years. Finally, while stopping-out places these students in a higher risk category for not completing their educational goals, 86% of the students also had at least one additional risk factor when they first left college for not completing their educational goals associated with non-traditional college students (Feldman, I993; Hoachlander et al., 2003). Research design. After IRB approval, I conducted 30 to 60 minute interviews with 48 participants from an urban, mid-western community college. Prior to the interview, I reviewed and received informed consent (Appendix F), introduced myself, reviewed the purpose of the study, and asked students to select their own pseudonyms. Pseudonyms were used to protect the anonymity of each student. 36 .VIN .L a. Students described why they left school and what their decision to return to college looked like. I drew upon possible selves theory to develop the protocol and in analysis to further understand how a student’s conceptions of his or her possible academic and career selves influenced the decision to return. Understanding how participants conceive of their possible academic and career selves prior to departure and when returning provided insight into changes in the perception of possible selves and strategies for the achievement of possible selves that are important in the decision to reenroll. Previous studies found that asking students to reflect on important events and decisions provides a context to more clearly describe their decisio— making processes and what was important within them (Baxter Magolda, 2001). To facilitate this inquiry, the interview questions guided students through their educational history; their goals, hopes, fears, and how they changed over time, as well as their decision making process to leave and return to college. A sample of questions asked include: Tell me about your experience in school before college; What did you see yourself like as a student? Did you have plans for how you were going to achieve those goals? What did you hope to be like as a student? What did you see happening when you left? Did your academic goals change while you were gone? Is how you see yourself as a student or in your future career different from before? As noted, the protocol served as a semi-structured guide for the interview, therefore I asked many questions and probed beyond the scope of the protocol when appropriate and necessary to gather more detail and information. In addition, all students completed a demographic questionnaire. The questionnaire asked 37 questions about age, race/ethnicity, enrollment and employment intensity, and educational history. The full protocol and questionnaire are included in Appendix G. At the conclusion of the interview, the students were given an opportunity to ask questions. Many students asked about my own interest in the topic, as a result I shared how I came to be interested in this topic. I restricted this discussion to the end of the interview in order to avoid unnecessarily influencing the participants’ responses. At the end of the interview I also paid the students $15 for their time and effort. Seven students expressed interest in the final results of the study; I asked these students if they were willing to provide feedback on the findings as they emerged through email. A majority of these students later provided feedback in an effort to establish trustworthiness for the study. Analysis. I used constant comparative analysis to analyze the data (Strauss & Corbin, 1998). I read the transcripts for codes and themes focusing on the reasons stridents’ returned to college, strategies and plans for returning, hopes and fears for academic and career selves, and the intersections among these areas. A first pass at coding revealed themes in accordance with those previously listed. Additional analysis exposed more detailed patterns among student responses that identified life events that influenced the decisions to leave and return to college, experiences that affected dispositions toward college and learning, and two different paths through college shaped by educational attitudes and possible selves. 38 Table 4 Themes and Subthemes for Qualitative Part of Study Theme Educational Path High School Experience Reasons for Leaving Reasons for Return Subthemes K-12 Experience High School Entry into College Experience in College First Time Leaving College Return to College Most Recent Experience in College “Burnt out” and Lazy Average Student Liked School/Good Student Not/Encouraged by School/Family College was Expected/What you did Did Not Set: Purpose Did Not Fit in Did Not See Purpose “Burnt Out”/Taking a Break Pregnancy Illness/Death/Family Emergency Wanted to “Experience the World” Job Opportunity/Make Money Not “College Material” Drugs/Alcohol Jail No Direction Personal Goal Job Stability Job Advancement Paid for by Government Program Example to Children College Now has Clear Purpose 39 Table 4 (Con’d) Learner Disposition Possible Selves Factors Change in Approach to Academics Now Care Learning to be Better Professional Change in Study Habits Take Advantage of College Resources Utilize Faculty and Students Learning Tied to Other Goals College/Non-College Possible Selves CPS=Plans for College CPS=Lifelong College Goal NPS=Poor HS Academic Experience NPS=Lack of direction Change in PS=Change in Roles Employment/Job F amily/Children Marital Status External Responsibilities Laid Off/New Job Finances Motivation Academic Self-Efficacy Belief in Purpose of College Personal Goals In an effort to establish trustworthiness, emerging themes and patterns were shared with seven participants for feedback (Lincoln & Guba, 1985). During the interviews many students expressed interest in learning more about the study’s findings. I took the opportunity to ask these students if they would be willing to review findings and provide feedback. Due to the non-traditional characteristics of the majority of students in the study (e.g., single parent, above the age of 25, married, part-time student, full-time employee, etc.), finding time to meet is a challenge, therefore I emailed students a short explanation of findings and themes and they emailed their feedback back. I used the feedback to evaluate the patterns and made adjustments to more accurately reflect 40 participants’ experiences. Most respondents affirmed the patterns I identified and suggested only minor additions or adjustments. In addition to evaluating the trustworthiness of the findings with the participants, I also compared the findings to the quantitative part’s outcomes. The quantitative study was an analysis of the relationship between characteristics and factors recognized in the literature as related to departure from college without completion of a degree or certificate and whether or not students returned to college after a period of unenrollment. While the survey does not include items that allow me to evaluate the outcome’s relationship to possible selves, I was able to compare the qualitative themes regarding the role of risk-factors in the decision to leave/return and reasons for initially entering college to the findings from the quantitative study. Mixed methods article While not a separate part of the study, the third article in this dissertation required additional analysis of the qualitative data. The quantitative findings were reiterated, but the qualitative data were examined more closely for themes and patterns specific to the independent variables used in quantitative part and other reasons students gave for leaving college. I used the same qualitative data for the additional analysis. Themes identified and discussed are included in Table 5. 41 Table 5 Additional Themes and Subthemes for Qualitative Part of Study Theme Subthemes Reasons for Leaving Did Not See Purpose “Burnt Out”/Taking a Break Pregnancy Illness/Death/Farnily Emergency Wanted to “Experience the World” Job Opportunity/Make Money Not “College Material” Drugs/Alcohol Jail No Direction Family/Children Marriage/Divorce/Widowed Financial need Parental Encouragement/Example Reasons for Enrolling Disadvantage Limitations Recognition of this study’s limitations can assist understanding of how to apply these findings and how further research can extend this work. First, while the racial demography of this study is roughly similar to that in the state, students of color are a relatively small percentage of the participant group and primarily African Americans. Second, the state and particularly the city in which the college is located was dominated by the auto industry. As deindustrialization began to affect auto manufacturers over the past twenty years, the economy and employment in this area have been struggling and continue to be affected by issues in the national economy. While this study will likely be applicable to institutions in areas with high unemployment and significant economic hardship, this level of dependence on manufacturing industries and the subsequent suffering is not necessarily reflected in other areas and populations across the country. 42 [531 up“ 93" l.)' . t... “uh! ‘ . ‘H‘N'sr .3, Rise: 1? f '1 ’1 Third, this is a retrospective account of student experiences. Therefore, the benefits of hindsight and casualties of memory do not always reflect how students may have reported the same experience when it happened. Finally, this study is focused on students from one community college and frndings may not reflect student experiences with stopout at other community colleges or types of institutions; generalizability is limited. Summary In this chapter I supplemented the methodological information provided in Articles One, Two, and Three. After a comprehensive description of the overall mixed methods design of the study, I discussed the details of the quantitative and qualitative parts. 43 CHAPTER 3: ARTICLE l—QUANTITATIVE STUDY Introduction In President Obarna’s first address to a joint session of congress, he challenged the nation to increase focus on education and raise the proportion of college graduates to the highest in the world by 2020 (2009). The Lumina Foundation responded to this " challenge with a response, stating the critical need for increased higher edircational attainment as a priority for US. economic development and stability in the short- and long-term. Lumina presented its own educational goal for the US. to improve the percentage of Americans with a degree or credential from 39% to 60% by 2025. Two of the suggestions given in support of this goal include aligning postsecondary education with the workforce development system and expanding adult learning opportunities. The suggestions call for strengthening community colleges as institutions central to these recommendations. In addition, the Lumina foundation suggests drawing on adult students who never attended college or finished their educational goals and dislocated workers for workforce retraining and skill development (Lumina Foundation, 2009). Reaching these goals requires increasing the number of students who attend and finish college directly out of high school, as well as drawing individuals back to college who never finished or attended school previously. This is not only important for the benefit of the student, but the nation’s economic future. As the job market experiences a shrinking manufacturing field and increasingly requires employees to have technical and knowledge-based skill, individuals are likely to return to postsecondary education for the development of job skills and retraining (Bragg, 2001). Community colleges already provide a significant proportion of skill trade and job 44 training education, which suggests that many returning adult learners will turn to two- year colleges for postsecondary education (Bragg, 2001). The increasing cost of college, tightening economic realities for families, and open access character of community colleges make this type of institution a more likely option for adults who are returning to college and therefore an authentic setting to study returning students. As returning adults, these students are likely to have personal and work lives that can pose a challenge to their educational goals (Horn, 1996). During their time out of school it is reasonable to believe that they established work and home lives that require time and effort to maintain. Community college students’ non-academic life demands often translate into a challenge to balance work, family, and academics resulting in student needs that are different from students at four-year institutions. Many of these characteristics (e.g., full-time work, part-time enrollment, dependents, financial independence, delayed college enrollment, GED) are associated with an increased risk of departure (Horn, 1996). Community colleges already serve a greater proportion of students who are at higher risk for departure (Horn, 1996) and as a result are poised to serve the increasing number of students that US. leaders hope to attract back to postsecondary education. Therefore, it is critical for community college faculty, staff, and administrators to understand the educational paths of non-traditional and returning adult students in order to best serve and support them. This recent call to focus on bringing students into or back to postsecondary education for workforce retraining and skill development means that in addition to supporting college students who are more likely to succeed, practitioners and researchers 45 must also understand the educational paths of students who leave and return to college. The findings reported in this article help to explain how and what factors relate to risk of departure. These findings contribute to a better understanding of students’ decisions to return to postsecondary education and add to foundational knowledge for creating educational best practices in retaining students or assisting them in their return. The phenomenon of students returning to college after a period of nonenrollment is termed stopout In most definitions, the period of nonenrollment is at least four months (Horn, 1998; Bonham & Luckie, 1993). In this article, first, relevant literature is reviewed and followed by a discussion of the study’s methods, design and analysis. The findings section reports an overall evaluation of the model and includes a discussion of the individual significant independent variables. Finally, implications and suggestions for research and practice are presented. Background Community college students A focus on attracting individuals returning to postsecondary education, particularly the two-year sector, requires a foundational understanding of students at community colleges. In 2006-2007 approximately 65% of students attended a four-year college and 35% of students in higher education attended a two-year institution (Provasnik & Planty, 2008). Examination of the characteristics of students attending two- year colleges paints a picture of a diverse population. Forty percent of community college students are dependents (i.e., under 24 and financially dependent), while the remaining 60% are financially independent--26% are older than 24, 20% are married with children, 46 and 15% are single parents. In general the median age of a community college student (24) is higher than those at four-year colleges (21). Community colleges also have higher percentages of low—income students (26% versus 20% at four-year colleges) and Black (15% versus 10%) and Hispanic students (14% versus 9%) (Horn & Nevill, 2006; Provasnik & Planty, 2008). Community college students’ academic goals also vary; 36% report intentions to transfer, 43% seek an associate’s degree, 17% are enrolled for a certificate, and 42% report returning for job skills (not mutually exclusive categories) (Provasnik & Planty, 2008). Furthermore, 54% of students who enter a community college have an increased risk of departure (Hoachlander, Sikora, & Horn, 2003) and lack academic preparation in reading and math (Borglum & Kubala, 2000; Nora, Attinasi, & Matonak, 1990; Summers, 2003). Persistence and stopout In addition to having higher levels of diversity across student background characteristics, community colleges also have higher percentages of departure (45%) than four-year colleges (17%) (Horn, 1998; Provasnik & Planty, 2008). Approximately 30% of students in all of higher education leave after the first year--63% are from community colleges, 22% are from four-year institutions, and 15% are from other types of schools (e.g., private, for-profit; public less-than-two year; or private, not-for-profit, less than four-year institutions). Specifically, among students who attend four-year institutions 16% depart, while 42% of students enrolled at two-year schools will leave. Of those who leave, over half (57%) will return to higher education within five years. (Horn, 1998). Yet, 50% of those who depart from the two-year sector return, while 64% of those who leave a four-year institution return within five years. This gap in returns indicates that 47 students who begin at a two-year institution and leave after their first year of school are less likely to return to college than students who begin at four-year institutions, but represent a greater number of students overall. In other words, out of 100 students who . leave during their first year 85 of them are from two- and four-year institutions. Of these 85 students, 46 will return within five years: 32 of those who return originate from a community college versus the 14 who begin at a four-year school. Furthermore, of the students who leave during their first-year and return within five years, a majority will return to community colleges with 79% of students who start at a two-year and 24% of students who start at a four-year institution returning to community colleges (this includes students who return to the same institution at where they began and students who transfer from another two- or four-year institution) (Horn, 1998). For example, out of 100 stopout students who begin at a two-year institution, 79 of them will return to a community college and the remaining 21 will attend four-year or other types of institutions. Additionally, out of 100 stopout students who began at a four- year institution, 24 of them will return to a community college and the remaining 76 will attend four-year or other types of institutions. Recent calls to increase graduation rates and encourage individuals to return to postsecondary education highlight the issue of stopout in the study of attrition and retention. Understanding factors associated with which students stopout or dropout can contribute to developing strategies for how to prevent students from leaving and creating pathways for student return to college. 48 Risk factors related to student departure The community college student population is more diverse compared to the student population at four-year institutions—the students are older, more racially and ethnically diverse, and less academically prepared. Many of the characteristics that contribute to the diversity at two-year institutions were also associated with traits linked to a higher risk of college departure. In a report profiling undergraduates, Horn and Premo (1993) identified seven characteristics commonly associated with a higher risk of non-persistence and nontraditional students. The traits were representative of three groups: 1) Nontraditional enrollment choices (e. g., delayed enrollment or attending part-time), 2) financial and family status (e. g., having dependents, working full-time, financial independence, or being a single parent), and 3) high school graduation status (e.g., receiving a diploma or GED.) (Horn, 1996; Horn & Premo, 1993). Each of these factors were linked to a potential increase in risk of attrition, therefore the number of characteristics students had was also an indication of risk levels (i.e., minimally, moderately, and highly at risk). Horn and Premo (1993) concurrently labels these risk factors as characteristics associated with non-traditional students. This reveals a bias toward a four-year inStitution perspective. Students with these characteristics are more likely to attend two-year institutions (Horn, 1996), therefore while community college students are more likely to be at higher risk of departure compared to students at four-year institutions they are very traditional for a two—year institutional context. The current study draws on the risk factors identified by Horn & Premo for independent variables, but since the sample is focused 49 solely on students who began at community colleges I choose to frame these factors only as indicators of risk versus levels of traditionalness. Additional research point to other student characteristics associated with being at higher risk for departure and lack of persistence toward educational goals. Previous studies report that while traditional-aged students out-paced nontraditional -aged students in enrollment growth between 1995 and 2006, this trend will likely reverse between 2006 and 2017 with nontraditional-aged students enrollment growing nearly twice as fast as enrollment for traditional-aged students (Snyder, Dillow, & Hoffman, 2009). In addition, students with more risk factors are concentrated at two-year institutions, while students with no risk factors are concentrated at four-year institutions (Horn, 1996). Also, higher risk students are less likely to earn a degree within five years of beginning college and more likely to leave without returning than traditional students. Thirty-one percent of higher risk students obtain a bachelor’s degree within five years compared to 54% of lower risk students, also higher risk students are twice as likely to leave in their first year (3 8%) than their lower risk counterparts (16%) (Horn, 1996). There is ample evidence that higher risk student status is associated with negative outcomes in higher education, such as early departure and lower rates of degree attainment (Horn, 1996; Horn & Premo, 1993; Hoachlander et al., 2003). Findings also indicate that the number of older and higher risk students on campus is growing and is concentrated at two-year institutions. With the appeal by President Obarna and the Lumina Foundation to increase the number of students receiving diplomas and returning to education, it is likely that the higher risk student population will continue to grow. Assuming that a significant number of students who return to college also have higher 50 risk student traits, gaining a better understanding of the association between risk factors for attrition, who returns to college after nonenrollment, and why they return is important. Following is a detailed discussion of the study’s design and methods. Method Research Questions This quantitative analysis is part of a larger mixed methods study focused on the overarching research question: Why do community college students return after a period of non-enrollment? The study described in this manuscript answered particular subparts of the overarching question. The research questions addressed by this analysis were: (a) Which factors are related to who returns after a period of non-enrollment? And (b) what is the influence of these factors on stopout? Data and Sample The Beginning Postsecondary Students Survey (BPS) is a longitudinal extension of the National Postsecondary Student Aid Study (NPSAS) conducted by the US. Department of Education National Center for Education Statistics (NCES). This nationally representative survey includes a cohort of students who began their postsecondary education during the first year of the NPSAS study. An advantage of this survey is that it collected information on any student whose first-year of study coincided with the NPSAS base year; this includes non-traditional students (e.g., delayed enrollment, single parent, non-dependents, received a GED) in addition to traditionally— aged college students (17 -19). Also, the survey was designed to collect detailed enrollment information, including information about enrollment and unenrollment spells, changing of institutions, and exiting postsecondary education without return during the 51 length of the survey (Wine et al., 2002). Therefore, BPS is appropriate for statistical analysis that includes both non-traditional students and stopout patterns. This study analyzed BPS: 96/01, which began with the 1995-1996 base year of NPSAS. Additional surveys were administered in 1998 and 2001. The approximate number of BPS: 96/01 respondents was 12,100. I excluded students for this study who . did not begin at a 2-year public institution or stop- or dropout during the six years the study to ensure that the data represented community college students who made a decision to remain out or return to postsecondary education. BPS utilized a stratified multistate cluster complex sampling design. The strata represented different sectors of higher education (institution-level based) and clusters represented geographic regions. The public 2-year sector was included in this sample. While the NSPSAS and BPS sampling designs ensured that the sub-sample was nationally representative, adjustments for oversampling bias and nonindependence of observations are discussed in a later section. I selected students for the sample if they began their postsecondary education at a public 2-year institution and either stopped-out or stayed-out during the length of the study. Stopout was defined for BPS as a break in enrollment of five or more consecutive - months. It begins on the first month of the break. A stopout is also the months between two enrollment spells (Wine et al., 2002). I considered students as stopping out if they indicated that they stopped-out one or more times (PRSTNU2B). A majority of the students who returned to college returned to two-year institutions, but the students may have also returned to other types of institutions (e.g., technical or four-year). I defined dropout as students who left postsecondary education without return and without 52 attainment of a degree between fall of 1995 and June 2001. While students who left without return during this period of time may return at a later date, for the purposes of this study I included them as dropouts. Also, I included} students who stopped-out prior to leaving without return as dropouts due to their lack of return. The decision to limit the , sample to students who began at a public 2-year institution sustained the focus on students who would be classified as a community college student and avoid the confounding element of students who may have attended a public 2-year institution at some point, but began at a 4—year institution. The resulting sample size for the study was 779 observations. Thirty-five percent (272) of students in the sample stopped-out, while 65% (507) stayed-out. In Table 1 are frequencies for descriptive characteristics of the sample. The “total” in parenthesis refers to the percentage of that category in the total sample. For example, male stopouts account for 29.7% of all the males (stopouts and dropouts) in the study and male dropouts represent 17.7% of all the males in the study. Table 6 Characteristics of the Sample gem Stopout Male Female 45.7% (231) 54.3%(275) (29.7% of total) (35.3% of total) Dropout Male Female 50.5% (138) 49.5%(135) (17.7% of total) (17.3% of total) Total Male Female 47.4% (369) 52.6% (410) 53 Table 6 (con’d) Stopout Dropout Total Stopout Dropout Total Stopout Dropout total) Total Analysis White 76.1% (207) (26.6% of total) White 68% (345) (44.3% of total) White 70.9%(552) ’ Not Disadvantaged 48.7%(1 3 1) (17% of total) Not Disadvantaged 34.2%(172) (22.3% of total) Not Disadvantaged 39.2% (303) Minimal Risk (0-1) 47.7%(129) (17% of total) Minimal Risk 35.2% (129) (22.6% of total) Minimal Risk 39.6% (301) Rage ' Black Hispanic Asian/Indian/Other 8.1%(22) 10.3%(28) 5.5%(15) (2.8% of total) (3.6% of total) (1.9% of total) Black Hispanic Asian/Indian/Other 15%(76) 12.4%(63) 4.5%(23) (9.8% of total) (8.1% of total) (3.0% of total) Black Hispanic Asian/Indian/Other 12.6%(98) 11.7%(91) 4.9%(38) Socioeconomic Index Minimally Moderately/Highly Disadvantaged Disadvantaged 38.3%(103) 13%(35) (13.3% of total) (4.5% of total) Minimally Moderately/Highly Disadvantaged Disadvantaged 53.1%(267) 12.7%(64) (34.6% of total) (8.3% of total) Minimally Moderately/Highly Disadvantaged Disadvantaged 47.9%(270) 12.8%(99) Risk Index Moderate Risk (2-3) High Risk (4-7) 30.1% (82) 22.4%(61) (10.8% of total) (8.0% of total) Moderate Risk High Risk 34.2% (167) 30.7%(150) (21.9% of total) (19.7% of Moderate Risk High Risk 32.7%(249) 27.7%(21 1) The complex sampling design used for BPS 96/01 (and most large—scale surveys) presented challenges that needed to be addressed for analysis. First, oversampling of students and institutions with particular characteristics (e. g., racial and ethnic student attributes, characteristics associated with non-traditional students, HBCU institutions) is common within large-scale surveys to ensure sufficient numbers are represented for 54 analysis (including BPS). While necessary for analysis of these populations, oversampling usually results in a distortion of the representation in the overall population 9 resulting in additional weight for these students and institutions that will bias results (Thomas & Heck, 2001). The weights constructed for BPS 96/01 address both subsampling and nonresponse bias within the survey. The sampling weight for longitudinal analysis of BPS 96/01 (B01LWT1) was selected to adjust for bias. The weighted sample resulted in over-estimated significance levels due to the larger population size, therefore a normalized weight was constructed to preserve the accuracy of the weighted estimates by reflecting the sample size (Thomas & Heck, 2001). To avoid removal of cases with normalized weight values of zero in analysis, a constant of .001 was added to the weights. This allowed for the cases to remain in the sample for analysis with a negligible affect (de Vaus, 2004). Second, the use of clustering in the latter stages of large-scale surveys contributes to the nonindependence of Observations and nonhomogeneity, violating the assumptions of simple random sampling. The typical statistical software treats data as if constructed lusing simple random sampling and not adjusting for effects from multistage clustering, resulting in variance estimation and standard error bias (Diemer, 2008; Thomas & Heck, 2001). In recognition of potential estimate bias as a result of a simple random sampling assumption, a more conservative alpha level was used to evaluate the model and predictor variables (p<01) (Thomas & Heck, 2001). While not reported, I initially conducted chi-square tests for independence to analyze the difference between students who stopped-out and students who stayed-out on the independent variable. Chi-square tests for independence were also used to evaluate 55 multicollinearity among predictor variables (see Appendix C) (de Vaus, 2004). No significant relationships were found. Finally, I created a dummy variable for missing and nonmissing observations for each independent variable and ran chi-square tests of independence with the dependent variable to evaluate for potential relationships or patterns (de Vaus, 2004). Again, none of these tests were significant, therefore additional missing data analysis was not conducted. Analyses were conducted using SPSS statistical software, version 17. Direct logistic regression analysis was conducted to observe the effects of factors associated with hi gh-risk for departure on student return to postsecondary education after a period of nonenrollment Numerous statistics are reported as an evaluation of the final model. SPSS uses the Omnibus Tests of Model Coefficients chi-square statistic to determine if the model provides an overall fit to data by providing improvement over the intercept-only model. Additionally, an examination of changes in the percentages of correct classifications between the null and final models is used as a validation of predicted probabilities. Finally, goodness-'of-fit of the model is presented with two different statistics. The Hosmer and Lemeshow statistic reports a chi-square test for goodness-of-fit that is more conservative and sensitive to the ways in which predicted probabilities are grouped (Hosmer & Lemeshow, 1989; Peng et al., 2008). When interpreting the significance it is important to note that alpha levels above .05 are considered an indicator of better fit Also, since a traditional R square cannot be determined using logistic regression, the Cox and Snell R Square and Nagelkerke R Square are pseudo-R square statistics used to estimate the percentage of the amount of variation in the dependent variable explained by the model. 56 Dcff’n‘jc {Biff-ten: mum id ' I 4 *1“ (”A“)! {rut Th l} .- iltflalure f Pffmll. 19 Ho: higher risk itgle parel While enrol also used in lirefore th. the variab. depart UIC. Dependent variable The dependent variable is a constructed binary variable (stop__stay2) that represents students who began at a public two-year institution and either left without return (dropout) or left with return (stopout). Stopout is the outcome of interest. Table 7 Dependent Variable Stop_Stay2 Students who began at a two-year institution that: (a) left college and returned during the course of the study or b) left without return to college prior to June 2001. Independent variables The primary variables of interest within this study reflect characteristics that literature has shown to be related to a hi gher-risk for departure (Horn, 1996; Horn & Premo, 1993). Horn (1998) identified the following seven characteristics as traits related to a higher risk of student departure: delayed college enrollment, received a GED/certificate, single parenthood, financial independence, responsible for dependents, wOrked fulltime while enrolled, and enrollment in school less than fulltime. These characteristics were also used in BPS 96/01 to construct the risk index for departure (Wine et al., 2002), therefore this index is a measure of both the level of student risk of attrition. I included these variables individually in this study as potential indicators of increased risk for departure. A high proportion of students at community colleges are at greater risk of departure and inversely a high number of higher risk students begin their postsecondary 57 education at a two-year institution (Horn, 1998; Stratton, O’Toole, & Wetzel, 2004). Therefore, it was important to include variables in this study that were related to increased risk levels for non-persistence. Four out of the seven variables were retained in the final model as significant or important for the model’s overall goodness-of-fit. The variables included in the final model were dependency status (DEPENDEN), marital status (MARRYDUM/DIVORDUM), high school diploma or GED/Certificate (HSDIPL02), and enrollment intensity (PARTDUM/MIXDUM). Those variables were either recoded into a binary format or a dummy variable was created for each value/category. I also included an additional three variables that literature has shown to potentially affect attrition. Previous research demonstrated that poor GPA. is a common reason given for leaving college, especially for traditional students (Barr, 2007; Burley, Butler, & Cejda, 2001; Hawley & Harris, 2005/2006; Mohammadi, 1994; Ogletree, 1992; Summers, 2003). BPS institutions reported GPA for the 95-96 year. I used GPA and recoded it into a continuous variable ranging from 4.0 to 0.0 in .5 increments. I did not use variables often employed to represent and predict student academic performance prior to college (e.g., ACT /SAT scores or high school GPA) because data was only reported for students who took the ACT /SAT and two-year institutions generally do not require an ACT/SAT score. Therefore, these data were not available for most students who began at a two-year institution. Another variable included in this study was students’ “reasons for enrolling” (SBENRPUR). There is some evidence that students’ goals entering into college are related to student departure (Choy, 2001). Additionally, the larger mixed methods 58 prrject of their colic; Th: infidel. Ge represents BLACKDl ASIANDI Katie, 011 to hare gut" Fin :DlSADV; littlest tdt Ml- Thre {”0 disadv; HDlSADV Dir {30.013 in . project of which this study is a part focuses on how students’ purpose and goals influence their college going decision-making. Three variables representing descriptive characteristics were controlled for in the A model. Gender (GNDERCOD) was a binary variable (male-=1). Race and ethnicity was represented by four dummy variables—WHITEDUM (White non-Hispanic), BLACKDUM (African-American Non—Hispanic), HISPADUM (Hispanic), and ASIANDUM. ASIANDUM included Asian/Pacific Islander, American Indian/Alaska Native, Other, and Non—resident Alien. The latter race/ethnicities were collapsed in order to have sufficient numbers for analysis. WHITEDUM was the reference variable. Finally, socioeconomic status was represented by a disadvantage index variable (DISADVAN). This index was constructed for BPS from thee other variables—parents’ highest education, percent of poverty level, and economic status of high school student body. Three dummy variables were constructed and used in the model—NDISADVAN (no disadvantage), MDISADVAN (minimal disadvantage), HDISADVAN(moderate/high disadvantage). NDISADVAN was the reference variable. Results Direct logistic regression was performed to assess the impact of a number of factors on a likelihood that respondents would return to college after a period of nonenrollment. The model contained 16 variables (see Table 8). 59 Table 8 Variables in Code $ariable Description Code Reference Stay_Stop2 Dropout/Stopout 0/1, 1=Stopout N/A ENDERCOD Gender M, 1=Female Male BLACKDUM Afiican-American/ O/ 1, 1=AA/Black White Black T—fi SPADUM Hispanic 0/1,1=Hispanic - White ASIANDUM Asian/Pacific 0/1, 1=Asian White Islander/American Indian/Alaska Native/Other/Non-resident Alien MDISADVAN Minimally Disadvantaged 0/1, 1=Minimally Not Disadvantaged Disadvantaged HDI SADVAN Moderate/High Disadvantaged 0/1, Mod/Highly Not Disadvantaged Disadvantaged WWUM Married 0/1, 1=Married ‘ Single DIVORDUM Divorced/Separated/ 0/1, 1=Divorced Single _ Widowed PARTDUM Part-time Enrollment 0/1, 1=Part-time Fulltime g Enrollment MIXDUM Mixed Enrollment 0/1, 1=Mixed Fulltime Enrollment HSDIPLOZ HS Diploma or 0/1, 1=GED/Certificate High School GED/Certificate Diploma mNDEN Dependent or Independent 0/1, 1=Independent Dependent JOBDUMMY Reason for Enrolling—Jobskill 0/1, 1=Jobskill _ Personal _\ Enrichment AACERTDM Reason for Enrolling— 0/1, 1=AA/Certificate Personal .\ AA/Certificate Enrichment TRANSFRDM Reason for Enrolling—Transfer O/l, 1=Transfer Personal .G\ Enrichment PARECOD GPA 95-96 4.0 4.0-3.75 N/A 7 3.5 3.75-3.25 3.0 3.25-2.75 2.5 2.75-2.25 2.0 2.25-1.75 1.5 1.75-l.25 \ 1.0 1.25-0.0 60 The full model was statistically significant with a chi-square value of 79.054 (df=16, n=779, p=.000), indicating that the model was able to distinguish between students who reported stopping-out and staying-out. Two pseudo R squared statistics were reported; while not identical to R in OLS regression, they indicate how much inclusion of the independent variables in the model reduces the variation (Menard, 2000; Peng et al., 2008). The model as a whole explained 11% (Cox and Snell R Square) and 1 5% (Nagelkerke R Square) of the variation in stopout. While not included in SPSS output, I also calculated McFadden’s index. I reported it here because it was “preferred over others for its conceptual similarity to the OLS coefficient of determination, its relative independence from the base rate, and its comparability across models comprised of different predictors yet applied to the same outcome variable and the same data” (Peng et al- , 2008, p.10). McFadden’s R squared was .088; values .2 to .4 are considered highly satisfactory (Cameron & Windmeijer, 1997). The percentage of correctly classified cases increased from 62.9% to 63.9% after the introduction of the independent variables to the ConStant-only model. Finally, the Hosmer and Lemeshow test for goodness-of—fit was Significant at 12.916 (df=8, n=779, p>-05)- Within the model, four predictors were found to provide a statistically significant Contribution at p<.01—minimally disadvantaged, divorce, mixed enrollment, and job Skill as reason for enrollment. The strongest predictor of returning to school was mixed eurollment, with an odds ratio of 1.759. This indicates that students who experienced both part- and full-time enrollment were 1.76 times more likely to return to college than t hose who were only enrolled part- or full-time; in other words, the odds of returning to 61 2"? college for students with mixed enrollment was approximately 76% higher than for students only enrolled part- or full-time. Divorce was also a stronger predictor, but of who was less likely to return. The odds ratio for divorce was .230, meaning that students who were divorced/separated/widowed were 77% less likely to return to college than those who were single or married. Students who went to college to develop job skills were also less likely to return than students who went to college for personal enrichment, associate’s degree, certificate, or transfer to a four-year college. The odds that a student who went to college for job skills will return to school after leaving was .502, or 49.8% less than students who went to college for the previously listed reasons. Finally, students who were minimally disadvantaged had decreased odds of returning compared to moderately/highly or not disadvantaged students, they were .606 or 39.4% less likely. Yet, when the odds of minimally disadvantaged students are individually compared to moderately/highly disadvantaged and not disadvantaged students, different odds ratios emerge. When the odds ratio for minimally and not disadvantaged students was calculated, minimally disadvantaged students were 34.6% (.654) less likely to return to college. Conversely, when minimally and moderately/highly disadvantaged students’ odds were compared, minimally disadvantaged students were 4.17 times more likely to return to college than moderately/highly disadvantaged students. 62 Table 9 Logistic Regression Predicting Likelihood of Returning to College After Nonenrollment Variables in the Equation 95% C.I.for EXP(B) ‘ B S.E. Wald df Sig. Exp(B) Lower Upper Step 1‘ GNDERCOD .054 .178 .092 1 .76 1.055 .745 1.495 BLACKDUM -.558 .287 3.773 1 .05* .572 .326 1.005 HISPADUM -.354 .286 1.537 1 .21 .702 .401 1.228 ASIANDUM -.117 .396 .087 l .76 .889 .409 1.934 MDISADVAN -.501 .197 6.454 1 .01** .606 .412 .892 HDISADVAN -.295 ..293 1.013 1 .31 .744 .419 1.322 DEPENDEN .601 .272 4.880 1 .02* 1.824 1.070 3.108 HSDIPL02 -.364 .274 1.764 1 .18 .695 .406 1.189 MARRYDUM -.676 .312 4.674 1 .03* .509 .276 .939 DIVORDUM -1.469 .506 8.423 1 .004** .230 .085 .621 PARTDUM -.453 .271 2.800 I .09 .636 .374 1.081 MIXDUM .565 .216 6.866 1 .009** 1.759 1.153 2.683 JOBDUMMY -.690 .281 6.046 1 .01** .502 .290 .869 AACERTDM -.312 .275 1.292 1 .25 .732 .427 1.254 TRNSFRDM -.012 .248 .002 1 .96 .988 .607 1.608 GPARECOD .174 .088 3.917 1 .04* 1.190 1.002 1.413 Constant -.537 .350 2.347 1 .12 .585 a. Variable(s) entered on step 1: GNDERCOD, BLACKDUM, HISPADUM, ASIANDUM, MDISADVAN, HDISADVAN, DEPENDEN, HSDIPL02, MARRYDUM, DIVORDUM, PARTDUM, MIXDUM, JOBDUMMY, AACERTDM, TRNSFRDM, GPARECOD. *p<.05 **p<.01 Discussion The overall model provides a better understanding of factors that are stronger predictors of the likelihood that students would return to college. The results that student 63 disadvantage, marital status, enrollment intensity, and reasons for enrollment were significant predictors of who dropsout or returns to college both supports and diverges from previous research. Most previous research on stopout was based on single institution studies and/or students at four-year institutions (Ahson, Gentemann, & Phelps, 1998; DesJardins et al., 1994; Herzog, 2004; Hoyt & Winn, 2004; Johnson, 2006; Porter, 2003/2004; Woosley, 2003/2004). Furthermore, with the exception of Stratton et al.’s (2004) analysis of stopout at four— and two—year institutions using the BPS 90/94, studies focused on community college students (Bonham & Luckie, 1993;‘Conklin, 1993; Grosset, 1993; Stratton et al., 2004) are older and more descriptive in nature. The fact that only two of the six variables identified as risk factors (the seventh variable, work intensity, was purposely not included to avoid multicollinearity), remained in the final model and were significant is surprising. Considering over 60% of the sample is at least moderately at-risk (using the risk index that is a compilation of the six risk factors used for this model), it is reasonable to expect that a number of the risk factors used in the model would provide a unique and significant contribution. As a result, one can assume that there are additional factors not included in the current model that would significantly contribute to an explanation of stopout as an outcome. This conclusion supported by the model statistics that demonstrated that while the model was significant at .000 and the percentage of correctly classified cases was above 50%, it only explained between 11% and 15% of the variance and correct classification of cases was only improved by one percent. Although the contribution of the model was not considerable, the individual variables that were significant provide additional information and contribute to a better understanding of stopout. 64 The in toroilege was par-time enro I993; Hoyt & In more prec times as 111131} the increased t could be that ll circumstances. drtumstances 1993; Conklm halt derelope. Efrem enrol} Til-math PCIIC ammmOdall‘U The finding that enrollment intensity was a significant predictor of student return to college was not unique in comparison to previous research. Multiple studies found that part-time enrollment was associated with an increased likelihood of stopout (Gosset, 1993; Hoyt & Winn, 2004); unique to this study was the findingthat mixed enrollment was more predictive of student return to college than either part-time enrollment (4.17 times as likely) or full-time enrollment (7.2 times as likely). One potential explanation for the increased odds of stopout among students who have a history of mixed enrollment could be that it indicates a higher level of adaptability and flexibility to life circumstances. Many studies found that students often cite conflict with work and life circumstances as primary reasons for leaving (Ahson et al., 1998; Bonham & Luckie, 1993; Conklin, 1993; Hoyt & Winn, 2004; Light, 1993). It is possible that students who have developed the flexibility in their enrollment pattems—willingness to increase or decrease enrollment—to accommodate conflicts within their non-academic life may approach periods of non-enrollment not as leaving school, but as a temporary accommodation for other life events. Being divorced/widowed/separated (div/wid/sep) was another significant variable within the model. Div/wid/sep students were 77% less likely to return to college than married or single individuals (together), and almost 100% less likely than married or single individuals analyzed as separate groups. Students who are divorced may have less support than students from either of the other two categories. Eighty—two percent of single students are dependents, while 100% of married students and 95.7% of divorced/separated/widowed students are financially independent. Therefore, single students may be more likely to have the financial and personal support of parents and 65 ill; tl‘.‘ 4:" Ussl mgr iL PM. i, “'15:. family and married students are also likely to have the resources and support of a spouse, while divorced/separated/widowed students are likely to be independent from family without the support of a spouse. Having less resources and support may contribute to the decreased odds of returning to school. Previous research has had mixed findings. Similar to the findings in this study, Stratton et al. (2004) found that marriage was associated with stopout for men and women and divorced women were more likely to dropout Singlehood was also associated with returning to college (Hoyt & Winn, 2004). But when there was a change in marital status, which was not addressed in this study, getting married was associated with a higher likelihood of dropout while divorce was associated with stopout (Stratton et al., 2004). Also, divorced and married students are in higher risk categories (i.e., level of risk for attrition/non-traditional status) than single students. Relative to the overall number of students at each risk level, the number of single students at each risk level decreases as the risk/non-traditional level increases (see Table 10). Conversely, for married and divorced/separated/widowed students the percentage of students at each risk level increases as the risk level increases, this is especially true for divorced/separated/widowed students. The additional risks may also contribute to the decreased odds of returning to college. 66 Table 10 The Percentage of Students Single, Married, and Divorced/Separated/ Widowed Students Who are Low, Minimal, and Moderate/High Risk of Attrition/Non-traditional Marital Low Risk Minimal Risk Moderate/High Status . (0-1 Characteristics) (2-3 Risk Characteristics) (4-7 Characteristics) Single (n=569) 53% ' 34% 13% Married (n=145) 0% 30% 70% Divorced/Separated/Widowed 2.1% 25.5% 72.3% (n=47) Note: The percentage of students in each risk category who are single, married, and divorced/separated/widowed is different from the percentages listed because of relative number of students in the study. The study’s findings indicate that students who enroll in college for job skills have decreased odds of return (50% less likely) compared to those who enroll for personal enrichment, AA/certificate, or transfer (together). When the odds of returning for students who enrolled for job skills is individually compared to students who enrolled for personal enrichment (.853), AA/certificate (.773), and transfer (.256), the former students remain less likely to return than individuals in the other three categories. While previous research indicates that student goals for enrollment are negatively associated with departure and positively associated with persistence (Choy, 2001; Summers, 2003), most studies on stopout do not include factors related to goals/reasons for enrollment. The inclusion of this variable sheds light on how goal intentions were related to students’ decisions to return to college. The result that students enrolled for job skills were less likely to return than students pursuing an AA/certificate or transfer makes intuitive sense. Students can attend college, gain the skills they need and leave having accomplished their goal but without a 67 degree/certificate or transfer. While students pursuing a degree/certificate or transfer would likely need to persist over a longer period of time and have clear achievement markers (i.e., degree or transfer) when they complete their goal. Specific goals may also provide clearer motivation for return when they leave college. On the other hand, the fact that students who enrolled for personal enrichment were also more likely to return to college than students who enrolled for job skills is also understandable. Unlike students returning for transfer/degree, personal enrichment students generally do not have a particular educational goal in mind, or one that does not include long-term enrollment/attendance, much like those attending for job skills. This may indicate students attending for personal enrichment may view two-year institutions as places to attend not only for a particular degree or job-related skill, but to develop skills/knowledge that are personally important. Therefore, students pursuing personal enrichment may be more likely to return because they view education as ongoing development versus attending college for a terminal purpose (e.g., job skills or degree). Finally, the only background characteristic that was also a significant predictor was that for students who were minimally disadvantaged. Initially the finding that minimally disadvantaged students were less likely to return to college than moderate/high or not disadvantaged (together) students is perplexing because it is reasonable to expect that students who were minimally disadvantaged would be more likely to return than students who were moderately/highly disadvantage. Yet, when the Odds ratios for not and minimally disadvantaged students (1.53) and minimally and moderately/highly disadvantaged students (4.17) were individually compared a clearer pictured emerged. Not disadvantaged students were one and a half times more likely to return to college 68 than minimally disadvantaged students, but minimally disadvantaged students were more than four times as likely to return to college than moderate/highly disadvantaged students. There is evidence in the literature that higher parental education is related continuous enrollment and persistence (Stratton et al., 2004; Swail, Cabrera, & Lee, 2004; Terenzini, Cabrera, & Bemal, 2001; Tinto, 2006; Titus, 2006; Walpole, 2003b). In studies of dropout and stopout, financial reasons (e. g., unmet need and cost of tuition) were often cited as reasons for departure (Ahson et al., 1998; Bonham & Luckie, 1993; Conklin, 1993; Herzog, 2004; Hoyt & Winn, 2003/4; Light, 1993). The findings in this study support previous research, indicating that students with more financial need are less likely to return to college or persist. More unique to this study is that the variable used to measure disadvantage is not only based on income (i.e., poverty index), but includes factors related to social capital (i.e., parents’ highest education and economic status of high school). Therefore, the variable representing this index not only has implications for a better understanding of the association between family income and stopout, but more broadly about the relationship between social capital and stopout. Social capital is the social relations and structures that form highly interconnected networks (e.g., family, school relationships, work) through which norms, expectations, and standards are developed (Bourdieu, 1987; Coleman, 1988; Kim & Schneider, 2005). It is particularly through parents and family networks that critical information and values for educational success are transmitted to children (Coleman, 1988; Kim & Schneider, 2005; Schneider & Stevenson, 1999). Research has demonstrated that social capital has an influence on the educational outcomes of children (Coleman, 1988; McNeal, 1999; Stanton-Salazar & Dornbusch, 69 1995; Sun 1999) and transition to postsecondary education (Kim & Schneider, 2005; McDonough, 1997 ; Outcalt, Tobolowsky, & McDonough, 2000; Walpole, 1998, 2003) independent of socioeconorrric status (SES). Therefore, the use of a variable that includes indicators Of social capital in addition to SES suggests that students with less social capital (not only less income) are less likely to return to college. The study’s findings in many ways affirmed previous stopout research. Yet, analysis of the significant variables also provided new information for understanding factors related to who returns to college. The findings and discussion concerning reasons for enrolling was a new contribution to understand who returns to college. Furthermore, while previous research discussed stopout and enrollment intensity, marital status, or levels of disadvantage, this study contributed and extended those discussions. Limitations There were a number of limitations to this study. First, the study is limited to and specifically focused on variables related to non-traditional student status. Two of the seven characteristics (having dependents and single parenthood) were eventually removed because they did not significantly contribute and I preferred not to impute due to the potential reduction to variation. Since there were no significant relationships between missing data within the independent variables and the dependent variable, I chose to exclude those variables. One variable was never included (work intensity) due to multicollinearity. Second, in an effort to remain focused on the relationship between community college student risk factors and stopout, I did not include variables often associated with students who attend four-year institutions—-academic integration, social integration, and 70 loans/financial aid (Beil, Reisen, Zea, & Caplan, 1999; Braxton, Sullivan, & Johnson, 1997; Borglum & Kubula, 2000; Cabrera, Nora, & Castaneda, 1993; Dowd, 2004; Napoli & Wortman, 1996, 1998). While this exclusion allows for a better understanding of a I specific set of factors, it also excludes other factors that could further explain the outcome. Third, I did not calculate the intracluster correlation coefficient (ICC) to measure the internal homogeneity of clusters within the design, a measure that could provide an indication of the degree of bias in estimating variances in data collected through cluster samples. (Thomas & Heck, 2001). Also, I did not compute the delta-p for the continuous variables in the model. This measure would indicate the change in probability per one- unit change in the independent variable (Peng et al., 2002). Additionally, while using a more conservative alpha-value and relative weight to alter the sample size are acceptable adjustments for calculating more accurate standard errors, using specialized software is preferable (Thomas & Heck, 2001). Finally, due to financial and resource constraints it was not possible for BPS researchers to survey the entire population of students and institutions, therefore it was necessary to use a sampling approach for estimation of trends and patterns across the data. To ensure appropriate representation of the population researchers oversampled selected subgroups of students and institutions. This results in a distortion of the overall _ representation that required adjustment through the application of a weight. In addition, the clustering design approach contributes to nonindependence of observations and nonhomogeneity, violating the assumptions of simple random sampling. A more conservative alpha level and normalized weight were applied to compensate. 71 Implications and Further Research First, the lack of variance explained and goodness-of-fit indicates that there is significant room within the model to explore other factors that contribute to understanding why some students return to college and others stay out. As mentioned previously, including more independent variables that previous research has demonstrated may be related to attrition for traditional and nontraditional students may result in a model that is more indicative of factors that are influential in students’ return to college. Potential variables include the full seven risk characteristics and indicators of academic integration (e.g., use of library, interaction with faculty, study groups), social integration (e.g., participation in social clubs, participation in intramural sports, spend time with other students outside of class), and financial aid (é. g., grants and loans). BPS has an academic integration index and social integration index compiled from individual variables identified as relevant to these areas of integration. The drawback to the social integration index is that many of the variables would be more applicable to students at four-year institutions (e.g., varsity sports, social clubs, intramurals). Variables associated with financial aid are plentiful within BPS as NPSAS was designed to collect data about how families afford and pay for college. Second, the significant variables within the model associated with odds of stopout and dropout were varied and complicated. Complexity within and between factors such as the interplay between financial resources and social capital within student disadvantage, the status of personal support and resources post-divorce, the perception of the role postsecondary education in the reasons for enrolling, and the differences between 72 A . Vei- rv. a P: Ill 5. bh“ if; ‘5 3‘5 . ‘1'“ I .111 ‘ "“1“ enrollment intensity types are challenging to fully understand and capture exclusively through quantitative approaches. Qualitative exploration of why students return to college, how they go about making their decisions, and what factors influence that decision can shed light on the nuances of this issue. The unexpected significance of mixed enrollment suggests that institutions should consider how they might utilize this type of enrollment to promote persistence and return to college among community college students. Often full- and part-time enrollment patterns are pitted against one another as being more or less supportive of persistence. For students who are unable to attend full-time Consistent, I would suggest to institutions that faculty and staff encourage students to approach enrollment flexibly. Therefore, rather than leaving school when school-work-life balance becomes a challenge, students should consider reducing and increasing enrollment as their lives allow. Finally, the findings and analysis within this article provide helpful information about which factors may be more influential than others in understanding who returns to college and afford insight into issues within student lives that contribute to their educational decisions and paths. For practitioners, this is a reminder that students’ lives are complex and therefore result in a variety of educational paths that may not include continuous enrollment. Such complexity of personal and academic issues is a reminder that there can be no “one program or strategy fits all” for student retention re-enrollment. For example, one student who works full-time, is a single parent, and is trying to go to school may be able to maintain enrollment by being open to part- or full-time enrollment and utilizing the on-campus daycare designed by the institution to support students with children. Conversely, another student may have a similar academic, work, and personal 73 situation, but is so overwhelmed that he or she is getting sick from the stress. It may be more appropriate for this student to take a semester off and work with the institution ‘ social worker to develop a plan for return. Acknowledging the uniqueness of each student’s circumstances allows institutions to better respond to attrition and create effective programs and plans to assist retention or return, but a number of challenges present when this type of recommendation is made. In order for individualized plans to be developed in an effort to address attrition, students must connect with the institution. In an open access institution where students attend for multiple reasons, many of which do not include a degree, certificate, or transfer, identifying students who may need support is a challenge. I suggest that, first, campus staff, admnristrators, and faculty need to take shared responsibility in connecting with students. Since most community colleges are commuter campuses, students may only connect with a campus official in the classroom, at financial aid, or once a semester with an advisor. Ensuring that campus staff and faculty view each of these points of connection as an opportunity to converse with the student on their goals, progress, and plans for enrollment may assist in reducing campus attrition. Second, implementing campus systems that require or prompt students who do not enroll for the next session (and are not transferring or graduating) to see an advisor will help create shared responsibility with the student and reduce some of the burden across the campus while individually connecting with students to better understand their departure and provide support. Another form of outreach to students would be to follow- up with students who leave within the year to encourage and support their return. 74 Summary This study explored the influence of factors associated with nontraditional student status and risk of departure on students’ return to college. While the overall model did not considerably contribute to an explanation of stopout, the significant variables provided a better understanding of which individual factors were more important for students’ decisions to stay out or retum to college. The results that mixed enrollment, divorce/widow/separation, enrollment for job skill development, and disadvantage contributed to increased or decreased odds of retum to college extend current understanding of why some students who leave college are more likely to return than Others. The results draw attention to the complexity of student lives and the multiple ’ academic and non-academic factors that they must consider when making decisions about their educational paths. 75 CHAPTER 4: ARTICLE 2— QUALITATIVE STUDY Introduction There are a significant number of students who do not achieve and accomplish their higher education goals. Students at higher risk of not graduating with a degree/ certificate are more likely to attend community colleges where uninterrupted attendance and persistence to educational goal achievement are less likely than for those students at four-year institutions (Horn, 1996). As institutions that serve a larger proportion of students who are more likely to leave higher education without a certificate or degree, community colleges are charged with supporting and assisting their students to aChieve a broad spectrum of educational goals (Berkner, Horn, & Clune, 2000). Furthermore, many of these students face numerous personal, academic, and institutional Challenges to their educational goals (Hoachlander, Sikora, & Horn, 1998; Horn, 1996). One choice students often make is to leave college and return later to continue their education (Horn, 1998). This phenomenon, stopout, is associated with multiple falCt()I‘s-—-—poor academic performance, financial burden, and family responsibilities are aniong the most common reasons (Horn, 1998). What is troubling to institutions about Stopout is that students who experience an interruption in enrollment are more likely to experience subsequence stopouts and are less likely to graduate (DesJardins, Ahlburg, & lVlccall, 2006). The research on stopout is focused on the reasons student leave, how Stopout affects student outcomes, and differentiating stopout from dropout behavior. Yet, llttle is known about why students return after a period of stopout and how they make the 76 decision to return. This information may be valuable to institutions as they attempt to retain students in jeopardy of dropping out, assist many students in breaking a cycle of enrolling and leaving college without achieving their educational goals, and support local community members in returning to college and finishing their education. The latter reason is particularly relevant to institutions, such as the college in this study, in states that have traditionally relied on manufacturing industries and whose labor force and economies are struggling to shift to a knowledge-based economy. The purpose of this study is to explore external and internal reasons and factors involved in individuals’ decisions to return to college after an extended absence. Specifically, drawing on the qualitative portion of mixed-method research, this study focuses on the following research questions: 1. Why do students return after a period of non-enrollment? 2 What factors impact this decision? 3. What is the influence of these factors on students' return to college? 4 What role do college and career possible selves and learner dispositions play in the decision to return to college after an absence? 5. How does stopout influence on individual 's academic and career possible selves? This study is informed by scholarship on stopout behaviors, factors related to dropout and retention, and career and academic possible selves (PSs). This paper will also address a II"Odel of stopout paths related specifically to the decision to return to college. A discussion of implications for students and institutions completes the paper. Background P - . er S lstence for community college students 77 While there are concerns about academic achievement of students at higher risk of dropping out across the educational pipeline, in the postsecondary sector a primary concern is retaining students through their academic goals and helping them to endure through the academic, organizational, and personal factors that challenge their ability to persist. Across postsecondary education, approximately 30% of students will depart from higher education (Horn, 1998). Yet, the relative percentage of students who leave from four-year institutions (16%) is not equal to the percentage of students who leave from two-year colleges (44%) (Horn, 1998). This inequality draws attention to the community college sector for the study of persistence (Horn, 1998). The open access, low-cost characteristics of two-year colleges have broadened access to higher education for higher-risk students facing challenges such as financial hardship, poor academic preparation, a lack of basic English language, or proficiency Skills (Grubb, 1999). As a result, approximately 54% of community college students enter with at least one characteristic that places them at risk for not completing their POStsecondary education (Hoachlander et. al., 2003). These characteristics are typically linked with nontraditional student status such as delaying enrollment between high school and college, part-time enrollment, working full-time while enrolled, completing high S(31'1001 by certificate or GED, having children, being a single parent, and being financially independent (Feldman, 1993; Horn, 1996; Horn & Premo, 1993; Lee, 1996). Furthermore, while personal and financial reasons are among the most often cited reasons for departure, a lack of academic preparation and poor academic performance are alSO cited as reasons for leaving higher education (Bonham & Luckie, 1993a; Lee, 1996; Ogletree, 1992). In addition, the majority of students with higher risk factors (e. g., 78 nontraditional student status, poor academic preparation, and poor academic performance) for early departure choose to attend community colleges versus four-year colleges for their first experience with higher education (Adelman, 1999; Horn, 1996). The figures previously mentioned demonstrate that (a) community colleges continue to attract students at a higher risk for departing from college before they achieve their educational goals and (b) community colleges are important and authentic settings to study students at higher risk of leaving postsecondary education prior to completing their educational goals. One of the challenges that these institutions confront when considering how to improve student retention is the stopout phenomenon. Stopout behavior A significant challenge to institutional retention and overall postsecondary educational persistence is a phenomenon that occurs when students leave college after their first year and return at a later date: stopout. Horn defined a first-year stopout as “a beginning student who interrupted his or her enrollment in the first year with a break of at leaSt four months before reenrolling. . .This includes students who finished their first year, bUt did not reenroll for a second year” (1998, p. 4). This study expands this definition to not only include students who lefi during or just after their first year, but also students Who were enrolled for more than their first-year when they first left college. In addition, many of the students in this study stopped out multiple times in their postsecondary ed“Cation career. Of concern to institutions is that students who experience an inteI‘I'Uption in enrollment are more likely to experience subsequence stopouts and are less likely to persist and graduate (DesJardins et al., 2006). These outcomes bring into focuS the potential problems with stopping out for students and raise the issue of how 79 institutions can help students maintain continuous enrollment and provide support for students to return to college. In recent studies of attrition and retention, researchers began to recognize and take into account this stopout phenomenon. Approximately 30% of students in all of higher education leave afier the first year, yet half (57%) of those who leave will return to higher education within five years. (Horn, 1998). Of those who leave postsecondary education, 63% are from community colleges, 22% are from four-year institutions, and 15% are from other types of schools (e. g., private, for-profit; public less-than-two—year; or private, not-for-profit, less than four-year institutions). Yet, 50% of those who depart from the two-year sector return, while 64% of those who lefi a four-year institution return within five years. This gap in returns indicates that students who began at a two-year institution and leave after their first year of school are less likely to return to college than students who begin at four-year institutions, but represent a greater number of students overall. In other words, out of 100 students who leave during their first year, 85 of them are fi'om two- and four-year institutions. Of these 85 students, 46 will return within five years; 32 of those who return originate from a community college versus the 14 who begin at a four-year school. Furthermore, of the students who leave during their first-year and return within five years, a majority will return to community colleges with 79% of students who started at a two-year and 24% of students who started at a four-year institution returning to community colleges (this includes students who return to the same institution at where they began and students who transfer from another two- or four-year institution) (Horn, 1998). For example, out of 100 stopout students who begin at a two-year institution, 79 of 80 them will return to a community college and the remaining 21 will attend four-year or other types of institutions. Additionally, out of 100 stopout students who begin at a four- year institution, 24 of them will return to a community college and the remaining 76 will attend four-year or other types of institutions. The previous figures indicate that a majority of students who depart higher education after the first year will leave from community colleges. Furthermore, of all students who leave a majority of them will return to two-year institutions. In light of this evidence, exploring stopout behavior at community colleges is sensible, as it is these institutions that most often encounter stopout behavior and, therefore, are likely to be most vested in responding to it. Furthermore, knowing and understanding not only why students leave but why they return, how to facilitate that return, how they think about that process, and what is most influential in that decision can be helpful to researchers and practitioners to (a) encourage and help students not to leave in the first place and (b) identify ways to support students to return to their educational goals. There are multiple important influences and factors that contribute to how and why students decide to return to college. Based on previous research related to aspirations and goal achievement, one of these factors is likely to include the academic and career hopes students have for themselves and the ways that students view their hopes and fears of achieving and becoming these possible selves. Their success in breaking a stopout cycle and avoiding subsequent dropout also appears to be tied to a student’s disposition toward learning. 81 Possible selves Research has demonstrated that the academic aspirations students hold for themselves are related to their ability to persist and achieve their goals (Kerpelman, Shoffner, & Ross-Griffin, 2002; Leondari, 2007; Oyserman, Bybee, & Terry, 2006; Pizzolato, 2006). One approach to this concept is the study of possible selves (PS). Possible selves are the cognitive representations and ideas of what individuals believe they might become, what they would like to become, and what they are afi'aid of becoming—the cognitive components of their hopes, fears, goals, and threats (Markus & Nurius, 1986). Markus and Nurius (1986) claim possible selves are important for two reasons. They function as incentives for future behavior and serve as a cognitive bridge between who an individual is now and who they can be in the future; possible selves are part of the connection between their motives and goals and the behavior/action necessary to achieve those goals (or avoid certain selves) and maintain their motivation. Motives are represented as “dispositions” to strive for particular incentives (goals) or avoid negative incentives (threats). Possible selves cognitively represent these motives and the plans/paths for achieving them. Markus and Nurius (1986) suggest that desire or motivation is not sufficient for motivating behavior, but that the desire must be translated into a vision of the self as healthy, active, and strong and must be accompanied by specific plans and strategies for becoming these possible selves. These possible selves are cognitive representations of the incentives for mastery, and without them there should be little instrumental behavior in the direction of mastery. (1986, p. 961) Furthermore, balance between hoped for and feared possible selves is related positively to academic persistence among undergraduates and African-American middle- 82 school students (Oyserman, Gant, & Ager, 1995) and a strong relationship between self- regulation strategies and academic possible selves is related to academic achievement and persistence (Oyserman, Bybee, Terry, & Hart-Johnson, 2004, Oyserman et al., 2006; Oyserman & F ryberg, 2006). Another important mechanism within possible selves theory is the concept of self-schemas. Self-schemas are cognitive representations of the procedures and concepts necessary to bridge a possible self to the behavior needed to attain or avoid it. Developed over time through past experiences, self-schemas provide a base knowledge and organization of one’s ability in a particular domain and give direction, form, and meaning to a person’s current evaluation of his or her abilities in the present. Pizzolato (2006) found that among hi gher-risk college students, not only are procedural schemas important, but conceptual schemas for the development of meaning and purpose are just as necessary. In relationship to the issue of stopping out, it is reasonable to believe that in the process of leaving and returning to postsecondary education, who a person believes he or she can be academically and in his or her career could play a part in what that student’s goals are and become, which academic goals he or she pursues (e. g., returning to school, changing a major upon return, not returning to higher education), and, ultimately, whether or not he or she returns to college after an absence. Learner disposition One of the “dispositions” likely associated with academic and career possible selves is a student’s perception of and beliefs about learning. Two components make up the concept of learner disposition» approaches to and perceptions of learning (Bloomer, 1996, 1997; Bloomer & Hodkinson, 1999, 2000). First, Bloomer (1996) found that when 83 students’ values, beliefs and expectations about learning are in contradiction with those of the instructors, students’ respond with the following approaches: strategic compliance, retreatism (absenteeism), rebellion (disruption), or innovation (achieving the learning outcomes outside of class). When student and instructor expectations are in line students conform to instructor objectives and expectations. Second, a student’s approach to learning opportunities stems from his or her perception of learning and education. There are multiple factors that influence students’ perceptions of learning: learners’ beliefs about the nature of knowledge, the purpose of secondary education, their abilities based on prior learning experiences, and the value placed on particular areas of study and learning experiences (Bloomer & Hodkinson, 1999, 2000). One’s learner disposition may represent the schema or plan he or she has developed for academic success and achievement. A schema is developed based on prior experience and observations. A student’s learner disposition is likely to influence the development of academic possible selves, in addition to potential changes to possible selves and associated schema over time. Messages about education from students’ families, their academic self-efficacy based on previous experiences, and the purpose and value ascribed to postsecondary education can shape whether or not the person envisions a collegiate possible self and, if so, what that possible self will look like. Entering college with a certain disposition and academic possible self is likely to shape that college experience and the decision to stay or go when a challenge presents itself. Furthermore, and salient to this paper, as events, environments, and context change for students, their perception, value, and approaches to education also change (Bloomer & Hodkinson, 2000). These shifts in learner disposition may have an affect on the student’s academic/career possible selves and his or her ability 84 to develop a plan to achieve them. As students make the decision to return to college and to end the stopout cycle of repeatedly exiting and re-entering college, a shift in learner disposition and PS5 may lead to persistence and accomplishment beyond what students have previously experienced. Methods After two pilot interviews, I conducted 30 to 60 minute interviews with 48 participants from an urban, mid-westem community college. The pilot interviews were used to form the interview protocol in conjunction with relevant literature. The interviews took place on school grounds using a semi-structured, open-ended protocol (Merriam, 1998). Students described why they left school and what their decision to return to college looked like. I drew upon possible selves theory to develop the protocol and in analysis to further understand how a student’s conceptions of his or her possible academic and career selves influenced the decision to return. Understanding how participants conceive of their possible academic and career selves prior to departure and when returning provided insight into changes in the perception of possible selves and strategies for the achievement of possible selves that are important in the decision to reenroll. I used three methods to recruit students: 1) I posted fliers advertising the study across the campus; 2) l recruited in a limited number of classes; and 3) an email was sent to the entire student body by the college on my behalf. A majority of the participants were recruited through the email. All participants were made aware of their rights and reimbursed $15 for their participation. 85 Data source Location. This study was conducted at an urban, mid-western community college (using pseudonym of Archer Community College) that has 10,456 students. Approximately 60% are female, 25% are students of color, and 64% are part-time students. The average age of students is approximately 28-29. I selected this college because it is relatively large compared to other community colleges in the state and its racial and ethnic proportions are similar to the state’s demographics. The city is approximately 53% Afiican American, 41% White, 3% Hispanic/Latino, and 1.1% of more than one race; all other racial and ethnic groups are at less than 1%. This demographic division of race is roughly reflected in the sample of study participants. It is the fourth largest city in the state, yet the population has continued to decline in the last decade with a reported population of 131,000 in 1998 and 114,662 in 2007. The median income for the city is $28,010 with 26.4% of the population living below the poverty line; 37 .4% of those under 18 are below the poverty line (City of Archer, 2008). Since the 19205, the city’s economy has been dominated by the auto industry as home to a major American auto company and the United Auto Workers union. As the American auto industry experienced decline and deindustrialization there was (and continues to be) a loss of industry and jobs in the area and, as a result, an increase of significant economic challenges for the city’s residents (Dandaneau, 1996). The study’s findings demonstrate the influence that the local economy had on many students’ decision to return to college. Participants. I interviewed forty-eight students about their decision and experience returning to a community college campus. Approximately a quarter of the 86 participants were students of color (primarily African American). The sample ranged from 20 to 50 years old, with a mean age of 32.5 and 52% of participants falling between 23 and 33 years old. Fifty-two percent of the participants were women and the rest were men, with one female-to-male individual. All the participants were pursuing transfer, a certificate, or an associate’s degree and were currently enrolled. A majority of students (66%) have only stopped out once, but that stopout ranged from one full semester (not including a summer semester) to 15 years. Finally, while stopping-out places these students in a higher risk category for not completing their educational goals, 86% of the students also had at least one additional risk factor when they first left college for not completing their educational goals associated with non-traditional college students (F eldman, 1993; Hoachlander et al., 2003). Pseudonyms are used for all institutions and participants. Participants selected their own names. Interview. Previous studies have found that asking students to reflect on important events and, especially, decisions provides a context to more clearly describe their decision making process and what was important within it (Baxter Magolda, 2001). To facilitate this inquiry, the interview questions guided students through their educational history; their goals, hopes, and fears and how they changed over time; and their decision making process to leave and return to college. A sample of questions asked included: Tell me about your experience in school before college; What did you see yourself like as a student? Did you have plans for how you were going to achieve those goals? What did you hope to be like as a student? What did you see happening when you left? Did your academic goals change while you were gone? ls how you see yourself as a student or in your future career different from before? Furthermore, in addition to playing a role in the 87 theoretical and conceptual frame for the study and protocol, I utilized possible selves theory in the analysis of the data. Analysis I used constant comparative analysis to analyze the data (Strauss & Corbin, 1998). I read the transcripts for codes and themes focusing on the reasons students’ returned to college, strategies and plans for returning, hopes and fears for academic and career selves, and the intersections among these areas. A first pass at coding revealed themes in accordance with those previously listed. Additional analysis exposed more detailed patterns among student responses that identified life events that influenced the decisions to leave and return to college, experiences that affected dispositions toward college and learning, and two different paths through college shaped by educational attitudes and possible selves. In an effort to establish trustworthiness, emerging themes and patterns were shared with participants for feedback (Lincoln & Guba, 1985). During the interviews many students expressed an interest in learning more about the study’s findings. 1 took the opportunity to ask these students if they would be willing to review findings and provide feedback. Due to the non-traditional characteristics of the majority of students in the study (e.g., single parent, above the age of 25, married, part-time student, full-time employee), finding time to meet is a challenge, therefore I emailed students a short explanation of findings and themes and they emailed back their feedback. I used the feedback to evaluate the patterns and made adjustments to more accurately reflect participants’ experiences. Most respondents affirmed the patterns I identified and only suggested minor additions or adjustments. In addition, once ten interviews were coded an 88 individual unrelated to the project was asked to review a set of three interviews with the previously constructed codes. I compared the coding results by the different reviewers and made adjustments to themes and codes according to differences. In addition to evaluating the trustworthiness of the findings with the participants, I also compared the findings to the outcomes of a semi-parallel quantitative study. The quantitative study was an analysis of the relationship between characteristics and factors recognized in the literature as related to departure from college without completion of a degree or certificate and the whether or not students returned to college after a period of unenrollment. I used the Beginning Postsecondary Surveyz96/01 to identify community college students who began college in 1996 and left with or without return prior to 2001. While this survey does not include items that allow me to evaluate the outcome’s relationship to possible selves, I was able to compare the qualitative themes regarding the role of risk-factors in the decision to leave/return and reasons for initially entering college to the findings from the quantitative study. The results indicate that students entering to develop job skills were less likely to return to school than those who enrolled for a degree/certificate, transfer, or personal enrichment. Further discussion of this outcome is in my dissertation Article Three. Limitations Recognition of this study’s limitations can assist understanding of how to apply these findings and how further research can extend this work. First, while the racial demography of this study is roughly similar to that in the state, students of color were a relatively small percentage of the participant group and primarily African Americans. Second, the state and particularly the city that the college is located was dominated by the 89 auto industry. As deindustrialization began to affect auto manufacturers, the economy and employment in this area has been struggling for the last 20 years and continues to be affected by the national economy. While this study will likely be applicable to institutions in areas with high unemployment and significant economic hardship, this level of dependence on manufacturing industries and the subsequent suffering is not necessarily reflected in other areas and populations across the country. Third, this is a retrospective account of student experiences. Therefore, the benefits of hindsight and casualties of memory do not always reflect how students may have reported the same experience when it happened. Finally, this study focused on students from one community college and findings may not reflect student experiences with stopout at other community colleges or types of institutions; generalizabiity is limited. Findings The purpose of this study was to explore the external and internal reasons students returned to college, the factors associated with return, and specifically the role that college and career possible selves and learner disposition has on the decision to return. Four themes emerged in an general stopout cycle model as a result of this study: 1) academic experiences in high school influence students’ entry into college and risk-factor associated with persistence; 2) students have multiple and intertwined reasons for leaving college--extemally and internally located; 3) as with the decision to leave, the reasons students return to college are complex and varied; and 4) whether or not students stopout once or multiple times, their ability and motivation to persist through to their educational goals (i.e.., certificate, associate’s degree, or transfer) without leaving again appears to be marked by a change in learner disposition and possible selves. Furthermore, students 90 move through this cycle along two distinct paths that are initially distinguished by their college PS and the learner disposition that they displayed in high school. To be evidenced in the discussion of themes, their beliefs about and attitudes toward college and learning in secondary school influenced their reasoning and decision to leave and return to college. One way that these themes became evident was through the recollection of participants’ educational experiences from K-l2 through their current enrollment. I will illustrate these themes through the general path that is represented by the model in Figure 1. The general model is expected-~earlier K-l2 academic experiences influence if and when students enter college for the first time. Furthermore, additional experiences in college affect the decision to leave and subsequently return. Students may move through stopout cycles of leaving and returning to college multiple times over many years. Finally, they will either continue in the stopout cycle, persist through and accomplish their educational goals or leave college all together. While this figure demonstrates the different paths that individuals might make from high school through college with the inclusion of stopouts, it does not represent the internal and external factors that influence student decisions through each step. The factors and specifically the role of learner disposition and possible selves are explored in more depth through the discussion of the themes. 91 Figure l: Stopout Cycle 4? Return Persist College srorour CYCLE mmugh High Diploma Experience Educational School '2'? Or l=> D Goals Experience GED or Leave College Permanently \ U Leave Leave High School Academic experiences in high school. Prior to college, students had varied experiences in the K-12 system that influenced when and how quickly they moved into college. For some students, going to college was not much of a question, it was the default, and they moved into college right after getting their diploma or GED. For these students, academics and college were important to them and were a clear part of their “hoped for” possible selves, “I had planned on just going right off to college. . .Me and my best fiiend was going to go and she was going to be a lawyer and I was going to be a lawyer. . .I always wanted to be like the person that other people looked up to. I always wanted to want to have a good grade.” (Tee). Many had parents who were supportive of college and peers who took a college path. Annette was told by her family that “school was number one.” In addition, she played sports in high school and went to a college where she could continue playing softball. Others were focused on college-going despite coming from less supportive family backgrounds. Chad stated that it was his abusive family background during his K— 92 12 educational years and “all my life being told I was stupid” that made him want to go straight into college. He was “gonna prove everybody wrong. . .it was pure anger.” College possible selves. Students along this path had college possible selves that were clear and had personal meaning and purpose beyond a simple expectation from parents or society. Their experiences in high school were generally positive and laid the foundation for learner dispositions that framed education positively. Their conceptual schema for college was connected to their identities and possible selves that they envisioned for the future, and were balanced with feared possible selves. While these students left college, they always had the intention of returning and, in general, had a higher number of stopouts because of the ongoing desire to return to college throughout the years. Non-college possible selves. Students along the other path had a less positive experience in school. Some students finished high school, but took time off in between. Beth had poor experiences in secondary school that affected her motivation, “I went from a 4.0 student to barely passing. . .’cause they said I was a girl, and I wasn’t that smart, and girls can’t play them sports. . .if they didn’t care about my education, why should I. . .so I gave up and didn’t care to go to college.” Bob left high school early, acquiring his GED later. While Bob described himself as an “underachiever” and “slacker” who didn’t care about school or see the purpose of college, “I didn’t, at the time, take it seriously, I’d think something will come up. I didn’t really have no plans. As far as a career, anything that paid good if] could get into it.” Many of these students did not see the need for college because they saw their families and peers making a good living doing trade work or working for the automotive industry. According to Blake, “. . .I knew I would find 93 work. There’s just stuff out there to do, so my father--an electrician, my grandfather-~an electrician. Everyone works for the automotive, especially in (state). I guess I felt that’s where I was established in life. It’s (college) not meant for us.” Blake always saw himself as mechanically inclined and had a lot of family who worked in the factories. People would ask him about college and he would tell them he wanted to go, but never made plans; he did not see himself as an “academic type.” Finally, other students along this second path left high school before graduating and completed a GED or diploma at a later time. Some of these students continued straight into college and others delayed enrollment. Most of these students reported negative experiences in school academically or being in the wrong peer group. Bill “hung out with the wrong crowds and ended up getting arrested and put in jail” and went through a rehabilitation program before getting his diploma. Autumn got pregnant her 9th grade year and did not have childcare to go back to her alternative school, later getting her GED. Influences on college-going. With the exception of the students who went straight to college from high school, the majority exited high school prematurely or delayed enrollment in college, automatically placing them in a group at hi gh-risk for departure (Hoachlander et al., 2003; Tinto, 1988, 1993). Furthermore, these negative experiences served to shape the students’ learner dispositions and academic and career possible selves as they headed to college. As demonstrated, college was not a priority for many because it was not emphasized in the home or did not serve a purpose translatable to the workforce greater than what they could do by working right after high school. Also, StUdents’ poor performance in high school made them question their ability to do college- 94 level work. With this learner perspective, most students at this stage approached their schooling with lack-luster effort and striving for mediocre grades; the goal was to get through school rather than to learn from school. In addition, this perspective and approach to learning influenced the development of college possible selves that were either non-existent or non-committal. The lack of clear and committed college/career possible selves that required college resulted in college-going that lacked resiliency and flexibility in light of Most of these students went to college because of the messages they received from parents and school that college was the expected next step. Jane planned on attending college right out of high school. Her father upheld education as an important experience, ‘Cause my Dad always—you know, I had 14 brothers and sisters, and he’s always talked to us about getting’ our education and not bein’ on welfare and all these kinda things like that. . .So that’s what I was plannin’ on doin’ too. . .And I seen my older brothers, they sittin’ up studying, and they’re in college and all this type of thing like that. Although she did not fulfill her original plans to immediately transition to postsecondary education, she encouraged both her children to attend college and is proud that her daughter has a master’s degree and her son is in nearly finished with school. Not only did the students receive messages from family, school, and society that led them to believe they “should be in college,” but they did not have a clear idea of what they would do otherwise. Despite the fact that most of them did not have a clear sense of why they wer e going to college or motivation to be there, it became the default. 95 At this early point in the stopout cycle, there are two categories of students whose high school experiences help to shape their college possible selves and their ongoing decisions about education: College Possible Selves (CPS) and Non-College Possible Selves (NPS). Students have multiple and intertwined reasons for leaving college College possible selves students ’ reasons for leaving college. Students gave many reasons for their decision to leave college ranging from family to financial to academic reasons, but a pattern emerged among the CPS students in the descriptions of their decision to leave. Students in this group tended to leave for unexpected reasons, such as pregnancy, illness, or unemployment. Autumn always had plans to go to college and reentered college many times only to leave again. She described the many events she encountered, “I became sick. My mother became sick. Got married, had two more kids, and it just didn’t seem to be in the cards. It seemed like every time I wanted to go to school something was going to hold me back.” These students also spoke a lot about competing priorities as a reason for leaving. Many of them, particularly women, left when they got married or had children. They took on new roles, competing possible selves, that were important to them and for which they had high expectations--to be a good wife, mother, or daughter. Lynn did well in high school and thought she would go to college, persisting without any breaks, but when she got married at 19 to her high-school sweetheart they decided to have children right away. She now has four children ranging from ages 11 to 17 She enrolled in various forms of post-secondary education six times over 10 years. Lynn described her dilemma, 96 I went with the intention of continuing each time I went. I think I just. . .well, it was just difficult on me because I had small children at home and a --we both (husband)—-we had two business that we were running, and so I was busy. But it was even harder on my family, and they generally didn’t take it really well, and so it just became priorities, I guess. . .you know, I was the one that chose to get married and have children, and they were my first responsibility. Others felt strongly that they “don’t want someone else raising my kids” (Autumn), not wanting to utilize childcare. Their other possible selves took priority over their college possible selves at that time. As mentioned before, students along this trajectory felt strongly about going to college. They did not report having a negative academic experience that drove them away, rather external events led them to make conscious decisions that prioritized other roles and possible selves over their college possible selves. Of the 48 participants, 12 of them demonstrated these experiences and attitudes. They number of times they cycled in and out of school ranged from two to six and occurred over a range of four to 15 years. Only two of these students were men. These students retumed to college a number of times throughout their years, hoping each return would “stick.” Non-college possible selves reasons for leaving. Conversely, students in the NPS group often left without plans or intentions to return to college. Reasons for leaving varied from financial concern to not “fitting in” at college. For men and women, financial reasons were a primary concern. Students emphasized the cost of school and the lack of eligibility for financial aid because they worked. The cost of leaving work to attend school was too high and the loans to costly. Chad left school because as his painting 97 business and potential earning power through the business grew, college became less of a priority. College did not serve a purpose because he could make more money growing his business. Marie echoed a similar sentiment, “I couldn’t afford to go. I mean, I was, like--I was workin’, I think, at a tanning salon at that time, and I was like ‘I’m not payin’ out of pocket to go,’ you know. . .and so I decided to take a break. I really didn’t think that I was ever gonna come back to school.” For many, the combination of being able to make “good money” at their current jobs, not seeing themselves as the college-type, and not being engaged in school made it easy for them to choose to leave. Marie left part-way through her first semester, I think it was just me still being young and not getting paid to go to a place every day. You know, and I wasn’t even full-time. You know, I just took a couple courses. . .But for me to drive there, and sit there, and learn this, and do this when I could be out somewhere else making money to survive on, then that was one thing. So I didn’t like the fact that I had to sit there and waste my time, ‘cause back then, to me, time was money. Bob spoke about how he saw himself as a student and the purpose of college, I just didn’t see me as being an academic-type, you know, somebody that would go to school. Just, I was comfortable at my job. . .I didn’t care for it (college) and I felt like I was going nowhere there. And when I got to where I was at, I was getting paid pretty good money. And when you got used to it, it was like for nothing. And why go to college and ruin that? Bob was happy with his job and income; his work did not require a college-level education and he did not identify as an “academic-type,” someone who belonged in 98 college. Therefore, there was little incentive to go to college or to promote a college possible self. These students’ conceptual schema for college was not well developed or defined. They often went to college because that is what they thought they should do or were being encouraged to pursue, but they lacked personal meaning or purpose for why they were attending. Their lack of commitment to college was also reflected in their learning disposition and academic strategies. Most students described their behavior in college as being similar to that in high school—an average to poor student who was bored in school and saw it as a waste of time. They generally did not have career possible selves that were clear or that required college, and were often content to work in minimum wage or factory jobs. Therefore, when something threatened their college-going, such as unexpected events, job prospects, or academic problems, they were happy to leave. As a result, there were few motivators or behaviors to prompt them to work out solutions or persist. As demonstrated, students reported multiple reasons for leaving, many of them occurring at the same time. The most salient of these were unexpected interruptions (often family) that took priority, economic and financial reasons, a lack of interest in college or seeing a purpose for it, and academic problems. These reasons for leaving are similar to findings fiom other studies that found both external (e.g., family responsibility, changes in job status, financial cost) and internal causes (e.g., lack of engagement, low academic self-efficacy, lack of clear goals/possible selves, cannot see purpose in college experience) (Ahson, Gentemann, & Phelps, 1998; Bonham & Luckie, 1993; Feldman, 1993; Horn, 1998; Hoyt & Winn, 2004). Also, while the reasons for leaving and attitude 99 toward school differed across groups, all students in the study experienced a misalignment between their possible selves, learning disposition, and academic expectations for success that resulted in a decision to leave school. Reasons students return to college All of the students in the study left college, but the amount of time in between enrollment and the number of times that students attempted to return varied. In addition, there was a range in students’ attitudes toward returning to college when they left. CPS students ’ reasons for return. The CPS students stated that when they left they had the intention of returning to college and fulfilling that goal. A small number of students with the intention of returning to college stopped-out between three and five times over at least ten years. This sub-group was unique because their primary reason for returning was their desire to accomplish their goals. Lynn discussed her decision to return despite concerns about returning to school once again at age 47, “It’s like, you know, Whether I’m 47 with a degree or 47 without a degree, the only difference is going to be the fact that I’ve accomplished that goal that I’ve set for my — I mean, I’ve wanted to do this since I was third grade.” Sally describes how important it is to her, “. . .’cause that’s my dream and I refuse to let go of it, no matter—even though I messed up so much. . .it was I want my degree because that — that - that will complete me in a way that no one can understand.” Their college possible selves were highly motivating and provided resilience that allowed the students to repeatedly return and try to accomplish their goals despite a lack of success in the past. Their academic success and the disposition toward learning during these trials varied across people and time. 100 Life changes. Among this group there were also many students who found their home lives and circumstances changing. For some this was a divorce or being widowed that left them without a means of income. Often relying on their husband’s income, these women found themselves without a college education and very little work experience as they retumed to the job market. Beth’s husband was the primary breadwinner, “. . .so there wasn’t necessarily a need for me to work, but then my husband got ill. . .and then he passed away, so now I’m just getting back to school.” Another common story among women was the effect that their children’s aging had on their decision to return to college. These women were primarily homemakers and did not return to their education because their priority was raising children and taking care of the family full-time. As their children entered school and matured, being able to understand why “mom wasn’t always available,” these students began to return to their own educational and career goals, often pursuing the college and career possible selves they had set aside while fulfilling other roles (as mothers and wives) that took priority. Lynn’s family was not always supportive of her attempts to return to college in the past, Well, I have two that are still at home, but one’s l9 and the other one’s 11, so on that level it’s a little easier. The ll-year-old’s still struggling with it a little, but she’s pretty independent, so I think we’ll get through that, and I think I’m a little better able to talk — because of the experience with the other three children, a little better able to explain to her why this is important and stuff than I probably was able to with my other children. And my husband’s much more behind me this time than he has been in the past. Finally, many of the women and men who have children emphasized their need to be a good role model for their children. They were planning for their children to go to 101 college and wanted to be a positive example. Sally explained, “But now it’s important - my grade point average is important to me, because my kids are watching and this is our future.” NPS students ’ reasons for return. Among the NPS group, one of the major external factors to influence their decisions was the exodus of manufacturing positions and jobs associated with the auto industry. As mentioned before, this area attracted many of the major auto and accompanying manufacturing companies. As a result, the city and area developed an economic dependence on this industry (Dandaneau, 1996). As deindustrialization spread, factories closed, and industry was moved out of the area or overseas, massive lay-offs of what were previously stable, well-paying jobs occurred. Many of the students in this study who had been at the same companies for 10 to 20 years suddenly found themselves facing unemployment and a potential change in career. These events prompted individuals who never thought that they would return to college, because they were not the “academic type” and did not enjoy school, to return, often with state or company funding for job retraining. Blake’s position in the auto industry was sent oversees and he was offered Trade Readjustrnent Assistance, “So I was like we’ll just see. I had a job opportunity to start at Delphi (another factory) back in the maintenance program making more money than I was making or go to school for free. Well, that was a decision because I didn’t like school the first time I tried it. . .and I was actually scared. . .So just talking it out and weighing all my options. . .I decided, all right, I’ll do it. . .and I haven’t regretted it yet.” Most of these students had matured and become spouses and parents since their last attempt at college and their return to college was linked to being a good spouse, parent, and provider. Tom always had a strong work ethic 102 and moved around to different work until his first child was born, he then began work at a factory where the position provided a good salary and security. At the time, he never considered college because there was no time or money. In 2006 when he was laid off he decided to take advantage of the Trade Retraining Allowance. Tom talks about the importance of doing well for his family, I guess I want to succeed, and I want my family to succeed, ‘cause I feel ifI don’t succeed as an individual, them my family suffers as a whole. . .It’s more than just financial too. There’s a whole psychological aspect of it also. You know, ‘What does your dad do?’ Oh, well, and the kids kind of clam up, you know. I don’t want my kids to be like that. I want my kids to be able to say, ‘Oh, my dad works with robots,’ or something like that. I want them to be proud of me. So I guess it’s a little bit of self-esteem and just wanting better for my family. As evidenced by the students who experienced changes in their personal lives and who had matured over time into new roles (e.g., spouses and parents), a combination of time and new events led to these roles as parents and spouses to serve as motivation to return to college. Whereas for some students the initial development of new roles in their lives caused conflict with their college possible selves and contributed to the decision to leave college, for many of these and other students, maturity into new roles contributed to their decision to retum. Students’ hoped for selves as good parents and providers aligned with their college possible selves. There was an integration and symbiosis between theses possible selves, I think I wanted my children to know, later in life, that it doesn’t really matter the kind of cards you’re dealt. It’s how you play them. . .And I don’t ever want my 103 kids to feel like they have anything to do with any kind of failure I might have, that they, by some chance, limited my life. Because if anything, without them, I may have been less determined to do well. . .Now I do have a reason to do well. I have a reason to make my life better, therefore making their lives better. . .So it’s always been something very big to go to college and have a college degree, so that my children can see it’s important. It’s important to stay in school and to learn always. (Mary Elisa) Development of perspective. Finally, among the NPS students, many referred to general life and work experiences that gave them perspective on what life would look like without a college education. Joseph described his work experience after leaving college, “I ended up working at Chuck E. Cheese for a year and decided—or seeing the people who worked there were either like ten years older than me or still in high school, I decided I didn’t want to end up being those people who are ten years older who’d been working there since they were my age.” For many of the younger students who left school because they believed it was a “waste of time” or were academically failing, being able to see their potential futures without a college education through other workers prompted them to readjust their possible selves and return to college. Upon return, college had more purpose and meaning; their conceptual schemas changed. Over time and through experiences many students gained clarity and focus about their academic and career goals. Many students saw college primarily for its utilitarian, career preparation purpose. Students who were unsure what career possible selves they wanted to pursue or even if they belonged in college, did not see college as an option until they gained more clarity about who and what they wanted educationally and as a 104 career. Therefore, as some of the participants found themselves developing more clarity about the career they wanted and the plan necessary to achieve that career, they also realized that they needed to return to college. Mother Over Everything spoke about her passion and its relationship to college, “. . .in order for me to be of service to individuals in this community. . .I need an education in order to do that. . .That is one of my life purposes now to make sure that I educate young people and/or senior citizens on the affects of AIDS.” While individually these events and life changes were enough to prompt a consideration to return to college, many of these events were happening simultaneously. For many students, during the time that passed while they were working in a factory or raising children or even in prison or rehabilitation, their life experiences were exposing them to potential career opportunities and trajectories that required a college education. For example, since Sally’s high school graduation in 1988, she has stopped-out approximately five times and in the last eight years struggled with addiction and abusive relationships. Sally’s experience in rehabilitation led her to study social work because she “want(s) to help the women with kids in treatment because it was so hard for me.” She now has aspirations for a master’s degree. Therefore, these life events prompted a consideration of what they wanted to do and what that would entail. For most, these contemplations and following clarity about career goals would confirm a need to return to college. Ability and motivation to persist through to their educational goals is marked by a change in learner disposition and possible selves. As students move through the stopout cycle, some multiple times, they will continue through one of three paths: 1) continue to 105 stopout—leave and return to college; 2) leave college permanently without completing educational goals; and 3) persist through to educational goals without leaving again. While the outcomes for many of the students in this study remain unknown because they have only recently returned to college, 44% of the students in the study are within one year of finishing with a certificate, associate’s degree, or transfer. Changes in possible selves and learner disposition The third theme discussed in this paper reports the reasons that students gave for retuming to college, demonstrating that most students experience a life change or event that prompts them to reconsider their career and educational possible selves, resulting in the need to return to college as part of the pursuit. This final theme raises the question-- now that the students have returned to college, what keeps them from leaving again? One of the more interesting findings from the study is the positive influence that a transformation and shift in learner disposition and college/career possible selves has on students’ academic behavior and ability to persist to their educational goals. I demonstrate that as students returned to college with renewed or changed possible selves, college and being a good student took on new meaning and importance. Students’ attitude toward learning, college, and study behaviors changed fiom prior experiences and attempts at college, as a result, students became more successful in college. This success reinforced their decision to return and their career and college possible selves. At this point, the two different paths converge. Change in possible selves. Most of the students were re-entering college having experienced a transformation in their career possible selves, who they could and wanted to be within a career, which also required them to reconsider their college possible selves, 106 if and who they wanted to be as a college student. Students were often re-entering with clearer career possible selves that excited and motivated them. In addition, this renewal of career possible selves reinvigorated students’ college possible selves. The return to college was a resurrection of a past identity and possible self that was not necessarily remarkable for some students, but others were intimidated and unsure about their ability to return. Many had tried to return to college before only to leave or never even enroll. For these students, the commitment and clarity to a career PS allowed them to cope with and push through the fear and anxiety of returning. Marie spoke about all the challenges since she returned to college, I work full—time right now, and I’m taking 16 credits. . .As of right now, I have a 2.7 GPA, which has been very difficult to bring back up. . .pretty much 3.5’s and 40’s is what I’ve received since. . .I’ve returned, and so, I mean, even with that, coming from a 1.67 GPA was really difficult. I mean, I had--I couldn’t get student loans. I couldn’t get any type of financial assistance. I had to pay for all of my classes out of pocket. . .Right now, I’m pretty dedicated. It’s taking me too long to get this far, but I’m very dedicated to get my degree, not just to get hired on by (Archer Community College), but to say I did complete my degree. Change in learner disposition. Furthermore, the students who were close to achieving their educational goals or maintained their enrollment for a longer period of time than previous enrollment, reported a change in their learner disposition--perspective of and approach to learning. As students re-entered with a stronger commitment to their career and college possible selves than previous college experiences, they also reported changing their learner disposition in order to enact academic behaviors that would allow 107 them to achieve their goals and possible selves, “I know I’m doing what I need to do and I’m doing the best of my ability. . .paying attention to the teacher and completing my work on time. Before I would wait until the last second, but now I’m getting it done right.” (Blake). One shift in learner disposition was not a change in students’ beliefs about the purpose of postsecondary education as career preparation, but in their perception of and meaning they ascribed to the college experience. With more focused career and educational goals, college became the path to a better, more stable life, and to a better self. In addition to becoming more personal significant, for many students it was the path toward changing their financial situation and sense of stability, 1 really didn’t like working the jobs that you could get right out of high school. There are a lot of jobs where the people above you will just treat you like crap because that’s how they were treated when they were in that position, and it’s just not a pleasant experience, and I wanted something where I was actually in demand rather than I can be replaced—work someplace that doesn’t have a giant turnover rate, and I started thinking about going back. . .And I wanted something where I felt good about what I was doing (Joseph). Bob recognized that in order to get a better job that would pay more, he needed to get a “jurnpstart” on his peers and returning to college was that jumpstart. In addition, Bob’s college PS changed as he demonstrated to himself that he could do the work and do it well. His college PS was reaffirm and encouraged. Alignment in reasons for attending college and perceived purpose of college Students’ belief about the purpose of postsecondary education was now in sync with what they were seeking from and the meaning they ascribed college. Bloomer and 108 Hodkinson (2000) also found that change in dispositions to learning were both prompted by change students’ context and internal beliefs and values, but that these changes also affected their meaning and perceptions of the context itself. Additionally, as career and college possible selves became more solidified and gained more importance for students’ lives, their approaches to leaming and academic behaviors also changed. The majority of students reported a major shift in how they viewed learning and their strategies for learning. Most described themselves in previous years as not taking school seriously, having priorities in other interests, not studying well or at all, not caring about learning the material, and being more concerned with the social aspect of college life. They returned to college as different students. They want to learn and become proficient at their chosen career, their study strategies have changed, their academic goals are higher, they are more realistic about what is achievable and are much more willing to ask for help from faculty, staff, and students. Bob talked about the changes he has experienced, Just as you get older you take things more seriously. And on top of that, just my study habits, I got to where I was doing pretty good with that. And when I started getting good grades, I was like ‘You know, you can do this stuff”. . .’cause I look back at my high school like ‘You didn’t try, you were an underachiever, and now you’re doing so good. . .(coming back to school) I could’ve went for another job, and I didn’t do it because I would’ve made less, I would’ve been stuck there the rest of my life. I thought, ‘Go to school. Get an education. Get the jumpstart ahead of these people who don’t do it.’ And that’s what made me do it. 109 Students’ schema for how to be successful in college is now more important, realistic, and in line with what is necessary for academic success and persistence. Research on self-schemas found that procedural schemas for how to become a college student were necessary, but conceptual schemas for the understanding the meaning and purpose of becoming a college student were just as important (Pizzolato, 2006). The students from this study not only developed more appropriate procedural self- schema for how to be a good college student and achieving their college and career possible selves, but also developed career and educational goals that were in line with their schema for the purpose and meaning of college, elevating the importance of those schema in their college success. Academic persistence and success. As a result of changes to their learning disposition, most students found more success within the classroom, felt that they belonged in college, and increased their academic self-efficacy and belief that they could “do” college-level work. As a result, the positive experiences reconfirmed their decision to return and pursue their career and college possible selves, which confirmed the learning disposition they adopted upon return to college. Shirley described this experience in school, “I’ve been getting good grades so that kind of pushes me to, you know, keep going and keep trying harder because I don’t want to. . .see my GPA drop or, you know, anything like that. So for what I’ve been doing for as good as I’ve been doing, I think it’s giving me incentive to keep trying and keep working hard.” While likely not the only factor in persistence, the students’ clarity in college/career PSs and changes in learner disposition enabled them to persist by engendering alignment between students’ educational goals and their beliefs about the 110 purpose of college, providing positive feedback about their ability and belonging as college students, and confirming their commitment to their college and career PSs. All of these positive outcomes contribute to their motivation to persist despite hardship and challenges. Summary The reflective nature of this study allowed for the variation in individual education paths that include stopout to take shape and presented in a model that included one’s high school academic experience, entrance into college, cycle of stopping-out and reentering college, and exit options. A depiction of the student paths as they move through these education events is Figure 2. Figure 2: Student Paths and Educational Events F HIGH SCHOOL L K EXIT COLLEGE oCPS: Students have a clear. deveIOped CS? 8: had a generally positive HS experience; went straight to college; oNPS: Students have more negative educational experiences; CPS is undeveloped 8t unclear; college-going has little purpose or mean' often left [-15 early and/or took time off before college K.__.____/ k K RETURN TO COLLEGE \ ~CPS: Students leave primarily due to external events-~illness, pregnancy marriage. Did not have negative college experience. Other competing PSs were prioritiezed. oNPS: Left due to poor grades & work opportunities; generally did not see purpose in college 8: didn't 'fit in.” \_____/ oCPS: Life goal In get degree; change in home life (e.g., divorce. children go to school): competing PSs now converge. oNPS: Change in home/work life cause students to reevaluate career PSs, goals now require college; college now has purpose] meaning. ¥____/ F OUTCOMES: PERSIST OR DROPOUT oCPS & NPS MERGE: Student now have developed college/career PS w/ meaningdearner disposition also shifts-want to learn 8: do well; engaged in school 8: change in study behaviors; as students are more successful they become more motivated and committed. reinforcing their college/career selves. and leading to more success in school. \______.J 111 I discussed the internal and external factors that influenced decision making along those paths and the role that career and college possible selves and learner disposition play in more detail throughout the paper. Students left high school with college/career possible selves and learner dispositions based on those experiences. As a result, many students entered college with a clear and well-developed college possible self. Others did not have a clear conception of what or who they hoped to be and their motivation waned. Furthermore, many students exited high school with low academic self-efficacy and a negative view of education, questioning whether they belonged there. These high school experiences shaped students’ perception of academics, college, and learning that resulted in two distinct patterns as students journeyed through the stopout cycle. Factors that influenced students’ decisions to leave and return to college fell into similar arenas. Externally, the economic environment and financial prospects heavily influenced all students; family responsibilities were primary, especially among women; and their academic success or failure were major themes. Internally, students’ beliefs about their academic ability and belonging in college, their level of commitment to and clarity of career goals, the efficacy of their learner disposition, and the alignment between their perception of college purpose and their own career/college goals. Students’ decisions to return to college were marked by significant events, life changes, and better timing that prompted individuals to reconsider and clarify their career and college goals, bringing them back to college with a stronger motivation and commitment because of their focus on their career paths. Finally, not only did a shift in career and college possible selves prompt students’ return to college, but it influenced 112 their ability persist in college. As students returned with a clearer commitment and focus on their career and college goals, their learner disposition also changed in order to align their approaches to learning with behaviors that would result in academic success. Positive feedback from the academic environment as a result of changes to their learning dispositions reconfirmed their commitment to their college and career possible selves and the schema they were now operating off of to achieve those goals. While these patterns and themes assist in understanding the stopout phenomenon in relation to persistence and retention, individual student paths are multi-layered and reflect intersections across reasons and motivation for stopping and starting attendance. Many of the factors and changes discussed happened simultaneously. Also, the range in age among participants demonstrates how these events, changes, and development occurs at different times of life for different people. hnplications for Theory, Practice and Further Research In this study, I built on existing research focused on stopout behavior in relation to retention (Bean & Eaton, 2000; Bonham & Luckie, 2003; Cofer & Somers, 2001; DesJardins et al., 1994, 1999; Hoachlander et al., 2003; Horn, 1996, 1998), possible selves (Leondari, 2007; Markus & Nurius, 1986; Oyserman, Bybee, Terry, Hart-Johnson, 2004; Oyserman, Bybee, Terry, 2006; Pizzolato, 2006; Plimmer & Schmidt, 2007), and learner disposition (Bloomer & Hodkinson, 1997, 1999, 2000). The findings from this study extend both the theoretical understanding of decision making around stopout behaviors and the factors that influence those decisions, as well as the ways that these findings can help institutions and practitioners assist students’ return to college after an absence, and to maintain enrollment. 113 Theoretically, the model presented here provides a greater picture of the educational paths that include stopout. Most studies of stopout have focused on the reasons students leave (Bonharn & Luckie, 2003; Cofer & Somers, 2001; DesJardins et al., 1994); this study extends that inquiry to how their reasons for leaving relate to potential return and maintenance of enrollment. In addition, understanding how external and internal factors affect decision-making along each step of the path give researchers and practitioners the opportunity to consider the best places along that pathway for intervention. Finally, findings that highlight the affect of development and change of college and career possible selves and learner disposition in the decision-making and persistence of students raise two issues. First, when students’ beliefs about the purpose of college and their own career and educational goals and possible selves are not aligned, students are more likely to leave college. Reciprocally, when they are aligned, not only are students more likely to maintain motivation, but also to adjust their approach to learning to enact attitudes and behaviors that are more likely to yield successful outcomes. Second, these findings underscore the importance of student’s psychosocial and cognitive factors in the decision to return and persist in college. Many students expressed that they continue to experience challenges to maintaining enrollment in college (e. g., financial, balancing responsibilities, unexpected events), yet this time they are more committed to achieving their educational goals (and many are very close to achieving them) than during past educational experiences. In other words, students’ ability to make decisions and enact behaviors that result in greater academic success and confirmation of their goals was a reflection of the change in clarity and commitment to possible selves and learner dispositions. ll4 Practically, this study has multiple implications for community college educators and institutions. First, the open access nature and multiple missions of community colleges create an educational environment that has students enrolling and leaving the institution with relative ease. Many students attend for lifelong learning education and skill development, rather than an educational certificate, degree, or transfer. As a result, it can be difficult to identify students who leave without completing their educational goals. Associated with stopout, students often attend different institutions, sometimes at the same time, unbeknownst to the institution. As a result, unless students attended the same institution or submit transcripts from another institution, or the institution collects demographic information about previous college attendance, it is very difficult to identify students who stopped-out. Presently, approximately 40 states have some form of enrollment tracking system of college students with basic demographic and institutional information (Ewell & Boeke, 2007), but the structure of these systems differ across states and generally do not share information. Although there is discussion of a national tracking system organized at the federal government level, concerns about scope, cost, and individual privacy serve as arguments against it (Hearn, McLendon, & Mokher, 2008) For both of these reasons, I recommend institutions devise a way of identifying these students that is not associated with their transcript. While it would be challenging to attempt to connect with each student who is entering or leaving, attempting to connect personally with students at critical points, such as advising or financial aid, with the intention of exploring why students are leaving or returning can help practitioners assist students in making educational plans and support in an attempt to keep them from leaving 115 SElK'ei re; ‘CI n‘ and maintain enrollment. Many students in the study emphasized the importance of face- to-face interaction and relative ease in identifying and accessing services as an important source of support and encouragement during their re-entry. Furthermore, the importance of possible selves in decision-making along college- going paths underscores the need to integrate this flame into program and policy development regarding student persistence and retention. As institutions that serve students with varied purposes and goals for attendance, adopting a possible selves at the leadership and administrative levels of community colleges is a natural alignment for considering how to individual serve and support student goal attainment. Second, misalignment between students’ view of the purpose of college and their own college and career goals was found to be a major issue in the decision to leave, return, and maintain enrollment in college. The challenge becomes how to address this disconnect with students. Again, I suggest a multi-pronged approach where conversations can be had with students. Since community college students’ goals are varied and, as demonstrated, there are multiple paths through college an effort to individually consider student needs is required. Critical areas of campus, such as financial aid and advising, where students can engage with a staff member one-on—one and in person could be a target area. Also, many institutions are developing student success courses. This topic would generally fit well into such a course. Third, based on the finding that changes in student college and career possible selves influenced students’ learner dispositions and academic success, I would recommend that institutions and researchers focus on the intersection between college PSs and learner disposition. Developing a better understanding of how students’ 116 CITECS’I‘UOI ‘ -r 1 , so ‘ fibudl¥ . a- 13"! :hh‘nlhbb" ,‘3H"'jfw 3111‘ “|\ 763d IIl V‘ r in.” ..r :1 ‘ Z- ‘i ~o lac-5A“ IH conceptions of their college PSs and motivation for attending college influences academic performance can be utilized to assist students in remaining focused and academically motivated. Furthermore, these findings suggest that faculty should work to help students understand why they are going to college and why those reasons are important for their academic work. Creating a tighter coupling between students’ motivation for attending college and doing well in school in students’ own minds could result in a stronger academic performance and college persistence. Future research. In light of the important role that learner disposition and possible selves played in students’ decision to return and persist, understanding what promotes change in these constructs would develop further understanding of stopout. Furthermore, understanding stopout and how to break its cycle is critical to promoting persistence among students, because alignment between students’ perceptions of institutional purpose and their own goals were of such importance. One option is to use large, longitudinal surveys that track enrollment and stopout patterns. Such data is available through the National Center for Education Statistic’s Beginning Postsecondary Surveys. The drawback to these surveys is the lack of variables that can measure developmental and psychological constructs. Therefore, an alternative would be to conduct survey work that specifically collected information on enrollment, stopout, possible selves, and learner disposition. Furthermore, I recommend continued exploration of stopout in relationship to different institutional types and characteristics. Student perceptions of different institutions and their characteristics or purpose may influence which factors are more salient and decision-making processes around stopout. Finally, 117 ...n .\. lfll this was a qualitative study at one institution. Testing the model quantitatively may give us a fuller understanding of the paths students take and the relationships between factors. 118 .,',_ ., ' l . dig-.111. 2655533 5.11.. Ll if in c: ’ wt \ Ma du~¥§L 55 33‘ Utah 3 [fliers his in; 3135(1ng Rhliol. E 33'th p _~r-“" WI 0 . Jester 751311365 ”5mm l0 CHAPTER 5: ARTICLE 3— MIXED METHODS Introduction In President Obama’s first address to a joint session of congress, he challenged the nation to increase focus on education and raise the proportion of college graduates to the highest in the world by 2020 (2009). The Lumina Foundation responded to this challenge, asserting that the critical need for increased higher educational attainment is a necessary priority for US. economic development and stability in the short— and long- terrn. Lumina presented its own educational goal for the US. to improve the percentage of Americans with a degree or credential flom 39% to 60% by 2025. Two of the suggestions provided to support this goal include (a) aligning postsecondary education with the workforce development system and (b) expanding adult learning opportunities. Central to these suggestions was the call to strengthen community colleges as institutions most likely to provide workforce development and serve adult learners. Also, Lumina advocated that adult students who never attended or finished college and dislocated workers who needed workforce retraining and skill development were the populations to draw from for new students (Lumina Foundation, 2009). Reaching these goals requires increasing the number of students who attend and finish college directly out of high school, as well as drawing individuals back to college who never finished or attend school previously. This article focuses on the experiences of students who returned to college after a period of nonenrollment. As the job market experiences a shrinking manufacturing field and increasingly requires employees to have technical and knowledge-based skill, individuals are likely to return to postsecondary education for the development of job skills and retraining (Bragg, 119 r“ .3“ fir’ 1 §laL ll .‘ _" ~ “4112:? 3135. _ a Laser: ”lift! 1 .JJ~~ 2001). Community colleges already provide a significant proportion of skill trade and job training education, which suggests that many returning adult learners will return to two- year colleges for postsecondary education (Bragg, 2001). The increasing cost of college, tightening economic realities for families, and open access character of community colleges make this type of institution a more likely option for adults who are returning to college and therefore an authentic setting to study returning students. As returning adults, these students are likely to have personal and work lives that can pose a challenge to their educational goals (Ahson, Gentemann, & Phelps, 1998; Bonham & Luckie, 1993; Hoyt & Winn, 2004). During their time out of school it is reasonable to believe that they established work and home lives that require time and effort to maintain. Returning students’ non-academic life demands often translate into challenges to balance work, family, and acaderrrics resulting in student needs that are different flom more traditional students (Horn, 1996; Horn & Premo, 1993: Polinsky, 2002/2003). Community colleges already serve a greater proportion of students who are at higher risk for departure (Horn, 1996; Stratton, O’Toole, & Wetzel, 2004) and as a result they are poised to serve the increasing number of students that US. leaders hope to attract back to postsecondary education. Therefore, it is critical for community college faculty, staff, and administrators to understand the educational paths of higher risk and returning adult students in order to best serve and support them. Previous research has explored attrition and retention (Bryant, 2001; Cofer & Somers, 2004; Hawley & Harris, 2005/2006; Horn, 1996; Napolli & Wortman, 1996, 1998; Polinsky, 2002/2003; Romano, 1995; Strauss & Volkwein, 2004; Summers, 2003), 120 but there is little insight into the phenomenon of students who leave and return to college after a period of nonenrollment, also known as stopout. This dearth of research is particularly true for community colleges, yet a majority of students who stopout leave flom and return to two-year institutions (Horn, 1998). This study aimed to examine the following research questions using a mixed methods research design: Why do students return after a period of nonenrollment? a) Which factors are related to who returns after a period of nonenrollment? b) What is the influence of these factors on students’ decisions to return to college? b) How do these factors affect educational decision making? The quantitative part of the study analyzed the relationship between characteristics associated with higher risk for departure and returning to college as an outcome. The qualitative part not only extended on the statistical portion by exploring which factors were more influential in students’ educational decisions, but also how these factors affected and shaped the students’ decisions. In this article relevant literature is reviewed and followed by a presentation of the methods, design, analysis and findings of the individual quantitative and qualitative parts of the study. The discussion section is an analysis of the intersection between the findings flom the quantitative and qualitative studies. Finally, implications and suggestions for research and practice are presented. 121 Background Community college students A focus on attracting individuals returning to postsecondary education, particularly the two-year sector, requires a foundational understanding of students at community colleges. In 2006-2007 approximately 65% of students attended a four-year college and 35% of students in higher education attended a two-year institution (Provasnik & Planty, 2008). Examination of the characteristics of students attending two- year colleges paints a picture of a diverse population. Forty percent of community college students are dependents (i.e., under 24 and financially dependent), while the remaining 60% are financially independent--26% are older than 24, 20% are married with children, and 15% are single parents. In general the median age of a community college student (24) is higher than those at four-year colleges (21). Community colleges also have higher percentages of low-income students (26% versus 20% at four-year colleges) and Black (15% versus 10%) and Hispanic students (14% versus 9%) (Horn & Nevill, 2006; Provasnik & Planty, 2008). Community college students’ academic goals also vary; 36% report intentions to transfer, 43% seek an associate’s degree, 17% enroll for a certificate, and 42% report returning for job skills (not mutually exclusive categories) (Provasnik & Planty, 2008). Furthermore, 54% of students who enter a community college have an increased risk of departure (Hoachlander et al., 2003) and lack academic preparation in reading and math (Borglum & Kubala, 2000; Nora, Attinasi, & Matonak, 1990; Summers, 2003). 122 Persistence and stopout In addition to having higher levels of diversity across student background characteristics, community colleges also have higher percentages of departure (45%) than four-year colleges (17%) (Horn, 1998; Provasnik & Planty, 2008). Approximately 30% of students in all of higher education leave after the first year--63% are flom community colleges, 22% are flom four-year institutions, and 15% are flom other types of schools (e.g., private, for-profit; public less-than-two year; or private, not-for-profit, less than four-year instiutions). Specifically, among students who attend four-year institutions 16% depart, while 42% of students enrolled at two-year schools leave. Of those who leave, over half (57%) will return to higher education within five years. (Horn, 1998). Yet, 50% of those who depart flom the two-year sector return, while 64% of those who leave a four-year institution return within five years. This gap in returns indicates that students who begin at a two-year institution and leave after their first year of school are less likely to return to college than students who begin at four-year institutions, but represent a greater number of students overall. In other words, out of 100 students who leave during their first year 85 of them are flom two- and four-year institutions. Of these 85 students, 46 will return within five years: 32 of those who return originate flom a community college versus the 14 who begin at a four-year school. Furthermore, of the students who leave during their first-year and return within five years, a majority will return to community colleges with 79% of students who start at a two-year and 24% of students who start at a four-year institution returning to community colleges (this includes students who return to the same institution at where they began and students who transfer flom another two- or four-year institution) (Horn, 123 1998). For example, out of 100 stopout students who begin at a two-year institution, 79 of them will return to a community college and the remaining 21 will attend four-year or other types of institutions. Additionally, out of 100 stopout students who begin at a four- year institution, 24 of them will return to a community college and the remaining 76 will attend four-year or other types of institutions. Recent calls to increase graduation rates and encourage individuals to return to postsecondary education highlight the issue of stopout in the study of attrition and retention. Understanding factors associated with which students stopout or dropout can contribute to developing strategies for how to prevent students from leaving and creating pathways for student return to college. Risk factors related to student departure The community college student population is more diverse compared to students at four-year institutions-~they are older, more racially and ethnically diverse, and less academically prepared. Many of the characteristics that contribute to the diversity at two- year institutions are also associated with traits linked to students at higher risk of departure. In a report profiling undergraduates, Horn and Premo (1993) identified seven characteristics commonly associated with a higher risk of non-persistence and nontraditional students. The traits were representative of three groups: 1) Nontraditional enrollment choices (e. g., delayed enrollment or attending part-time), 2) financial and family status (e. g., having dependents, working full-time, financial independence, or being a single parent), and 3) high school graduation status (e. g., receiving a diploma or GED.) (Horn, 1996; Horn & Premo, 1993). Each of these factors were linked to a 124 potential increase in risk of attrition, therefore the number of characteristics students had was also an indication of risk levels (i.e., minimally, moderately, and highly at risk). Horn and Premo (1993) concurrently labels these risk factors as characteristics associated with non-traditional students. This reveals a bias toward a four-year institution perspective. Students with these characteristics are more likely to attend two-year institutions (Horn, 1996), therefore while community college students are more likely to be at higher risk of departure compared to students at four-year institutions they are very traditional for a two-year institutional context. The current study draws on the risk factors identified by Horn & Premo for independent variables, but since the sample is focused solely on students who began at community colleges I choose to flame these factors only as indicators of risk versus levels of traditionalness. Additional research point to other student characteristics associated with being at higher risk for departure and lack of persistence toward educational goals. Previous studies report that while traditional-aged students out-paced nontraditional —aged students in enrollment growth between 1995 and 2006, this trend will likely reverse between 2006 and 2017 with nontraditional-aged students enrollment growing nearly twice as fast as enrollment for traditional-aged students (Snyder, Dillow, & Hoffman, 2009). In addition, students with more risk factors are concentrated at two-year institutions, while students with no risk factors are concentrated at four-year institutions (Horn, 1996). Also, higher risk students are less likely to earn a degree within five years of beginning college and more likely to leave without returning than traditional students. Thirty-one percent of higher risk students obtain a bachelor’s degree within five years compared to 54% of 125 lower risk students, also higher risk students are twice as likely to leave in their first year (38%) than their lower risk counterparts (16%) (Horn, 1996). There is ample evidence that higher risk student status is associated with negative outcomes in higher education, such as early departure and lower rates of degree attainment (Horn, 1996; Horn & Premo, 1993; Hoachlander et al., 2003). Findings also indicate that the number of older and higher risk students on campus is growing and is concentrated at two-year institutions. With the appeal by President Obama and the Lumina Foundation to increase the number of students receiving diplomas and returning to education, it is likely that the higher risk student population will continue to grow. Assuming that a significant number of students who return to college also have higher risk student traits, gaining a better understanding of the association between risk factors for attrition, who returns to college after nonenrollment, and why they return is important. Following is a detailed discussion of the study’s design and methods. Research Design This study utilized a mixed methods, explanatory design approach. A mixed methods study involves the collection or analysis of both quantitative and/or qualitative data in a single study in which the data are collected concurrently or sequentially, are given a priority, and involve integration of the data at one or more stages in the process of research. (Creswell, Plano Clark, Gutrnann, & Hanson, 2003, pg. 212 ). Mixed methods research can provide a better understanding of the problem than either method, quantitative or qualitative, would yield alone (Creswell & Plano Clark, 2007). 126 The purpose of this study was to explore stopout behavior and which factors and processes are potentially related to and influence the decision to reenroll after an absence flom postsecondary education. The quantitative and qualitative parts of this study were designed to examine different aspects of stopout related to the overarching research question, “Why do students return after a period of non-enrollment?” Part A was a quantitative study of risk factors associated with the odds of returning to college or staying out, while Part B was a qualitative inquiry into the educational paths and decision making processes of community college students who stopped out. The qualitative part was a purposeful expansion on the findings flom the quantitative study. While the quantitative study indicated which risk factors were significant, I used the qualitative study to more deeply investigate how such factors influenced students’ educational decisions and persistence. The design, methods, and findings of the quantitative and qualitative studies are reported in separate sections below. Part A: Quantitative Methods The quantitative portion of the overall study was designed to address the following research questions: (a) which factors are related to who returns after a period of non-enrollment? And (b) what is the influence of these factors on stopout? An analysis of which risk factors for attrition were related to which community college students return after non-enrollment provides a basis for understanding how students’ background characteristics (e.g., gender, race and ethnicity, socioeconomic status), academic choices (e. g., high school diploma or GED, enrollment intensity, reasons for enrolling), and life 127 choices (e.g., marital status, dependency status, parental status) may have an influence on which students make the decision to return to college. This study is a secondary analysis of data flom the Beginning Postsecondary Students Survey (BPS): 96/01. Data and sample The Beginning Postsecondary Students Survey (BPS) is a longitudinal extension of the National Postsecondary Student Aid Study (N PSAS) conducted by the US. Department of Education National Center for Education Statistics (N CES). This nationally representative survey includes a cohort of students who began their postsecondary education during the first year of the NPSAS study (1995-1996) with two follow-up surveys administered in 1998 and 2001. The survey was designed to collect detailed enrolhnent information, including information about enrollment and unenrollment spells, changing of institutions, and exiting postsecondary education without return during the length of the survey (Wine et al., 2002). Students were selected for the sample for the quantitative part of the study if they began their postsecondary education at a public two-year institution and either stopped- out or stayed-out during the length of the study. Stopout was defined for BPS as “a break in enrollment of five or more consecutive months. It begins on the first month of the break. A stopout is also the months between two enrollment spells” (Wine et al., pg. ,2002). Students were considered a stopout if they indicated that they stopped-out one or more times (PRSTNU2B). Dropout was defined as students who left postsecondary education without return and without attainment of a degree between fall of 1995 and June 2001. While students who left without return during this period of time may return at a later date, for the purposes of this study they were included as dropouts. 128 The resulting sample size for the study was 779 observations. Thirty-five percent (272) of students in the sample stopped-out, while 65% (507) stayed-out. In Table 13 (in Appendix A) are flequencies for descriptive characteristics of the sample. Approximately 53% were female and 71% were White, 12.6% Black, 12% Hispanic, and 5% Asian/Indian/Other. Sixty-one percent of student were at the least minimally disadvantaged. Finally, the level of risk of attrition was distributed relatively evenly across the sample—39.6% minimal risk, 32.7% moderate risk, and 27.7% high risk. Analysis BPS utilized a stratified multistate cluster complex sampling design. The strata represented different sectors of higher education (institution-level based) and clusters represented geographic regions. The public two-year sector was included in this sample. The complex sampling design used for BPS 96/01 (and most large-scale surveys) presented challenges that needed to be addressed for analysis. First, the weights constructed for BPS 96/01 address both subsampling and nonresponse bias within the survey. The sampling weight for longitudinal analysis of BPS 96/01 (B01LWT1) was used to construct a relative weight to preserve the accuracy of the weighted estimates by reflecting the sample size (Thomas & Heck, 2001). To avoid removal of cases with normalized weight values of zero in analysis, a constant of .001 was added to the weights. This allowed for the cases to remain in the sample for analysis with a negligible affect (de Vaus, 2004). Second, the use of clustering in the latter stages of large-scale surveys contributes to the nonindependence of observations and nonhomogeneity, violating the assumptions of simple random sampling (Diemer, 2008; Thomas & Heck, 2001). In recognition of potential estimate bias as a result of a simple random sampling 129 assumption, a more conservative alpha level was used to evaluate the model and predictor variables (p<.01) (Thomas & Heck, 2001). Direct logistic regression analysis was conducted to observe the effects of factors associated with hi gh-risk for departure on student return to postsecondary education after a period of nonenrollment. SPSS uses the Omnibus Tests of Model Coefficients chi- square statistics to determine if the model provides an overall fit to data by providing improvement over the intercept—only model. Additionally, an examination of changes in the percentages of correct classifications between the null and final models is used as a validation of predicted probabilities. Finally, goodness-of-fit of the model is presented with two different statistics. The Hosmer and Lemeshow statistic reports a chi-square test for goodness-of-fit that is more conservative and sensitive to the ways in which predicted probabilities are grouped (Hosmer & Lemeshow, 1989; Peng et al., 2008). When interpreting the significance it is important to note that alpha levels above .05 are considered an indicator of better fit. Also, since a traditional R square cannot be determined using logistic regression, the Cox and Snell R Square and Nagelkerke R Square are pseudo-R square statistics used to estimate the percentage of the amount of variation in the dependent variable explained by the model. Dependent variable The dependent variable is a constructed binary variable (stop_stay2) that represents students who began at a public two-year institution and either left without return (dropout) or left with return (stopout). Stopout is the outcome of interest. 130 Table 11 Dependent Variable Stop_Stay2 Students who began at a two-year institution that: (a) left college and returned during the course of the study or b) left without return to college prior to June 2001. Independent variables The primary variables of interest within this study reflect characteristics that literature has shown to be related to a hi gher-risk for departure, particularly for non- traditional students (Hom, 1996; Horn & Premo, 1993). The seven characteristics identified in literature as related to a higher risk of attrition are delayed college enrollment, receiving a GED/certificate, single parenthood, being married or divorced/widowed/separated, financially independent, have dependents, working fulltime while enrolled, and enrollment in school less than fulltime (Horn & Carroll, 1997). These characteristics were also used in BPS 96/01 to construct the risk index for departure (Wine et al., 2002), therefore this index is a measure of both the level of nontraditionalness of the student and the risk of attrition. These variables were individually included in this study as potential indicators of increased risk for departure. Four out of the seven variables were retained in the final model as significant or important for the model’s overall goodness-of—fit. The variables included in the final model were dependency status (DEPENDEN), marital status (MARRYDUM/DIVORDUM), high school diploma or GED/Certificate (HSDIPL02), and enrollment intensity (PARTDUM/MIXDUM). Those variables were either recoded into a binary format or a dummy variable was created for each value/category. 131 An additional two variables that literature has shown to potentially affect attrition were also included. Previous research has shown that poor G.P.A. was a common reason given for leaving college, especially for traditional students (Barr, 2007; Burley, Butler, & Cejda, 2001; Hawley & Hanis, 2005/2006; Mohammadi, 1994; Ogletree, 1992; Sommers, 2003). Another variable included was students’ “reasons for enrolling” (SBENRPUR). There is some evidence that students’ goals entering into college is related to student departure (Choy, 2001; Polinsky, 2002/2003). Additionally, the larger mixed methods project that this study is a part of focuses on how students’ purpose and goals influence their college going decision-making. Gender, race and ethnicity, and socioeconomic status were controlled for in the model. Race and socioeconomic status were represented by dummy variables. Socioeconomic status was represented by a disadvantage index variable (DISADVAN). This index was constructed for BPS flom thee other variables—parents’ highest education, percent of poverty level, and economic status of high school student body. Three dummy variables were constructed and used in the model—NDISADVAN (no disadvantage), MDISADVAN (minimal disadvantage), HDISADVAN (moderate/high disadvantage). Results Direct logistic regression was performed to assess the impact of a number of factors on a likelihood that respondents would return to college after a period of nonenrollment. The model contained 16 variables (see Table 13 in Appendix A for variables). 132 The full model was statistically significant with a chi-square critical value of 79.054 (df=l6, n=779, p=.000), indicating that the model was able to distinguish between students who reported stopping—out and staying-out. Two pseudo R squared statistics were reported; while not identical to R in OLS regression, they indicate how much inclusion of the independent variables in the model reduces the variation (Menard, 2000; Peng et al., 2008). The model as a whole explained 11% (Cox and Snell R Square) and 15% (N agelkerke R Square) of the variation in stopout. While not included in SPSS output, I also calculated McFadden’s index. I reported here because it was “preferred over others for its conceptual similarity to the OLS coefficient of determination, its relative independence flom the base rate, and its comparability across models comprised of different predictors yet applied to the same outcome variable and the same data” (Peng et al., 2008, p.10). McFadden’s R squared was .088; values .2 to .4 are considered highly satisfactory (Cameron & Windmeijer, 1997). The percentage of correctly classified cases increased flom 62.9% to 63.9% after the introduction of the independent variables to the constant—only model. Finally, the Hosmer and Lemeshow test for goodness-of—fit was significant at 12.916 (df=8, n=779, p>.05). . Within the model, four predictors were found to provide a statistically significant contribution at p<.01—minimally disadvantaged, divorce, mixed enrollment, and jobskill as reason for enrollment. The strongest predictor of returning to school was mixed enrolhnent, with an odds ratio of 1.759. This indicates that students who had both part- and full-time enrollment were 1.76 times more likely to return to college than those who were only enrolled part- or full-time; in other words, the odds of returning to college for students with mixed enrollment was approximately 76% higher than for students only 133 enrolled part- or filll-time. Divorce was also a stronger predictor, but of who was less likely to return. The odds ratio for divorce was .23 0, meaning that students who were divorced/separated/widowed were 77% less likely to return to college than those who were single or married. Students who went to college to develop job skills were also less likely to return than students who went to college for personal enrichment, associate’s degree, certificate, or transfer to a four-year college. The odds that a student who went to college for job skills will return to school after leaving was .502, or 49.8% less than students who went to college for the previously listed reasons. Finally, students who were minimally disadvantaged had decreased odds of returning compared to moderately/highly or not disadvantaged students, they were .606 or 39.4% less likely. Yet, when the odds of minimally disadvantaged students are individually compared to moderately/highly disadvantaged and not disadvantaged students, different odds ratios emerge. When the odds ratio for minimally and not disadvantaged students was calculated, minimally disadvantaged students were 34.6% (.654) less likely to return to college. Conversely, when minimally and moderately/highly disadvantaged students’ odds were compared, minimally disadvantaged students were 4.17 times more likely to return to college than moderately/highly disadvantaged students. 134 Table 12 Logistic Regression Predicting Likelihood of Returning to College After Nonenrollment Variables in the Equation 95% C.I.for EXP(B) B S.E. Wald df Sig. Exp(B) Lower Upper Step 1“ GNDERCOD .054 .178 .092 1 .76 1.055 .745 1.495 BLACKDUM -.558 .287 3.773 1 .05* .572 .326 1.005 HISPADUM -.354 .286 1.537 1 .21 .702 .401 1.228 ASIANDUM -.117 .396 .087 1 .76 .889 .409 1.934 MDISADVAN -.501 .197 6.454 1 .01** .606 .412 .892 HDISADVAN -.295 .293 1.013 1 .31 .744 .419 1.322 DEPENDEN .601 .272 4.8801 .02* 1.824 1.070 3.108 HSDIPL02 -.364 .274 1.764 1 .18 .695 .406 1.189 MARRYDUM -.676 .312 4.674 1 .03* .509 .276 .939 DIVORDUM -l.469 .506 8.423 1 .004** .230 .085 .621 PARTDUM -.453 .271 2.800 1 .09 .636 .374 1.081 MD(DUM .565 .216 6.866 1 .009** 1.759 1.153 2.683 JOBDUMMY -.690 .281 6.046 1 .01** .502 .290 .869 AACERTDM -.312 .275 1.292 1 .25 .732 .427 1.254 TRNSFRDM -.012 .248 .002 l .96 .988 .607 1.608 GPARECOD .174 .088 3.917 1 .04* 1.190 1.002 1.413 Constant -.537 .350 2.347 1 .12 .585 a. Variable(s) entered on step 1: GNDERCOD, BLACKDUM, HISPADUM, ASIANDUM, MDISADVAN, HDISADVAN, DEPENDEN, HSDIPL02, MARRYDUM, DIV ORDUM, PARTDUM, MIXDUM, JOBDUMMY, AACERTDM, TRNSFRDM, GPARECOD. *p<.05 **p<.01 While the full model did not provide an extensively more comprehensive understanding of stopout as an outcome, it did indicate which factors were significantly related to and more influential. Students with mixed enrollment had an increase in their odds of return, while minimally disadvantaged, divorced, and attending college for job skills were associated with decreased odds of returning to college. The following section 135 presents findings flom the qualitative portion of the study that provide a more detailed understanding of what is important to students in the educational paths leading to stopout. The qualitative data not only illuminates the quantitative data, but the sum of the study data contributes to a better answer to the overall research questions. Part B: Qualitative Method The qualitative part of the mixed methods study was designed to explore the internal and external factors involved in an individual’s decision to return to college. Additionally, this part of the study aimed to better understand how students managed these factions in relationship to decisions along their educational paths. This study focuses on the following research questions: 1. Why do students return after a period of non-enrollment? 2. What factors impact this decision? 3. What is the influence of these factors on stopout? 4. How do students manage these factors along their educational paths? Data source Location. This study was conducted at an urban, mid—western community college (using pseudonym of Archer Community College) that has 10,456 students. Approximately 60% are female, 25% are students of color, and 64% are part-time students. The average age of students is approximately 28-29. The city is approximately 53% Afiican American, 41% White, 3% Hispanic/Latino, and 1.1% of more than one race; all other racial and ethnic groups are at less than 1%. It is the fourth largest city in the state, yet the population has continued to 136 decline in the last decade with a reported population of 131,000 in 1998 and 114,662 in 2007. The median income for the city is $28,010 with 26.4% of the population living below the poverty line; 37.4% of those under 18 are below the poverty line (City of Archer, 2008). Participants. I interviewed 48 students about their decision and experience returning to a community college campus. Approximately a quarter of the participants were students of color (primarily Aflican American). The sample ranged flom 20 to 50 years old, with a mean age of 32.5 and 52% of participants falling between 23 and 33 years old. Fifty-two percent of the participants were women and the rest were men, with one female-to-male individual. All the participants were pursuing transfer, a certificate, or an associate’s degree and were currently enrolled. A majority of students (66%) have only stopped out once, but that stopout ranged flom one filll semester (not including a summer semester) to 15 years. Finally, while stopping-out places these students in a higher risk category for not completing their educational goals, 86% of the students also had at least one additional risk factor when they first left college for not completing their educational goals associated with non-traditional college students (Feldman, 1993; Hoachlander et al., 2003). Pseudonyms are used for all institutions and participants. Participants selected their own names. I recruited students using three methods: 1) I posted fliers advertising the study were posted across the campus; 2) I recruited in a limited number of classes; and 3) an email was sent to the entire student body by the college on my behalf. A majority of the participants were recruited through the email. All participants were made aware of their rights and reimbursed with $15 for their participation. 137 Interview. To facilitate inquiry, the interview questions guided students through their educational history; their goals, hopes, and fears and how they changed over time; and their decision-making processes to leave and return to college. A sample of questions asked included: Tell me about your experience in school before and in college; Why did you leave college? What did you see happening when you left? Did your academic goals change while you were gone? Why did you return to college? What happened in your life in the interim? Two pilot interviews were used to form the interview protocol. Students described why they left school and what their decision to return to college looked like. I drew upon possible selves theory, stopout, and attrition literature to develop the protocol and in analysis to further understand how a student’s conceptions of his or her possible academic and career selves influenced the decision to return. Analysis I used constant comparative analysis to analyze the data (Strauss & Corbin, 1998). I read the transcripts for codes and themes focusing on the reasons students’ left and returned to college, strategies and plans for returning, factors that challenged and supported educational paths, and the intersections among these areas. The factors associated with attrition and non-traditional students in literature served as a guide in analysis (Horn, 1996, 1998; Horn & Premo, 1993; Napoli & Wortman, 1996, 1998; O’Toole, Stratton, & Wetzel, 2003; Stratton, O’Toole, & Wetzel, 2004 ). A first pass at coding revealed themes in accordance with those previously listed. Additional analysis exposed more detailed patterns among student responses that identified how particular factors associated with non-traditional students influenced student decisions at particular 138 points on their educational paths—life prior to and attending college, leaving postsecondary education, and the return to college. In an effort to establish trustworthiness, emerging themes and patterns were shared with participants for feedback (Lincoln & Guba, 1985). I emailed students who expressed interest in learning more about the study’s findings during the interview a short explanation of findings and themes and four students emailed their feedback back. I used the feedback to evaluate the patterns and made adjustments to more accurately reflect participants’ experiences. In addition, once ten interviews were coded an individual unrelated to the project was asked to review a set of three interviews with the previously constructed codes. I compared the coding results by the different reviewers and made adjustments to themes and codes according to differences. Findings All of the participants in the qualitative portion of this study were previously enrolled in a two- or four-year college and left school at least once; in addition, all of the participants returned to postsecondary education to resume their studies and were current students at Archer Community College. Investigation into the ways that factors associated with risk for attrition influence participants’ educational decision-making not only provides insight into why students leave, but what influences their decision to return. In an effort to remain focused on the role of individual factors associated with attrition, the following themes reflected the factors that were most influential in the decision-making of the participants. The participants’ educational paths varied across individuals, but generally they all experienced three major decisions with regard to their 139 postsecondary education: initially attending college, leaving college at least once, and returning to college at least once. Each factor was not prominent at each step along students’ paths. Therefore, the findings are organized according to the major decision areas and within those events the most influential factors are focused upon. College Going Family background and expectations. Analysis of students’ stories revealed a pattern of which students went to college directly out of high school and what students’ attitudes toward college were in regard to family support and examples of college going. Students whose family encouraged or expected them to go to college or had models of parents and siblings who attended college were more likely to want to attend college right out of high school. College had meaning to them beyond a familial or societal expectation, these students believed it was important to go to college in order to achieve a better career and life for themselves. 4B6I74 (the student chose her own pseudonym) had one parent who went to college and one who did not. She describes how this influenced her attitude toward college, 4B6I74: I’ve always known that I’m gonna go to college. That was just factual for me. Interviewer: Why is it so important to you? 4B6I74: Well, my mom didn’t, first of all, and for everything that I wanna do, it’s pretty much a prerequisite. Interviewer: And what did the fact that your mom hadn’t gone to college mean to you and that decision to go to college? 140 4B6I74: Just the fact that she was always working at things like retail, or food service, or some job that was overly stressful. It just didn’t seem worth it, whereas my dad, on the other case, who had an associate’s early on in my life, and got a bachelor’s later, was working in the cube farm. . .most of the time that work is stable. Going to college was also important to these students on a personal level. It was what they expected of themselves and represented a personal achievement. Shirley left school to have a family, but returning to college was important to her for herself and for her grandmother, “I always figured I’d be the first one to graduate college. Well, that didn’t happen, but my grandma, who passed away in 2000, it was really important for us, to her, that we go to college. So I always promised myself. . .that I would go back and I would graduate.” Students internalized the belief that going to college was necessary and important and as a result they had clear and well-developed goals and desires to go to college. All but one of the students that attributed such meaning to postsecondary education had families that supported and encouraged going to college. The personal and professional meaning attributed to postsecondary education was also essential to these students’ repeated decision to return to college. Finances and enrollment. The majority of participants paid for college themselves, including a majority of students whose parents encouraged college. Three- fourths of students who entered college with a well developed conception of and purpose for college started as full-time students, of those students two-thirds switched to part-time enrollment upon their return to college. Most students indicated that they reduced their enrollment to part-time in order to acclimate back to school slowly and, after learning 141 flom past experiences, recognized the need to adjust their lives to be more successful this attempt. Student 4B6I74 became very overwhelmed the first time she was in college. She found that fulltime and online or open classes were very challenging for her, Well, I’m approaching it differently. I’m trying to stay in classroom classes because I know I do a lot better if I have to sit down and actually listen. . .(and) I realize that I can’t handle taking so many credits at once with certain classes, so I do them slowly. In contrast, a majority of students who lacked a sense of purpose for and in college began school part—time and over half of those students transitioned to full-time enrollment upon return to school. Like most participants, these students expressed a commitment to and sense of purpose for college that was not present in prior attempts at school. Partial evidence of this commitment was their willingness to enroll full-time, reduce or eliminate their work hours, and put their financial resources (e. g., loans or out- of-pocket funds) toward school. Garrett went through a female-to-male transition and as a result had trouble getting financial aid because of confirsion with the paper work. At the time of the interview he was homeless and living in his truck. When asked why he remains in school he stated, I could hire in probably to a company that’s going to get me a paycheck. . .but no guarantee that that’s gonna be there next fall as the economy’s still sliding. So if I go ahead and get my education, get a skill that I think is always gonna be there. . .it’ll be a better choice than to take the quick fix. Reasons for college going. Most of the participants originally indicated that they enrolled to get a certificate, Associate’s degree, or Bachelor’s degree, yet further 142 investigation revealed that their clarity and purpose for why they were in college varied. Some students stated that they enrolled in college to get a Bachelor’s or Associate’s degree and could articulate why this was their goal and what the purpose of college was in their lives. Tee enjoyed high school and was a good student, “I had planned on just going right off to college. . .Me and my best fliend was going to go and she was going to be a lawyer and I was going to be a lawyer.” As could be expected, a majority of the students who entered with a well-defined sense of purpose and goal for themselves to be college students could identify why they were enrolled and these reasons for enrolling were related to their decision to return. On the other hand, among participants with a less clear purpose for college and themselves as students, although they often enrolled in degree/certificate programs they expressed not having a clear direction for a career or major. Rather, they indicated that they came to college because it was what was expected of them, it was what their peers were doing, and it was often better than the altemative. Most of these students experienced a shift and developed clearer reasons for their eventual return to college. A number of students also took courses to improve their job opportunities. They were often testing the waters in preparation for potentially enrolling in a training or degree program. Multiple times Delane took individual college course to help prepare him for skill and aptitude tests required for jobs. Although at the time Delane had little desire to enroll in a degree/certificate program, he took individual courses to improve his job skills for training and job opportunities. Similarly, a few students took one or two courses for personal enrichment or to prepare themselves for greater enrollment. Beth had poor experiences with the school 143 during high school and initially had no interest in college. When she did try a college class she “just figured maybe I might just take a class for fun here or there” and found “they were actually fun and surprisingly I got four points in them, so I’m like, ‘I could do this.’” Originally, Beth did not have a goal or an interest in pursuing a degree/certificate, but the experience of taking one or two college courses for entertainment led to greater educational ambition and a greater commitment to returning to college. The reasons students originally enrolled in college often appeared straight forward, but upon investigation reflected more complex motivation and decision making. Students would enroll, but flequently lacked the motivation or desire to be there. Others would enroll for reasons unrelated to degree/certificate attainment. Additional discussion of why students leave and return to school displays how reasons for enrolling influence later educational decisions. Leaving College The primary theme throughout this section reflects how regardless of the reason students left college, they were choosing to prioritize other roles and concerns in their lives over going to school. They saw these other issues and roles as operating in contradiction to their college enrollment. Work and finances. One of the primary reasons students gave for leaving school was the decision to take a full-time job. Most of the time students did not try to pursue classes at the same time as working, rather they had the idea that if they had a good job that did not require education there was no purpose to continue paying for and pursuing college. John was in college when he was hired at a local automotive manufacturing facility. He felt the pay and benefits were good and there was not much incentive to 144 continue with school, “There was a lot of fringes that went with it, so I was thinking, ‘Hey, this isn’t too bad. Maybe I’m going to stick with this.’ . . .I was taking a philosophy class, and then when I was done with that. I kind of stopped.” This was particularly the case for a student who did not hold college-going as an important goal for his— or herself. Children and family. For many students the decision to leave school and pursue work was not solely about making enough to live on and college no longer having a role, it also was related to students’ role as provider and caregiver and the decision to prioritize that role over that as student. A reason given by participants for taking time flom college was life transition—specifically the decision to get married and/or have children. As participants were preparing to be spouses and parents, school was another aspect that needed to be managed and the area that was least important and purposeful for life during that time. Having children was a particularly poignant issue for women. For many, getting pregnant was unplanned, therefore leaving college was part of an effort to manage and triage such a surprise. Renee had a series of marriages and a pregnancy that contributed to a cycle of leaving and returning to college. This was often the situation for students with clearer college goals who generally enjoyed school and did not have plans to leave, but found themselves in a difficult position. Other women planned their pregnancies and leaving college was part of being a better parent. Lynn returned to college multiple times only to leave each time because of a sense of responsibility to her family, I went with the intention of continuing each time I went. It was difficult on me because I had small children at home and we had two businesses that we were running, and so I was busy. But, it was even harder on my family, and they 145 generally didn’t take it really well, and so it just became priorities. . .I was the one that chose to get married and have children, and they were my first responsibility. To these students the role of mother took priority over that of student. Personal. Finally, similar to students’ responses to an unplanned pregnancy, participants experienced life situations that flom their perception conflicted with being in school. Drug abuse and illness were among the most often cited issues. Substances were more important than school and took priority of academic work. There were also a number of deaths and illnesses that participants had to deal with, often choosing to prioritize their roles as caretakers over students. Returning to College Work. Whereas the opportunity of a fulltime job prompted many students to leave college, a lost job and seeking job stability drew students back to school months and years later. The auto industry’s deindustrialization affected this region particularly hard. Many of the students within this study were laid off and had many years of unstable employment prior to losing their positions. Despite a lack of funding, they often chose to return to college to pursue careers that were more stable and had higher salaries. Jane demonstrates this in the recounting of her return to college, So, I ended up workin’ in the warehouse and losin’ my job, so had my son in college, and I was like, well, you know what I’m gonna do? I’m gonna go back to school, you know, instead of goin’ out here gettin’ me a little $10.00 an hour job workin’ full-time, I’d rather just go ahead and go to school full-time and try to pay my little bills as I can find payment to pay ‘em with. And then, after God bless me and get my degree and I can make some money, then I’ll pay them off. 146 And rather than me makin’ $10.00 and spendin’ the rest of my life tryin’ to pay them something out of $10.00. Some students received funding through a trade retraining program for fulltime tuition and living expenses. Many saw it as an opportunity to pursue a more stable line of work. Blake was laid off and receive trade retraining funding, I mean it was—I had a job opportunity to start at (a local factory) in the maintenance program making more money than I was making or go to school for flee. Well, that was a decision because I didn’t like school the first time I tried it. . .So just talking it out and weighing all my options. . .I decided “okay, I’m gonna go at this. . .My brother, he encouraged me because he’s been going to school. And my wife said it was a good opportunity. My father said it was a good opportunity. . .So I did it and I haven’t regretted it. The funding served as motivation to return. Postsecondary education took on purpose and priority in the lives for most of these students that was not present previously. As a result, they were willing to sacrifice money, time, and effort to pursue an education; a price that was too costly in previous attempts at school. Children and Family. Similarly, participants previously viewed commitment to family and children in opposition to going to college. They believed they could not be good mothers and parents while also being a student. Yet, as life circumstances changed, participants began to see going to school not only as a possibility, but as necessary to good parenthood. The changes in individuals’ lives varied, ranging flom divorce to resolved family illness to children aging and transitioning to school. As issues that were 147 challenges to college going resolved, individuals began to consider returning to college. Beth spoke about how the illness and passing of her husband affected her enrollment, My husband at the time was making $15.00 an hour, so there wasn’t necessarily a need for me to work (while going to school), but then he got ill last year and I ended up having to drop out last year to stay home and take care of him, and then he passed away. So I had to take another semester off just to deal with that, so I’m just now getting back after another year off school. Furthermore, parents now saw it as their responsibility to demonstrate the importance of going to college through their own examples. This demonstration included their children seeing their college campus, going to class, and doing homework. Delane describes a conversation with his son, I was talking to my son and I asked him what did want to be and be was telling 1 me wanted to be a doctor. . .I was like, ‘I have to lead and be an example to him so he can say, ‘Well my dad has a degree, so I’m gonna go to school.’ But he’s not gotta have a choice in the matter. For others it was about not being hypocritical and emphasizing the importance of going to college to their children while not attending themselves. Autumn emphasized how important it was that her children go to college and explained why her children were a primary reason for returning to college herself, “Setting a new example for the most part, because I stress, ‘you will go to school.’ A diploma’s just not enough anymore. Come on, you’ve gotta have a high school diploma just to work at McDonald’s. How can I set an example if I haven’t done it myself?” The participants wanted to be good parents 148 both by pursuing a career that was more stable and required postsecondary education and by providing an example of college going. Personal Enrichment. Finally, students who initially had well developed goals to be in college gave a reason for return that was unique to this subset of students. They stated that part of their reason for returning to college (multiple times) was not only for career and family, but to fulfill a personal goal to finish college. Sally returned to college many times over the years. She spoke about why, “’cause that’s my dream and I refuse to let go of it, no matter—even though I messed up so much. . .it was I want my degree because that - that - that will complete me in a way that no one can understand.“ From their first enrollment in college, these students knew that they wanted to be college students and why it was important to them. Although, the made choices that drew them away flom attending school, they retained the goal to some day finish their postsecondary education goals. Retuming to college after nonenrollment and pursuing those educational goals was for personal enrichment and fulfillment as much as it was about creating a better life for themselves and others. While students left college because their goals and life roles conflicted with going to college, returning to school occurred when participants saw college and postsecondary education as an integral part of other priorities. College became purposeful and necessary for stable employment and being a good provider, parent and fulfilled person. The quantitative part of the study analyzed the relationships between factors associated with risk of departure flom postsecondary education, while the qualitative part explored in more detail the role of those and other factors in students’ educational decision- 149 making. The discussion section draws on these findings to address how this information intersects to answer the study’s research questions. Discussion The qualitative and quantitative findings come together to speak to the study’s overall research question—why do students return after a period of enrollment?—by answering the two sub-questions: Which factors are related to who returns after a period of nonenrollment? And, how do these factors affect educational decision-making with regard to leaving and returning to college? Examined on its own, the quantitative data indicated that mixed enrollment was associated with a higher likelihood of returning to college, while students were divorced/widowed/separated, enrolled for job skill development, and disadvantaged were less likely to return to college. Yet, when the qualitative data is introduced a far more complex picture emerges in response to the research questions, indicating that these factors can both contribute to and dissuade students to leave and return to college depending on the students’ environment, context, and beliefs about college. The qualitative data also suggested a number of other factors that were influential in students’ decisions to leave and return to college, but were not significant or included in the quantitative data. The additional issues explored include work decisions, children, and finances. Each factor was related to and influenced students’ stopout paths in multiple ways. The following discusses what the data, quantitative and qualitative, says about how these factors influence student stopout. 150 How was mixed enrollment related to the return to college? How did it influence the decision to leave and return to college? In the model mixed enrollment was a positive significant predictor of student who returned to college. In fact, it was more predictive than either part-time (4.17 times as likely) or full-time (7.2 times as likely) enrollment alone. There are studies that found part-time enrollment associated with increased likelihood of return to college (Gosset, 1993; Hoyt & Winn, 2004), but the result that mixed-enrollment was most predictive is unexpected. The qualitative data supported the indication that there may be a relationship between returning to school and enrolling both part- and full-time on different occasions. Over half of the participants experienced mixed enrollment over the course of their educational paths. The remainder of participants was evenly split between students that enrolled exclusively part-time and exclusively full-time. The students who switched flom part-time to full-time explained the decision to shift enrollment intensity as a part of the commitment to retuming to school and achieving their educational priorities. Those that shifted flom full-time to part-time discussed the decision as evidence of making adjustments to be more successful this enrollment than in past attempts. This could mean taking fewer classes to re-acclimate to being in school or taking on less responsibility to avoid being overwhelmed. Enrollment shifts in both directions reflect recognition on behalf of participants of what may have not worked during previous attempts at school. Such willingness to recognize one’s abilities and limitations and to be flexible enough to implement changes (e.g., changing enrollment patterns to accommodate family and work life) may indicate an increased commitment to return to and succeed in school. 151 How were decisions about work related to the return to college? How did they influence the decision to leave and return to college? Related to this issue, what was not included in the model but was a theme throughout the qualitative study was the importance of working and how it figured into the educational decision making of students. As mentioned previously, many of the students in the qualitative portion of the study left to pursue fulltime work and or attempted to balance school, work, and family but became overwhelmed. Students often did not see a purpose in remaining in school when they could make a living wage without postsecondary education and/or saw leaving school to work as prioritizing the needs of family as a provider. Previous research on stopout found that community college students often cited conflict between work and school as a main reason for leaving, compared to students at four-year colleges who cited financial concerns and grades more often (Ahson, Gentemann & Phelps, 1998; Bonham & Luckie, 1993; Conklin, 1993; Hoyt & Winn, 2004; Stratton, O’Toole, & Wetzel, 2004; Woosley, 2003/2004). What was not discussed in previous studies was that the intersection between work, school, and family was also a primary incentive for students to return to college. Additional postsecondary education was now seen as an opportunity to gain more stable employment, provide better for family, and serve as an example to children. How was marital status related to the return to college? How did it influence the decision to leave and return to college? Being divorced/widowed/separated (div/wid/sep) was another significant variable within the model. Div/wid/sep students were 77% less likely to return than married or 152 single individuals (together), and almost 100% less likely than married or single individuals analyzed as separate groups. Students who are divorced may have less support than students flom either of the other two categories. Eighty-two percent of single students are dependents, while 100% of married students and 95.7% of divorced/separated/widowed students are financially independent. Therefore, single students are likely to have the financial and personal support of parents and family and married students are also likely to have the resources and support of a spouse, while divorced/separated/widowed students are likely to be independent flom family without the support of a spouse. Having fewer resources and less support may contribute to the decreased odds of returning to school. Within the qualitative portion of the study there were seven participants who were divorced and two who were widowed. Since all of the students in the qualitative portion of the study returned to college, it is challenging to compare who was more likely to return to college in regard to marital status. That said, the qualitative data did suggest that marital status and events concerning marriage/divorce/widowed did factor into when and why students returned to school. First, marriage, particularly for women, prompted many women to leave school. They spoke about the transition to marriage being overwhelming and then moving quickly into having and caring for a family. The students who were divorced or widowed did not speak about those events as reasons for dropping out or significant baniers to re-enrolling, rather they were part of the ongoing events of life that included changes in work and family. Once students decided they wanted to return to school, either as a newly developed or a persistent goal, becoming divorced or widowed served not as a deterrent to returning but rather a reason 153 to delay the return. John Wayne took courses at a community college for years before leaving for a fiilltime position in a factory. When his wife passed and his daughter was four-years old he “got to thinking, ‘Things aren’t going well in the automotive industry. Maybe I need to start looking back into going back to school.” John initially returned to school part-time in 2006 and switched to full-time in 2008, but it took time to return to school and make that transition. He spoke about the challenge of managing his home life, It was difficult. Well, a lot of the difficulty came flom my home situation, being a single parent. I had to make sure Grandma could watch the little one, or a babysitter here or there. . .The concern was how I was going to manage the home life. How am I going to pay the bills? How am I going to make sure my little one has clothing. Things like that. Being divorced or widowed represents a major change in an individual’s life that often takes time to recover flom, therefore it is reasonable to believe that without the additional financial or personal support of a spouse the person would be less likely to return or to delay returning to college than a married or single person. This supports the quantitative findings and aligns with previous research on marital status and stopout/dropout. Stratton et a1. (2004) found that marriage was associated with stopout for men and women and divorced women were more likely to dropout. Singlehood was also associated with stopout (Hoyt & Winn, 2004). But when there was a change in marital status getting married was associated with a higher likelihood of dropout while divorce was associated with stopout (Stratton et al., 2004). 154 How were the reasons students originally enrolled in college related to the return to college? How did it influence the decision to leave and return to college? The study’s findings indicate that students who enroll in college for job skills have decreased odds of return (50% less likely) compared to those who enroll for personal enrichment, AA/certificate, or transfer (together). When the odds of returning for students who enrolled for job skill development is individually compared to students who enrolled for personal enrichment (.853), AA/certificate (.7 73), and transfer (.256), the former students remain less likely to return than individuals in the other three categories. Analysis of the qualitative data resulted similar findings. Students who had clear, purposeful collegiate goals, such as a degree or certificate, maintained their desire and attempts to return to college after they left. Conversely, students who did not have a clear purpose for being in college or were in college to improve odds of finding a job were more likely to leave and not retum for a significant period of time. Students who left because they did not have a clear vision of themselves as a college student or for postsecondary education did not return until they had a well-developed purpose for college in their lives. Previous research also found that students were more likely to persist in college when they had clear goals for enrollment (Choy, 2001; Summers, 2003). Furthermore, many students began college because they were told they could find a stable job with postsecondary education (job skill development), but did not have a clear major or academic outcome in mind. Therefore, when they found a fulltime job, they no longer had a purpose for college that kept them flom leaving school. For these 155 students, as long as they had a stable position that met their financial needs, they did not see a purpose in returning to college. This reflects the results in the quantitative portion that found students enroll for job skill development to be less likely to return than students who were there for personal enrichment, transfer, or a degree/certificate. The qualitative data provided more insight into the ways students’ initial goals for enrolling in postsecondary education influence their educational paths. It also brings into question how reasons and intention for attending college are measured. The BPS survey allowed for four options (i.e., job skills, degree/certificate, transfer, and personal enrichment) but it became clear through the qualitative data that although a student may select degree/certificate, transfer, or job skills, how well college fits into the larger context of their lives and how purposefully they attend may vary. A majority of students in the qualitative study lacked a purposeful conception of why college was important to them and left when in their perception they got a better option or had life events that took priority over school. Yet, many of these students had been told that they should go to college to get a degree/certificate and a better job. They were savvy about what their families, peers, and schools expected of them and likely would have selected one of the survey question options that matched those expectations, even if was not a goal that they were purposefully and personally committed to. This brings into question the ability to measure intentions that have so many dimensions. How was student disadvantage related to the return to college? How did it influence the decision to leave and return to college? The final significant variable in the model was minimal disadvantagedness. Initially the finding that minimally disadvantaged students were less likely to return than 156 moderate/high or not disadvantaged (together) students is perplexing because it is reasonable to expect that students who were minimally disadvantaged would be more likely to stopout than students who were moderately/highly disadvantage. Yet, when the odds ratios for not and minimally disadvantaged students (1.53) and minimally and moderately/highly disadvantaged students (4. 17) were individually compared a clearer pictured emerged. Not disadvantaged students were one and a half times more likely to return than minimally disadvantaged students, but minimally disadvantaged students were more than four times as likely to return than moderate/highly disadvantaged students. The variable used to measure the level of disadvantage within BPS is not only based on income (i.e., poverty index), but takes into account factors related to social capital (i.e., parents’ highest education and economic status of high school). Therefore, the variable representing this index has implications for a better understanding of the association between family income and stopout and more broadly the relationship between social capital and stopout. Social capital is the social relations and structures that form highly interconnected networks (e.g., family, school relationships, work) through which norms, expectations, and standards are developed (Bourdieu, 1987; Coleman, 1988; Kim & Schneider, 2005). It is particularly through parents and family networks that critical information and values for educational success are transmitted to children (Coleman, 1988; Kim & Schneider, 2005; Schneider & Stevenson, 1999). Research has demonstrated that social capital has an influence on the educational outcomes of children (Coleman, 1988; McNeal, 1999; Stanton-Salazar & Dornbusch, 1995; Sun 1999) and transition to postsecondary education (Kim & Schneider, 2005; McDonough, 1997; 157 Outcalt, Tobolowsky, & McDonough, 2000; Walpole, 1998, 2003) independent of socioeconomic status (SES). Therefore, the use of a variable that includes indicators of social capital in addition to SES suggests that students with less social capital (not only less income) are less likely to return to college after leaving. This is important for understanding how the quantitative and qualitative data intersect with one another. Themes within the qualitative data indicated that the example of parents who went to college and parental encouragement to attend college had an influence on the development of students’ attitudes toward college and their decisions about where and when to start college. Students who developed a clearer concept of college going and believed that acquiring a degree/certificate was personally important to them more often had parents/siblings who went to college and families who encouraged them to utilize postsecondary education. A majority of them also started at four-year colleges. This attitude and commitment toward college stayed with these students throughout their educational paths. Most of the students who had a clearer concept of college-going also returned multiple times to college to attempt to fulfill their educational goals. Although their previous efforts were not successful, they never lost or changed the desire to obtain a degree/ certificate. This reaffirrns the quantitative results. The more students had examples that college going was important and were encouraged to attend themselves, the more they internalized the value that postsecondary was a primary goals for their professional and personal lives. These priorities fueled students’ continued desire to return to college. There is substantial educational and sociological evidence in the literature that higher parental education and expectations are related to continuous 158 enrollment and persistence (Coleman, 1988; Kim & Schnieder, 2005; McDonough, 1997; McNeal, 1999; Outcalt, Tobolowsky, & McDonough, 2000; Staton-Salazar & Dornbusch, 1995; Stratton et al., 2004; Walpole, 2003a, 2003b). This study affirms these findings and extends on them by illustrating not only that parental education and expectations are important, but also the influence they have on the development of student attitudes toward college and subsequent educational decision-making. How were personal finances related to the return to college? How did they influence the decision to leave and return to college? Additionally, the qualitative study also demonstrated that students’ financial situation entering and progressing through postsecondary education influences their educational decisions. While students shared multiple reasons for delaying enrollment after college, almost all of them were responsible for paying for their education and had families that were not demonstratively supportive of college going. In addition, students’ perceptions of their financial stability contributed to decisions to leave college and return. Although the statistical results that higher levels of disadvantage are less likely to return to college after leaving are reflected within the qualitative study findings, the variable indicating delayed enrollment was removed flom the model because it was consistently not significant and did not sufficiently add to the model’s explanation of stopout indicating it was a less important variable in determining who might return to school. In previous research, low family income was related to increased departure (Johnson, 2006; Swail, Cabrera, & Lee, 2004; Terenzini, Cabrera, & Bemal, 2001; Tinto, 2006; Titus, 2006; Walpole, 2003b) and in studies of dropout and stopout, financial reasons (e.g., unmet need and cost of tuition) were often cited as reasons for departure 159 (Ahson etal., 1998; Bonham & Luckie, 1993; Conklin, 1993; Herzog, 2004; Hoyt & Winn, 2003/4; Light, 1993). Similar to the case of parental education, the quantitative and qualitative findings support the belief that income and finances are critically important to college access and going (Swail, Cabrera, & Lee, 2004; Terenzini, Cabrera, & Bernal, 2001; Tinto, 2006; Walpole, 2007), but it also emphasizes that a students perception of the purpose and role of college in their professional and personal lives is necessary to consider. Summary In the discussion section I drew upon the study’s qualitative and quantitative data to better understand which factors were most salient in students’ stopout experience and how those factors influenced their decision-making. This attempt to address the overall research questions for the study suggested a number of conclusions. First, a majority of the variables that were significant within the model were also supported within the qualitative data, yet a number of themes within the qualitative data were not significant within or retained within the model. I believe this demonstrated the complexity of educational decision making that students experience. The model revealed which risk factors were more predictive of student retum to postsecondary education and the qualitative data extended on this information by showing how and what role these factors played in student departure and return. The decisions students made were multidimensional, taking into account many issues at the same time and illustrated how many of these factors intersected with one another. Second, prior to initially entering college, students generally had a preformed conception of what the purpose of college was and how this fit into the context of their 160 13:58:11 lflprP 'gf'r‘t‘ISt’ 1 - 1‘. 1131 {01.6 and '1 ram 11:5 1:. 9081580 hcew and The quallf f0r at Shell and 11 53m: “tert- that t lives. Students that believed college was important to them professionally and personally were persistent in their desires and efforts to return to school, while students who lacked purpose in/for college and did not want to be in school had to develop a sense of purpose for college that led them back to school at a later time. Finally, the decision making process for leaving and returning to college revolved around the evaluation of priorities and identifying where college fit in among them. A majority of students left school because they prioritized other issues and factors in their lives (e. g., providing for family, full—time work, illness, pregnancy, etc.) above a postsecondary education. In their perception, college did not fit in their lives because there was either no room or it was not necessary. It was the rearranging of these priorities and the integration of college into them that led students back to school. Limitations There were a number of limitations within this study. First, stopout is conceptualized somewhat differently in the quantitative and qualitative studies. Specifically, the time flame between leaving and returning is defined different in order to qualify for stopout. The BPS survey duration was six years, therefore students who left for at least four months and returned within six years experienced a stopout. I was less specific about the timeflame for stopout in the qualitative study. I am interested in shorter and longer terms of stopout. Therefore, I limited the shortest length of absence to one semester (not including summer), but did not restrict how long the absence could be. Therefore, when comparing the two types of data there could be differences in students that take shorter versus longer periods of time to return to school. 161 Second, the model did not include all of the risk factors associated with non- traditional student status. While some factors were not included because they did not substantial contribute to the model, work intensity was never included because of potential multicollinearity with enrollment intensity. In addition, I identified a number of factors and themes within the qualitative data that were not included in the quantitative study. Finally, when comparing the two sets of data, the analysis is limited because the quantitative study was designed to examine the influence of factors on students who returned to college (stopout) and students who left college without return (dropout), while the qualitative study focused exclusively on students who returned to college. Implications and Further Research This study has a number of implications for further research and practitioners. First, additional statistical analysis should include a more comprehensive list of variables associated with high risk of departure. Additionally, I would attempt to include variables representing the themes discussed in the qualitative portion of the study. Second, patterns associated with a potential influence of gender on students’ reasons for educational decisions require additional analysis due to limited exploration within the qualitative findings. Finally, I would suggest exploring an intervention study based on the study’s findings. The results suggested that non-traditional students’ lives are complicated and attending college cannot be separated flom the context of lives both logistically and for educational decision-making. Also, the importance and purpose that students’ assign to postsecondary education is critical as they prioritize the many responsibilities and goals 162 grand and w lrecor 3:301 schedl anew qr grammar national p 11:5 are not 1 9311161115. per- 1: 1111111 idual 1531191111116 j 3110111613de 1 reel 11‘1ng for 1191‘ face: in their lives. Assisting students to clarify their beliefs about the purpose of college in general and within their lives can assist students in the decision making about college. I recommend an intervention program that is designed to identify students who are not scheduled to return and work with the student to evaluate why they are leaving, addressing questions such as: What are the student’s reasons for leaving? How does postsecondary education fits within the student’s priorities? And what is an appropriate educational plan? The intervention must include student purpose and priority. Students’ lives are not linear or neat. They are complicated and changing depending on context of students’ personal lives—this is why addressing retention and the return to college must be individualized and multi-pronged. Issues challenging college going for some students may require problem solving and resources in order to help them remain in school, while another student dealing with similar issues may need to take time off. I recommend that the following components must be included in a retention program for non-traditional students in order to work with the challenges and needs that they face: 1. Assessment: Why is the student leaving? What is going on within his or her life? How does college fit within life, goals, and plans? The context of students’ academic and personal lives must be assessed for how postsecondary education fits within it. Consider not only what is going on in their lives, but also how they perceive the purpose of postsecondary education and its priority in their lives; 2. Resources: Can this student be retained with supportive resources? Evaluate with the student what resources are available to help them stay in school—— through the school and in their lives? Again, help the student assess the purpose 163 Com of college in their lives. The students in the study that saw a purpose for college going and believed it was a significant priority for multiple reasons, were much more likely to make efforts and sacrifice time, effort, and money to remain in school. 3. Make a plan: If the student is trying to stay in school, a short term plan for next semester should include academic goals and a plan, an identification of challenges to remaining in school, a plan for addressing those challenges and what resources are necessary. Ideally, along term plan for the students’ educational goals will also be developed. The student’s plans should not only include his or her academic goals, but also draw upon the internal and external support available to the student. A student may also decide to leave or take a break. If he or she decides to leave, help the student develop a plan for return. 4. Institutional plan: Finally, the institution should have a system in place for following-up with students who have left, such as a personal letter or call. Community colleges disproportionately serve students who are at higher risk for attrition. These students have different educational experiences and life contexts that contribute to their decision-making regarding postsecondary education. Typically educational institutions are designed to serve traditional students—who generally have less complex financial, family, and educational backgrounds and are experiencing different developmental change. I am recommending that institutions that serve a large proportion of non-traditional students must examine whether their services and structure are designed to help these students succeed. 164 "I": 1‘ 3.; Q . 11. h... &\ Conclusion The recent attention focused to the issue of increased educational attainment in postsecondary education by President Obama’s address to Congress (2009) and the Lumina Foundation’s subsequent suggestion to target students who have left college for reenrollment, heightens the need for additional research and implications for practice that focus on nontraditional students and stopout. This study attempted to contribute to this need for information through the analysis of mixed methods data. The findings revealed a number of factors associated with increased or decreased odds of students returning to college and explored how they influenced students’ educational decisions. The results suggested that the complexity of students lives coupled with students’ motivation, attitude, and beliefs about college influence the why they choose to leave and/or return at certain periods of time. 165 V «- 1330b»: nong held: 0111115 1 r:5:arcl IELTCGSC L's-sins cctific: papilla: Strategy 20091.‘ is scam. Pfilod ( 11115 (81 1 Students 311111211” 111611 re! finders“ preSeme CHAPTER 6: CONCLUSION Introduction Retention and postsecondary degree attainment remains an ongoing concern among higher education practitioners and researchers. However, the recent call by President Obama (2009), echoed by the Lumina Foundation, (2008) enhanced awareness of this issue at the national level and intensified the need to respond to the call with new research and suggestions for practice. President Obama (2009) set a goal for the US. to increase the proportion of college graduates to the highest in the world by 2020, while the Lumina Foundation (2008) advocates for 60% of Americans to have a degree or certificate by 2025. Yet, current estimates indicate approximately 40% of the US. population have a degree/certificate (Lumina Foundation, 2008). Attracting students back to college to complete their education is part of the strategy suggested to increase degree/certificate attainment rates (Lumina Foundation, 2009). Yet the research on students returning to college, particularly community college, is scant. Understanding how and why students return to continue their education after a period of nonenrollment is critical to improve the degree/ certificate attainment rates of this (at risk) population. A deeper comprehension of and appreciation for returning students’ educational paths has the potential to develop institutional support systems and structures that will prevent students flom departing higher education and will facilitate their return after a period of absence. This study was designed to contribute to the research on returning students, an understudied population. It was a mixed methods study in two individual parts and presented in three journal article-style pieces. The overarching focus of the study was an 166 .4 3C10 prac bent SIX-y Stud} exploration of the reasons students returned to college and the factors that influenced that decision. Throughout the study, I use the term “stopout” to refer to students returning to college. In the study’s first part I conducted a quantitative analysis of the relationship between factors associated with a high risk for departure and nontraditional student traits and community college students who returned to school after departure. The second part of the study was a qualitative inquiry into the educational paths of community college students who stopped out, but were currently enrolled. I presented the findings flom the quantitative and qualitative parts of the study in the first and second articles. The third article was an exploration of the intersection between the quantitative and qualitative results. The research questions addressed across all three articles were: a) Why do community college students return after a period of nonenrollment? b) Which factors influence the decision to return? c) What is the influence of these factors on the decision to return to return? Each study focused on a different aspect of these questions. Following is a summary of each article. I then discuss the overlapping themes across the parts of the study and present some general suggestions for future directions in practice and research. Article 1: Quantitative Study Summary This article reported the results of a quantitative analysis of the relationship between community college students who stopped out at least once over the course of a six-year study and factors associated with a high risk of departure flom college. The study focused on the following research questions: Why do community college students return after a period of nonenrollment? 167 11011811 1 char {tidy design or communl Tn: outcom arming st hall lean-in; Chamcreris' Msennc' ldi‘lllllled associated Characteri departure QUICOme SignifiCa Consld-u not Inch 510130111 Sigmfic the (lee a) Which factors were related to who returns after a period of nonenrollment? b) What is the influence of these factors on stopout? I examined data from the Beginning Postsecondary Survey 96/01, a large-scale study designed and conducted by the National Center for Education Statistics, focusing on community college students who departed college between 1996 and 2001 (n=779). The outcome of interest was students who returned to college during that time; by returning students complete a stopout cycle of leaving and retmning to college, rather than leaving and staying out. I was particularly concerned with the relationship between characteristics associated nontraditional student status and students’ return to postsecondary education. I selected independent variables that previous research identified as associated with a higher risk for departure. These variables were also associated with nontraditional student status; therefore a student with a higher number of characteristics associated with nontraditional status will also have a higher risk of departure. l utilized direct logistic regression to analyze the relationship between the outcome and independent variables. Results indicated that while the overall model was significant, its contribution to the explanation of who returns to school was not considerable (11% of variance explained) and suggests that there are additional factors not included in the final model that would significantly contribute to understanding stopout as an outcome. That said, the individual variables in the model that were significant provided additional information about which factors were more influential in the decision to reenroll. 168 (153115 2 med tier d: were 1 that m: The final model had four significant variables associated with who stays out and who returns to school— mixed enrollment intensity, divorced/widowed/separated marital status, job skill development as a reason for initial enrollment, and minimal student disadvantage. First, a result I found surprising result was that students who experienced mixed enrollment (part- and full-time enrollment) were more likely to return to college after departure than students with exclusively part- and full-time enrollment intensity. One possible suggestion for this finding is that students who utilized mixed enrollment developed more flexible enrollment patterns in order to accommodate and balance life events. Second, results indicated that students who were divorced/widowed/separated were less likely to return to college than their single or married counterparts. I suggested that married and single students may have more emotional and financial support flom family and spouses than divorced/widowed/separated students and may contribute to their ability to return to college. Third, students who enrolled in college for job skill development had decreased odds of return compared to students who enrolled for personal enrichment, AA/certificate, or transfer (compared as a group and individually). The significance of this variable points to the importance of student goals and intentions for enrollment on students’ departure and enrollment decisions, and reflects different conceptions for the purpose of college. The specific goal of enrolling in school for a degree/certificate attainment may provide a clearer motivation for return, while students enrolled for job skill development may complete their educational goals without acquiring a degree/certificate and depart without a need to return. Furthermore, students attending for 169 personal enrichment may be more likely to return because they view education as a process of ongoing development rather than for a terminal purpose. Finally, the finding that students who were minimally disadvantaged were more likely to return than students with no disadvantage, but less likely to return than those who were moderately/highly disadvantaged was to be expected based on previous research (Kim & Schneider, 2005; McDonough, 1997; Outcalt, Tobolowsky, & McDonough, 2000; Walpole, 1998, 2003). Since the variable for student disadvantage is based on an index compiled flom data on parent education, economic status of high school, and a poverty index, the result not only supports previous research that indicates the importance of financial need on student persistence, but also suggests that students’ social and educational capital are also influential. In light of the study’s results, I provided a number of suggestions for research and practice. First, additional studies of stopout would benefit flom the inclusion of additional variables including factors associated with academic integration, social integration, and financial aid. Second, a qualitative study of the relationship between risk factors and stopout could extend on the role of these factors in the process of educational decision- making. Finally, the complexity of student experiences indicates the need to develop campus support services that are the responsibility of faculty, staff, and administrators across the campus, but are structured to connect with the individual. Article 2: Qualitative Study Summary The second article presents the findings flom the qualitative portion of this study. The purpose of the study was to explore the external and internal reasons and factors involved in students’ decisions to return to college after an extended absence. This study 170 specifically sought to explore the role of students’ concepts of who they might be (or want to avoid becoming) in the college and career domains of their lives, students’ career and college possible selves. Specifically, I addressed the following research questions: Why do students return after a period of nonenrollment? a) Which factors impact the decision to return to college? b) What is the influence of these factors on the students’ decision to return to college? c) What role do college and career possible selves and learner dispositions play in the decision to return to college after an absence? (1) How does stopout influence individual’s academic and career possible selves? I interviewed 48 students flom a community college with an enrollment of approximately 10,500 in an industrial, mid-west city with a population of approximately 115,000. The interviews were 30 to 60 minutes and students were reimbursed $15 for their time. Students were included in the study if they met three criteria: they were previously enrolled in college and took a leave of absence for at least one semester; they were currently enrolled at the site institution; and they had academic goals of either transfer to a four-year institution or an associate’s degree/ certificate. I used a grounded theory approach, using a semi-structured, open-ended protocol. I asked students to respond to questions focused on their educational history; their goals, hopes, and fears and how they changed over time; and their decision making process to leave and return to college. I used constant comparative analysis to develop codes and themes related to the reasons students returned to college, strategies and plans for return, 171 hopes and fears for academic and career selves, and intersections among these areas. I shared emerging themes and patterns with participants for feedback and to establish trustworthiness. Analysis of the interview revealed two different educational paths that were partially shaped by students’ college possible selves as they initially entered college. The different educational paths and influence of possible selves and other important factors came to light as students discussed critical decision-making points along their journeys. I created a model of these points to illustrate the experiences that all students moved through and the various choices they made. What is distinct about this model is that it includes stopout as a juncture for these students. The primary experiences represented were high school experiences, the initial entry and exit flom college, and subsequent reentry (and at times additional exits) into higher education. I found that students’ high school academic experiences contributed to their attitude and perception of the purpose of college as they exited the K-12 system and entered into higher education. Students with more positive academic and social experiences in college tended to enter college with clear and well-developed college possible selves. College was important to them and these students internalized it a personal goal. Many students exited high school with low academic self-efficacy and/or feeling “burnt out.” These students did not have a purpose for being in college, lacked motivation, and did not have a clearly defined college possible self. This was the inception of two educational paths that included students returning to college. Participants’ reasons for leaving college included external factors such as financial burdens, job prospects, academic failure, and family responsibilities, and 172 internal reasons including beliefs about academic belief and belonging, lack of purpose for college, and the prioritization of other issues in their lives. Students with clearer college possible selves tended to leave due to unexpected events (e. g., illness or pregnancy) or the prioritization of other roles in their lives (e.g., mother, provider, worker). On the other hand, participants with less developed college possible selves often did not want to be in college and did not see a purpose for it. Therefore, they chose other roles and paths (e.g., got married or found a good paying job); they did not see a conflict between their goals and not having a college education. I noted that changes in individuals’ lives often prompted a shift in students’ college possible selves that marked an eventual return to college. Events such as work instability or lay-offs, children starting school, rehabilitation and getting out of jail motivated students to reconsider their lives and choices, resulting in clearer, more well- developed college possible selves. Students who maintained college possible selves flom previous enrollments did not necessarily experience a shift in those possible selves, but found that changes in their lives prompted them to be able to return to college at that time and see how the goal of going to college now fit in with other possible selves (e. g., efficiently providing for family or being a good example to children). Students with non- college possible selves were often developing these college possible for the first time and now saw how going to school fit in with other goals as well. For a majority of participants, returning to college with more committed career/college possible selves meant that they were taking their academic work more seriously. They changed their attitudes toward the purpose of learning, education, and 173 learning strategies for academic success. As a result, they did better in school, which led to an increased commitment to school, and additional effort at academic work. I suggested that future research on this tOpic should include additional exploration of mechanisms of change for possible selves and learner disposition and understanding stopout in relation to different institution-types and regional environments. Also, it would be beneficial to test the paths presented in this study through survey work and path analysis. The findings also resulted in recommendations for practice. This study draws attention to the importance that student purpose for enrolling and college possible selves when entering college has for persistence and educational decision-making. Considering the many reasons that students attend community colleges, determining why students are entering or leaving an institution can be challenging. Therefore, I recommend developing a campus-wide system for identifying students who are not enrolled for the next semester and emphasize face-to-face contact for students. Utilizing critical points of service, such as financial aid and advising, instead of relegating this task to one office and support this approach and promote a sense of shared responsibility for the support of students as they return and to avoid departure. Article 3: Mixed Methods Analysis Summary The purpose of this article was to present the results of the quantitative and qualitative studies and analyze the intersection of the data. The article focused on the following research questions: Why do students return after a period of nonenrollment? 3) Which factors are related to who returns? 174 b) What is the influence of these factors on stopout? b) How do these factors affect educational decision making? While the results reported for the quantitative portion of the study remain unchanged, the data flom the qualitative study were reanalyzed and coded to better understand and expand on the ways that the significant variables in the quantitative model appeared in and influenced decision-making among the participants in the qualitative study. The primary goal was to expand on and more fully comprehend which risk factors for departure were most relevant within the quantitative and qualitative studies and to develop a broader picture of their relationship to the departure and return to college. Previously, I reported the same quantitative results that were discussed in Article 1. Yet, the qualitative findings were different flom those presented in Article 2. I used the same data, but I specifically analyzed it for the presence and role of risk factors for departure within student educational paths. I examined how the significant variables flom the quantitative study factored into college enrollment, departure, and return. There was some overlap with the findings flom Article 2, but the emphasis was on the individual factors and their roles. Due to the fact that there were no students in the qualitative study who did not return to college, comparison of data was challenging. The qualitative data supported the quantitative results that students who utilized mixed enrollment pattems may be more likely to return to college. The participants who changed enrolhnent over the course of their educational careers reflected their recognition that previous attempts at college at their prior levels of enrollment did not contribute to (or even detracted flom) their academic success. The shift in enrollment 175 showed a willingness to adjust to a different approach and an effort to avoid past mistakes. While it was challenging to use the qualitative data to determine why divorced/widowed/separated students would be less likely to return to college, students who experienced divorce/widow/separation spoke about the event as a reason for delaying their return. In the qualitative study marriage was also a reason that students left college as students prioritized their new role of wife (generally) over student, but marriage was not significant within the quantitative model. Findings regarding initial reasons for college enrolhnent also found intersection among the qualitative and quantitative data. Operating under the assumption that students who attend college for personal enrichment value learning and see a purpose for college and that students attending with plans for a degree/certificate have a clear, it is not surprising that the qualitative findings that students with well-developed college goals were more consistent in their efforts to return to college reflects the quantitative results that students who enrolled for job skill development were less likely to return than those who enrolled for personal enrichment, transfer, or degree/certificate. Finally, in the quantitative model the minimally disadvantaged subjects were more likely to return than respondents with no disadvantage, but were less likely to return than respondents who were moderately/highly disadvantaged. This was partially mirrored within the qualitative data. The index for disadvantage is a compilation of parental education, poverty index, and high school economic status variables. For example, students whose parents encouraged them to attend or went to college were more likely to develop a clearer, internalized concept of college going . Additionally, most of those 176 students began at a four-year college and were persistent in their desire to achieve their educational goals after departure. While not included in the quantitative model, students’ financial situations also influenced their educational decisions. All of the students who delayed their college education (approximately half) were financially responsible for their college education. Also, a majority of students returned because they believed that a college education could provide a more stable career that paid better than what they could earn without an education. A significant theme within the qualitative data that was not included in the statistical model was the importance of employment. Work intensity was not included as a variable due to multicollinearilty, but appeared in students’ stories as a motivator to leave and to return. When students did not perceive college to be necessary for job pursuits, individuals often left to pursue work that they believe met their needs. Whereas, many students returned to college when their work lives became unstable or they were laid off. College became purposeful and necessary in order to find work that was stable and paid a living wage. It was important to participants, especially as parents and providers. The primary suggestion flom this article was the suggestion that an intervention study would be a next appropriate step. Across both studies the complicated and complex decision making and balancing and prioritizing of various factors within students lives is clear. Therefore, I emphasize the importance of not making assumptions of about what a student needs or can do, but working with students to identify what they need, the resources available to them, and develop a plan tailored to their needs. 177 Themes across Articles Across all elements of the study there are themes with the potential to provide greater insight educational knowledge and practice. Findings flom one part of the study complimented and extended upon the other parts. Following are themes revealed and expanded in the analyses across the study. . First, students’ external lives very much factor into their academic lives and decisions. Each article demonstrated that students’ personal and work lives were as important, if not more so, than their academic experience. Finances, work opportunities and parental and marital roles were ongoing factors in student lives that required negotiation with their academic lives. These factors influenced their attitudes toward education and college; they contributed to their decisions to leave and return to school. As students work and home lives changed, their understanding of the purpose of higher education and how it fit within their lives changed as well. Students generally left school when there was an imbalance between their academic and non-academic lives. For many, there was so little motivation to remain in school there was no effort to try to create balance. Second, the purpose that students ascribed to higher education was key in their academic choices. Students who believed college was important for their careers and their personal goals were most likely to internalize degree/ certificate attainment as a sustained aspiration. Conversely, the students who did not see a purpose for college and had poor prior educational experiences were not only likely to depart college, but depart without plans or goals to return. 178 Third, students’ decision-making about college departure and return hinged on their priorities at that time and where college-going ranked among other goals. These decisions illustrated students’ life and academic goals, their prioritization of academic goals within their lives, and where the purpose of college fell among other life goals. For example, a female student may be in community college and getting married with the goal of starting a family right away. If this student does not believe that a college degree/certificate is necessary, she may not believe that trying to balance getting married, working, and a new family is worth the effort and chose to leave college. On the other hand, another student in a similar situation may have been told that college is necessary for a good job and maintains degree/certificate attainment as a personal goal. Depending on where college falls within her priorities, this student my choose to stay in school but temporme reduce her enrollment to part-time or leave school with intentions to return or stay in school full-time and delay starting a family. The number of different combinations of factors, goals, priorities, and the conception of college purpose, makes predicting who and why students will leave very challenging. The information and themes discussed throughout these articles suggests that there are an infinite number of educational paths that students may take and developing practice based on the assumption that most those paths can be responded to uniformly ignores the complexity of students lives and their decisions. Following are suggestions for why and how practitioners might integrate an individualized approach to education within their institutions. 179 Implications and Suggestions for Practice Through all these articles I made multiple suggestions for additional research and implications for practice. In this section I bring those suggestions together for an overarching proposal as to how practitioners and institutions might approach issues of retention and attrition on their campuses. First, the finding that students’ purpose for enrolling in and conception of college influenced its priority within students’ lives has implications for how issues of retention and attrition are approached on campuses. Institutions want to have high retention and graduation/transfer rates and have students who prioritize college, yet this becomes more complicated at community colleges. Two-year colleges’ open access and multiple mission characters means that students attend community college for all types of purposes, many of which do not lead to a degree/certificate or transfer. It also means that students are able to enter and leave the institution with relative case. As a result, tracking students and knowing which students are truly departing without having achieved their educational goals is challenging. Therefore, I recommend that campus retention and persistence be flamed around the concept of purpose—not just the statistics of who comes and goes flom a campus, but why students are choosing to leave. This flamework would involve creating a culture across campus that prompts faculty and staff to explore with students the questions: “Why are you going to college?” Why is college important to you?” “What role does college play in your life?” and “Why are you leaving?” This type of inquiry prompts students to consider the purpose of college in their lives and what it might mean to depart. It also supports students in the development of clear and well- 180 developed academic/career possible selves in an effort to create internal conceptions of present and future selves that promote postsecondary education. Second, using “purpose” as a flamework for thinking about retention and college going induces people to think about retention and persistence as part of personal, individualized decisions that that students make about their education. Each student has different goals, attitudes, and sets of circumstances that contributes to their decisions to leave, stay in, or return to college. Addressing students’ different situations encourages faculty and practitioners to consider student retention flom a tailored support versus a one-intervention—fits-all approach. In light of these two recommendations for flaming persistence, I have a number of suggestions for ways to create supportive services on campus utilizing the notions of “purpose” and “individualization.” First, to create an individualized support system for a campus, institutions must create ways to identify students that are dropping or stopping out. This requires creating a system, likely coordinated with campus enrollment, that flags students who are not enrolled for the next term. Identifying why students are not enrolled and what their plans are will assist in determining which students are dropping or stopping out and which students have completed their goals. The challenge in carrying this out is in finding ways to contact students and in getting students to respond. Second, once students who intend to leave higher education without completing their educational goals are identified, I recommend approaching student support flom an individualized, but coordinated, multi-departmental approach. Students are often (a) unaware of resources on campuses, (b) unable or unwilling to advocate for themselves, and (c) overwhelmed and intimidated by the amount of bureaucracy involved with many 181 of the support services at institutions. Therefore, for students in danger of dropping out, create a campus support system where individuals or experts that represent various offices (e.g., academic support, academic departments, child care center, financial aid, advising) can come together to work with the student to identify student needs and resources and to create a plan. This approach recognizes individualized students’ purposes and goals for college going; works with students to identify the challenges in their academic, home, and work life that threatens persistence; provides an opportunity for brainstorming potential resources on campus, at home, and in the community with knowledgeable individuals; and allows for the development of a coordinated plan for the short and long term. This type of problem-solving, coordinated approach can help create a supportive plan for students, facilitate communication, and acknowledges the need to balance students’ academic, home and work lives. This type of coordinated approach is not unheard of in education. In the K-12 educational system, students needing additional support due to physical, behavioral, and developmental delays and disabilities are required to have an IEP (Individualized Education Plan). An IEP is not only a formalized plan and contract for services that will be provided by the school, but is developed at a multi-disciplinary, coordinated meeting that is attended by relevant specialists and decision-makers (e.g., psychologists, behaviorists, teacher, administrators, parents, child, audiologist, physical therapists) (Heumann, Warlick, & Richards, 2000). I am not recommending the formality or institutional responsibility involved in an IEP, but rather am advocated the concept of individualized, coordinated support for students at risk for departure. 182 Third, not all students will be retained. Some students will not be prepared or motivated to do the academic work and others will have responsibilities at work or home that will be too difficult to balance with academic study. Regardless of the reason, I recommend that the students who choose to leave college without completing their educational goals should have a plan in place for return, and the institution should have a plan for following up with these students. Finally, the recommendations I made require a significant commitment of time and effort by staff and faculty. There must be sufficient buy-in and a willingness to coordinate across campus. Therefore, I suggest a retention center on campus that is responsible for tracking student enrollment, identifying and contacting students who are potentially dropping out, and working with students to identify the reasons for leaving, purposes for higher education, potential resources for support, and plans for persistence or return. Having retention specialists responsible for these activities would ensure a coordinator who is responsible for the retention of students. That said, the activities I recommended would require a significant commitment of resources, staff, and funding. In this economic climate, this is a challenge to any institution. Yet, the focus on degree/certificate attainment of college students by the current presidential administration places this issue at the foreflont of postsecondary education’s priorities. Enrollments of traditional students have continued to grow over the years, and experts are now recommending reaching out to students who left or never attended college to achieve these attainment goals. An increase in this population would likely mean an increase in nontraditional students that will need to consider how to manage the many demands and priorities in their lives as they return to college. To avoid 183 a potential increase in departure rates as enrollment rates grow, services must be in place to support nontraditional student in continued persistence and return to college. Future Research The findings and implications throughout the parts of this study suggest possibilities for future research. First, the lack of variables available to measure possible selves or other cognitive and social emotional developmental constructs in the BPS limited my ability to test the relationship between possible selves, stopout, and characteristics related to nontraditional students. I recommend that additional surveys be developed and administered to more fully explore the relationships and influence among these factors. Furthermore, the quantitative part of this study would benefit flom analysis that included additional factors (e.g., employment intensity, academic integration, social integration, finances) that may help to explain more of the outcome. Second, the findings flom the qualitative part of the study deserve more exploration flom a qualitative and quantitative approach. The different educational paths shaped by students’ possible selves would benefit flom testing using path analysis or similar quantitative approaches. In addition, I suggest further exploration of the relationship between possible selves, learner disposition, and enrollment maintenance once students return to college. Related to this issue, the findings indicated that a shift in possible selves and learner dispositions prompted students’ return to college and persistence, therefore I recommend exploring the mechanisms (environmental and developmental) that contribute to such change. Finally, the implications for practice propose that a flamework that utilizes student purpose should be applied to better understand students’ educational decisions 184 and create individually based retention programs. Interventions and programming initiatives must be accompanied by research plans to evaluate their effectiveness and determine what is successful and ineffectual. Furthermore, such assessment can provide additional insight into the relationship between possible selves, stopout, and characteristics associated with high risk of departure that can assist institutions and practitioners with developing more influential organizational and programming strategies for persistence. Conclusion The purpose of this mixed methods study was to explore the reasons why students return after a period of nonenrollment, what factors influence this decision, how the factors influence students’ return to college, and the role that college/career possible selves play in educational decision-making. The results suggest that external (e.g., family, employment, economy) and internal factors (e.g., current and changing possible selves/learner disposition) are critical to students’ attitudes toward college and its value in their lives. The beliefs students hold about college, its purpose, and its function within their lives influences the priority they ascribe to postsecondary education in comparison to other draws on their time. This influences the decisions students make to leave and return to college and the timing of those choices. The challenges that students face in an effort to balance their educational, employment, and family priorities and make educational decisions is evident across the results. The complexity of student lives reinforces the individual and unique nature of students’ decisions. The findings yield new insight into the research questions and suggested additional approaches for practice and research related to stopout and overall student persistence. 185 APPENDIX A: FULL DESCRIPTION OF VARIABLES FOR THE QUANTITATIVE STUDY Table 13 Descriptions of Variable and Codes Variable Description Reference Stay_Stop2 Dropout/Stopout GNDERCOD Gender BLACKDUM African-American/ Black HISPADUM Hispanic ASIANDUM Asian/Pacific Islander/American Indian/Alaska Native/Other/Non- resident Alien MDISADVAN Minimally Disadvantaged HDISADVAN Moderate/High Disadvantaged MARRYDUM Married DIVORDUM Divorced/Separated/Widowed PARTDUM Part-time Enrollment MIXDUM Mixed Enrollment HSDIPL02 HS Diploma or GED/Certificate DEPENDEN Dependent or Independent JOBDUMMY Reason for Enrolling--Jobskill AACERTDM Reason for Enrolling—AA/Cert TRAN SFRDM Reason for Enrolling—Transfer GPARECOD GPA 95-96 186 Code 0/1, 1=Stopout 0/1, 1=Female 0/1, 1=AA/Black 0/1, 1=Hispanic 0/1, 1=Asian 0/1, 1=Minimally Disadvantaged 0/1, Mod/Highly Disadvantaged 0/1, 1=Married 0/1, 1=Divorced 0/1, 1=Part-time 0/1, 1=Mixed O/l, 1=GED/Cert 0/1, 1=Independent O/l , l=Jobskill 0/1 , 1=AA/Certificate 0/1, 1=Transfer 4.0 4.0-3.75 3.5 3.75-3.25 3.0 3.25-2.75 2.5 2.75-2.25 2.0 2.25-1.75 1.5 l.75-1.25 1.0 l.25-0.0 N/A Male White White White Not Disadvantaged Not Disadvantaged Single Single Fulltime Enrollment Fulltime Enrollment HS. Diploma Dependent Personal Enrichment Personal Enrichment Personal Enrichment N/A APPENDIX B: GLOSSARY AND DESCRIPTION OF VARIABLES Stopout/Dropout Binary (Dependent) Stay_Stop2 Indicates if the student left college without return between 1995 and 2001 (created flom PRENYRZB). These students were coded as Dropouts. Stopouts were identified as individuals that experienced at least one stopout (left and returned to college) (created flom PRSTNU2B). Gender GNDERCOD Gender as indicated by student. Male Female Race and Ethnicity RACECODE Responses to question: What is your race? “Other” was provided as an option to all other race/ethnicities available. “Non-resident Alien” was also an option available to students who did not identify with one of the listed categories, although a respondent could be a non-resident alien in addition to one of the race/ethnicity categories. SBRACECI was recoded to combine Asian/Pacific Islander, American Indian/Alaskan Native, Other, and Non-resident Aliens. Dummy variables were created for White, Black, Hispanic, and Asian/Pacific Islander values (BLACKDUM, HISPADUM, ASIANDUM, WHITEDUM). White, non-Hispanic A person having origins in any of the original peoples of Europe, North Afiica, or the Middle East (except those of His- panic origin). Black, non-Hispanic A person having origins in any of the black racial groups of Afiica, who is not of Hispanic origin. Hispanic A person of Mexican, Puerto Rican, Cuban, Central or South American, or other Spanish culture or origin, regardless of race. Asian/Pacific Islander A person having origins in any of the peoples of the Far East, Southeast Asia, the Indian subcontinent, or Pacific Islands. This includes people flom China, Japan, Korea, the Philippine Islands, Samoa, India, and Vietnam. American Indian/Alaskan Native A person having origins in any of the original peoples of North America and who maintains cultural identification through tribal affiliation or community recognition. I87 Other Alternative option for students who do not identify with one of the existing categories. Non-Resident Alien Socioeconomic Diversity Index DISADVAN Represents an index of socioeconomic diversity from 0-2, based on the status of students on three indicators of socioeconomic disadvantage: total family income as a percentage of the 1994 federal poverty level, the highest educational level completed by either parent, and the proportion of the student body in the student's high school eligible for the flee or reduced-price lunch program in 1994-95. If more than one indicator was missing, DISADVAN was set to missing. Dummy variables were created for each value (NDISADVAN, MDISADVAN, HDISADVAN). Not disadvantaged Minimally disadvantaged Moderately/highly disadvantaged Primary Reason For Enrolling 95 -96 ENROLRSN Respondents enrolled at 2-yr or 1ess-than-2-yr institution were asked: Are you enrolled for a job-related reason or some other reason? Or what is your primary reason for enrolling in this school? This question was not asked of students attending four-year institutions. Furthermore, the wording of this question varied depending on the student's expectations regarding the receipt of a degree flom the NPSAS institution. SBENRPUR was recoded as ENROLRSN. Dummy variables were created for each value (J OBDUMMY, AACERTDM, TRNSFRDM, ENRICHDM). Job skills Degree or certificate Transfer to a two-year school, four-year school, or not sure where Personal enrichment Marital Status 95-96 MARRYREC Marital status during month when first enrolled in 1995-1996. Variable recoded flom SBMARRYl. Dummy variables were created for each value (SINGLEDUM, MARRYDUM , DIVORDUM). Never married Student was never married. Married Student was married. Divorced/separated/widowed Student was either married, but separated flom his or her spouse, widowed, or divorced. 188 Grade Point Average 95 -96 GPARECOD Indicated student’s GPA in 1995-1996 as reported by the institution. BPS recoded the student’s GPA into a standardized 0-4.0 scale. The original variable (GPA) used the value categories of “Mostly A’s; A’s and B’s; Mostly 8’3; 8’3 and C’s; Mostly C’s; C’s and D’s; Mostly D’s or Below; and No Grades or Pass/F ail.” I recoded these categories back into their 0-4.0 scale and excluded the “No Grades or Pass/F ail” category. 4.0 4.0-3.75 3.5 3.75-3.25 3.0 3.25-2.75 2.5 2.75-2.25 2.0 2.25-1.75 1.5 l.75-l.25 1.0 l.25-0.0 Enrollment Intensity Pattern Through 2001 ENIPTT2B Indicated if the student’s primary enrollment intensity 1995-2001. Dummy variables were created flom the values (FULLDUM, PARTDUM, MIXDUM). Full-time Part-time Mixed Received Diploma or passed GED HSDIPL02 Indicated if student received a high school diploma or passed the GED, received a certificate, or did not complete high school. The original variable (HSDIPLOM) was recoded to aggregate students who completed at GED, certificate, or did not complete. High school diploma GED/certificate/did not complete Dependency Status 95-96 DEPENDEN Indicated if the student was a dependent or financially independent during the first month of enrollment in the 1995-1996 academic year. Recoded flom SBDEPlYl. No Yes Variables Explored, but Excluded flom the Model 189 Have Dependents 95 -96 KIDSRCODE The original variable (SBDPNYl) indicated the number of children the individual had, zero to nine. This was recode to reflect that either the individual did or did not have dependents. No Yes Single Parent 95 -96 SINGPAR Indicated if individual was a single parent (original variable was SBSINGYl). No Yes Delayed Enrollment After High School ENDELAY Indicated if individual delayed enrollment in postsecondary education after high school. No Yes Hours Work Per Week While Enrolled 2001 QCENRHRS Respondents answered the questions: About how many hours do you work each week while you are enrolled? Or, about how many hours did you work each week during the last term you were enrolled (as an undergraduate)? The values were a continuous scale flom 0 to 80. ‘ Remedial Courses Taken 95-96 REMEDIAL Indicated if the student took any remedial courses during the 1995-1996 school year. No Yes Received Student Loan 95 -96 AHLOANl Did the student take out any student loans for the 1995-1996 school year. No Yes 190 Risk Index 95-96/Nontraditional Status RISKINDX Based on an index of nontraditional characteristics from 0—7 composed of 7 characteristics known to be adversely related to persistence and attainment. Characteristics include delayed enrollment, no high school diploma (including GED recipients), part-time enrollment, financial independence, having dependents other than spouse, single parent status, and working full time while enrolled. The original variable (SBRSKlYl) was recoded into RISKINDX. Dummy variables were created for each value (LRISKDUM, MRISKDUM, HRISKDUM). Traditional or minimally nontraditional (Low) Student had 1 risk factor or none. Moderately nontraditional (Medium) Student had 2 or 3 risk factors. Highly nontraditional (High) Student had 4 or more risk factors. Academic Integration Index 1995-1996 ACADIN T This variable indexes the overall level of academic integration the respondent experienced at the NPSAS institution during the 1995-96 academic year. It is derived based on the average of the responses indicating how often they had done the following items: participated in study groups (CMSTUDGP), had social contact with faculty (CMSOCIAL), met with an academic advisor (CMMEET), or talked with faculty about academic matters outside of class (CMTALK). Non-missing values for these items were averaged and the average multiplied by 100. The value was continuous, ranging flom 100 to 300. Dummy variables were developed for each of the variables used to create the index. Social Integration Index 1995-1996 SOCINT This variable indexes the overall level of social integration the respondent experienced at the NPSAS institution during the 1995-96 academic year. It is derived based on the average of the responses indicating how often they had done the following items: attended fine arts activities (CMARTS), participated in intramural or nonvarsity sports (CMINTRAM), participated in varsity or intercollegiate sports (CMVARSTY), participated in school clubs (CMCLUBS), or gone places with friends flom school (CMFRIEND). Non-missing values for these items were averaged and the average multiplied by 100. The value was continuous, ranging flom 100 to 300. Dummy variables were developed for each of the variables used to create the index. 191 APPENDIX C: FREQUENCY AND CHI-SQUARE TABLES Table 14 Frequency T able for Variables (n = 7 79) Variable Number Percent Dependent Dropout 507 65 Stopout 272 35 Total 779 100 Missing 0 0 GenderMale 369 47.3 Female 410 52.7 Total 779 100 Missing 0 0 Race White 552 70.9 Black 98 12.6 Hispanic 91 11.7 Asian/Indian 38 4.8 Total 779 100 Missing 0 0 SES Index Not Disadvantaged 303 38.9 Minimally Disadvantaged 370 47.5 Moderate/Highly Disadv 99 12.7 Total 772 99.1 Missing 7 .9 Risk Index Low Risk 301 38.6 Medium Risk 249 31.9 Moderate/High Risk 212 27.2 Total 761 97.7 Missing 18 2.3 Dependency Dependent 468 60. 1 Independent 293 37.6 Total 761 97.7 Missing 18 2.3 192 Table 14 (cont’d) Delayed Enrollment No Yes Total Missing Single Parent No Yes Total Missing Dependents No Yes Total Missing HS Diploma HS Diploma GED/Cert/No complete Total Missing Marital Status Single Married Div/Sep/Widowed Total Missing Enrollment Intensity Full-time Part-time Mixed Total Missing 367 383 749 30 683 78 761 18 595 166 761 18 677 102 779 578 154 47 779 172 220 387 779 193 47.1 49.1 96.2 3.8 87.7 10 97.7 2.3 76.4 21.3 97.7 2.3 86.9 13.1 100 74.2 19.8 6.1 100 22.1 28.3 49.6 100 Table 14 (con’d) Reason to Enroll GPA Job Skills Degree/Certificate Transfer Personal Enrichment 121 Total Missing 4.0-3.75 3.75-3.25 3.25-2.75 2.75-2.25 2.25-1.75 l.75-l .25 125-00 Total Missing Remedial Courses No Yes Total Missing Received Student Loan No Yes Total Missing 180 156 253 710 69 83 91 126 114 96 43 207 779 510 195 705 74 724 55 779 15.5 194 23.1 20.0 32.5 91.1 8.9 10.6 11.6 16.0 14.6 12.4 5.5 26.6 100 65.5 25.0 90.4 9.6 93.0 7.0 100 Table 15 Chi-Square Tests for Independence: Dependent variable (Stay_Stop) and Independent Variables Independent N df Chi-Square p-value Variable Statistic Gender 779 1 1.706 .192 Race 779 3 9.361 .025* SES Index 772 2 17.406 .000* Dependency Status 762 1 7.473 .006* High School Diploma 780 1 2.226 .136 Marital Status 778 2 15.904 .000* Enrollment Intensity 780 2 43.934 .000* Reason for Enrolling 710 3 22.331 .000* GPA 759 6 17.123 .009* Note: * indicates significant, p<.05 195 APPENDIX D: RECRUITMENT POSTER Looking for Research Project Volunteers! ! What? A research project looking at the experience of students who have left college for a period of time and then re—enrolled. Volunteers will participate in a 60 minute interview on campus. Who? Potential participants must meet the following criteria: Been enrolled in college previously (any college——4-year or 2- year) and taken a leave of absence that lasted at least one semester (not including summer) to six years; They must be currently enrolled at ACC; Have academic goals of either transfer to a 4-year institution, an associate’s degree, or a certificate. When? During the Fall 2008 semester. Why? Sharing your experiences and opinions is critical to the process of helping students come back to school and learning how we can better support students. Volunteers will receive $15 for their time. If interested, please contact Casey Ozaki at ozakicar@msu.edu 196 APPENDIX E: RECRUITMENT EMAIL TO STUDENTS To I am a doctoral student flom MSU’s College of Education conducting a research study on stopout behavior—when students enroll, leave for a period of time, and re-enroll in college. This study is being conducted with the support of ACC’s Office for the Vice President of Academic Affairs. The study will consist of interviews with students who have stopped out, but are currently enrolled at ACC. The interview will focus on why students leave, why the retum, and the decision-making process to return after an extended absence. Potential participants must meet the following criteria: Been enrolled in college previously (any college—4-year or 2-year) and taken a leave of absence that lasted at least one semester (not including summer) to six years; They must be currently enrolled at ACC; Have academic goals of either transfer to a 4-year institution, an associate’s degree, or a certificate. The interview will last between approximately 60 minutes and take place on campus at a time convenient to you. Volunteers will receive $1 5 for their time. I will be asking about your college and stopout experiences. I am looking to hear about your educational story. The interview is completely voluntary and you are able to opt out of any question, part of, or the entire study at any time. The information gained flom these studies will be used to help students’ re-entry to college after a period of absence and assist them as they persist toward their particular educational goals. Sharing your experiences and opinions is critical to this process and learning how we can better support students. If you are interested and willing to participate in an interview, please email or call Casey Ozaki at ozakicgr@msu.edu. Please include the days/times you are most available. Thank you for your time and attention, Casey Ozaki Michigan State University 197 APPENDIX F: RESEARCH PARTICIPANT INFORMATION AND CONSENT FORM You are being asked to participate in a research project. Researchers are required to provide a consent form to inform you about the study, to convey that participation is voluntary, to explain risks and benefits of participation, and to empower you to make an informed decision. You should feel free to ask the researchers any questions you may have. Study Title: Reasons Students Return To College After A Period of Nonenrollment Researcher and Title: Dr. Kristen Renn, Associate Professor Department and Institution: Higher, Adult, & Lifelong Education — Michigan State University Address and Contact Information: 428 Erickson Hall, Michigan State University, East Lansing, MI 48823, 517-353-5979, renn@msu.edu. OR Researcher and Title: Casey Ozaki, Graduate Student Department and Institution: Higher, Adult, & Lifelong Education — Michigan State University Address and Contact Information: 513 Erickson Hall, Michigan State University, East Lansing, MI 48823, 517-432-2804, ozakicar@msu.edu. 1. PURPOSE OF RESEARCH: You are being asked to participate in a research study of students who started college, leave, and retum—stopout. The study will focus on why students leave, why they return, and the decision-making process to return after an extended absence. You have been selected as a possible participant in this study because you have stopped out in the past for one semester to six years, but are currently enrolled at Lansing Community College or Mott Community College. In the entire study, about 30 people are being asked to participate in interviews. Your participation in this study will take about 60 minutes. If you are under 18 you cannot be in this study. 2. WHAT YOU WILL DO: Participation in this study involves your completion of a demographic survey an in- person interview. The interview will last approximately 45 to 60 minutes. During the interview you will be asked about your educational experiences, why you left college, why you returned, what your decision-making process to return looked like, and how your academic hopes and fears factor into your decisions. With your consent, the interview will be audio recorded. Finally, at the end of the interview you will be asked to fill out a short demographic survey. The form will take about 5 minutes. If you would like to receive the findings of this study in the aggregate—meaning the findings flom across all participants in this study, rather than individual findings, please 198 contact an investigator on the team via email (ozikicar@msu.edu), and these findings will be sent to you once the project is complete. 3. POTENTIAL BENEFITS: For your participation in the study you will receive $15. Your participation in this study may also contribute to the development of policies or strategies that can help prevent students from leaving and/or ease and improve their success upon reentry to higher education. 4. POTENTIAL RISKS: The potential risks of participating in this study are psychological discomfort or distress because you will be asked questions that may make you feel uncomfortable. If you experience distress or discomfort and would like to seek professional counseling referrals can be provided by the investigator. 5. PRIVACY AND CONFIDENTIALITY: Every effort will be made to maintain your anonymity and privacy within this project. The investigator will be aware of your identity, but LCC or MCC personnel will not have information about which students chose to participate in the study. For reporting purposes you will be asked to select your own pseudonym. The list linking your pseudonym to your name will be kept in a locked filing cabinet in a locked office separate flom the data. The data will be recorded and stored on a password protected computer until it is transcribed. Once transcribed, the hard copies will be kept in a locked filing cabinet in a locked private office. Your pseudonym will be used during the interview, but your actual name and institution will not be used. Your signed consent form will be kept in another locked filing cabinet. The only people who will have access to the data are: the researchers associated with this project, and the Institutional Review Board (IRB). The results of this study may be published or presented at professional meetings, but the identities of all research participants will remain concealed with the use of a pseudonym. 6. YOUR RIGHTS TO PARTICIPATE, SAY NO, OR WITHDRAW Participation in this research project is completely voluntary. You have the right to say no. You may change your mind at any time and withdraw. If you would like to withdraw your participation alert the investigator to your wish and your participation will be stopped immediately. You will be given the original data collected flom you to that point, and you will be directed to the paper shredder so you can shred your information. You may choose not to answer specific questions or to stop participating at any time. You may also choose to continue the interview without being audio recorded at any time. 7. CONTACT INFORMATION FOR QUESTIONS AND CONCERNS If you have concerns or questions about this study, such as scientific issues, how to do any part of it, or to report an injury, please contact the researcher Dr. Kristen Renn, 428 199 Erickson Hall, Michigan State University, East Lansing, MI 48823, 517-353-5979, renn@msu.edu or Casey Ozaki, 513 Erickson Hall, Michigan State University, East Lansing, MI 48823, 517-432-2804, ozakicar@msu.edu. If you have any questions or concerns about your role and rights as a research participant, or would like to register a complaint about this research study, you may contact, anonymously if you wish, Michigan State University Human Research Protection Program at 517-355-2180, FAX 517-432-4503, or e-mail irb@msu.edu, or regular mail at: 202 Olds Hall, MSU, East Lansing, MI 48824. 12. DOCUMENTATION OF INFORMED CONSENT. Your signature below means that you voluntarily agree to participate in this research study. Signature Date You will be given a copy of this form to keep. 13. DOCUMENTATION OF AUDIO RECORDING CONSENT. Your signature below means that you voluntarily agree to be audio recorded during this research study. 200 APPENDIX G: INTERVIEW PROTOCOL AND DEMOGRAPHIC QUESTIONAIRE Interview Protocol 1. 10. ll. 12. l3. 14. 15. Tell me about your experience in school before college-«lead to other experiences What were your college experiences like? How would you describe yourself as a student in HS, college, now? What did/do you hope to be like as a student? What do you expect to be like as a student? What kind of student do/did you fear becoming? Why do you think you’ve changed? . Major baniers to returning? Educational Goals? What were your plans for achieving your goals then? Now? When you left did you have plans to come back? Career Goals? What was it like returning to college? Are there resources/experiences that have helped or you think would have helped your transition back? Is there anything you I haven’t asked you about that would help me understand your story better? 201 10. ll. 12. l3. 14. 15. 16. 17. Demographic Questionnaire Sex Age . Race or ethnicity Do you work full- or part-time? Are you married? Do you have children? Are you financially independent? . Are you a full-or part-time student? What are your educational goals? What is your (anticipated) major? How many credits did you earn? What is your cumulative GPA? What is your mother’s highest level of education? Father? How would you describe your social class growing up? Did you graduate flom high school? If so, when? If not, did you earn your GED? When? When did you first enroll in college? Have you taken more than a semester off of college? Please list each period of nonenrollement—when left college and when returned. 202 REFERENCES Adelman, C. (1997). Diversity: Walk the walk and drop the talk. Change, 29(4), 34-45. Adelman, C. (1999). Answers in the tool box: Academic intensity, attendance patterns, and Bachelor ’3 degree attainment. J essup, MD: US. Department of Education, Office of Educational Research and Improvement. Ahson, N. S., Gentemann, K. M., & Phelps, L. (1998). Do stopouts return? A longitudinal study of re-enrollment, attrition, and graduation. A paper presented at the Annual Forum of the Association for Institutional Research, Minneapolis, MN. Allison, PD. (2001). Missing data. In M.S. Lewis-Beck (Series Ed.), Quantitative applications in the social sciences. Thousand Oaks, CA: Sage Publications. Barr, J. (2007). Freshmen dropouts. Journal of Applied Research in the Community College, 14(2), 105-1 13. Baxter Magolda, MB. (2001). Making their own way: Narratives for transforming higher education to promote self-development. Sterling, VA: Stylus. Bean, J ., & Eaton, S. B. (2001/2002). The psychology underlying successful retention practices. Journal of College Student Retention, 3(1), 73-89. Bean, J. P., & Eaton, S. B. (2000). A psychological model of college student retention. In J. M. Braxton (ed.), Reworking the student departure puzzle (pp. 48-61). Nashville, TN: Vanderbilt University Press. Beil, C., Reisen, C. A., Zea, M. C., Caplan, R. C. (1999). A longitudinal study of the effects of academic and social integration and commitment on retention. NASPA Journal, 3 7(1), 376-85. Berkner, L., Horn, L., & Clune, M. (2000). Descriptive summary of 1995-96 Beginning Postsecondary Students: Three years later (N CES 2000-154). Washington, DC: US. Department of Education, National Center of Education Statistics. Bloomer, M. (1996). Education for studentship. In J. Avis, M. Bloomer, G. Esland, D. Gleeson, & P. Hodkinson, Knowledge and nationhood: Education, politics and work (pp. 140-163). London: Cassell. Bloomer, M. (1997). Curriculum making in post-16 education: The social conditions of studentship. New York: Routledge. 203 Bloomer, M., & Hodkinson, P. (1997). Moving into FE: The voice of the learner. London: Further Education Development Agency. Bloomer, M., & Hodkinson, P. (1999). College life: The voice of the learner. London: Further Education Development Agency. Bloomer, M., & Hodkinson, P. (2000). Learning careers: Continuity and change in young people’s dispositions to learning. British Educational Research Journal, 26(5), 583-597. Bonham, L. A., & Luckie, J. A. (1993a). Taking a break in schooling: Why community college students stop out. Community College Journal of Research and Practice, 1 7(3 ), 257. Bonham, L. A., & Luckie, J. I. (1993b). Community college retention: Differentiating among stopouts, dropouts, and optouts. Community College Journal of Research and Practice, 1 7, 543-554. Borglum, K., & Kubala, T. (2000). Academic and social integration of community college students: A case study. Community College Journal of Research and Practice, 24, 567-576. Bourdieu, P. (1987). Distinction: A social critique of the judgment of taste. Boston: Harvard University Press. Bradbum, E. M., Berger, Rachael, Li, X., Peter, K., & Rooney, K. (2003). Baccalaureate and beyond: A descriptive summary of 1999-2000 Bachelor ’s degree recipients, 1 year later—with an analysis of time to degree (N CES 2003-165). Bragg, DD. (2001). The past, present, and future role of federal vocational legislation in US. community colleges. Journal of Applied Research in the Community College, 9(1), 57-76. Braxton, J. M., & Lien, L. A. (2000). The viability of academic integration as a central construct in Tinto’s interactionalist theory of college student departure. In J. M. Braxton (ed.), Reworking the student departure puzzle (pp. 11-28). Nashville, TN: Vanderbilt University Press. Bryant, A.N. (2001). ERIC review: Community college students: Recent findings and trends. Community College Review, 29(3), 77-93. Burley, H., Butner, B., & Cejda, B. (2001). Dropout & stoput patterns among developmental education students in Texas. Community College Journal of Research and Practice, 25, 767-782. 204 Cabrera, A. F., Nora, A., Castaneda, M. B. (1993). College persistence: Structural equations modeling test of an integrated model of student retention. The Journal of Higher Education, 64(2), 123-139. Cameron, A.C., & Windmeijer, F.A.G. (1997). An R-squared measure of goodness-of-fit for some common nonlinear regression models. Journal of Econometrics, 77(2), 329-342. Choy, S. P. (2001). Students whose parents did not go to college: Postsecondary access, persistence, and attainment (NCES 2001-126). Washington, DC: US. Department of Education, National Center of Education Statistics. City of Flint. (n.d.). Quick facts. Retrieved on 10/15/08, flom www.cityofflint.com. Cofer, J ., & Somers, P. (2001). What influences student persistence at two-year colleges? Community College Review, 29(3), 56-76. Coleman, IS (1988). Social Capital in the Creation of Human Capital. American Journal of Sociology ,94, S95-S120. Conklin, K.A. (1993). Leaving the community college: Attrition, completion, or something else? Community College Journal of Research and Practice, 1 7, 1-11. Creswell, J. W., & Plano Clark, V. L. (2007). Designing and Conducting Mixed Methods Research. Thousand Oaks, CA: Sage Publications. Creswell, J. W., Plano Clark, V. L., Gutrnann, M. L., Hanson, W. E. (2003). Advanced mixed methods research design. In A. Tashakkori & C. Teddlie (eds.), Handbook of Mixed Methods in Social & Behavioral Research (pp. 209-240). Thousand Oak, CA: Sage Publications. Cross, S. E., & Markus, H. R. (1994). Self-schemas, possible selves, and competent performance. Journal of Educational Psychology, 86(3), 423-438. Dandaneau, SP. (1996). A town abandoned: Archer, XXX, confronts deindustrialization. New York: SUNY. DesJardins, S. L., Ahlburg, D. A., & McCall, B. P. (1994). Studying the determinants of student stopout: Identifying “true” flom spurious time-varying effects. Paper presented at the 34th Annual Forum of the Association for Institution Research, New Orleans, LA. DesJardins, S. L., Ahlburg, D. A., & McCall, B. P. (1999). An event history model of student departure. Economics of Education Review, 18, 375-390. 205 DesJardins, S. L., Ahlburg, D. A., & McCall, B. P. (2002). A temporal investigation of factors related to timely degree completion. The Journal of Higher Education, 73(5), 555-581. DesJardins, S. L., Ahlburg, D. A., & McCall, B. P. (2006). The effects of interrupted enrollment on graduation flom college: Racial, income, and ability differences. Economics of Education Review, 26(6), 575-590. de Vaus, D. (2004). Surveys in social research (5th ed.). In M. Bulmer (Series Ed.), Social research today. New York: Routledge. Diemer, M. A. (2008). Large scale survey data in career development research. Journal of Career Development, 35(1), 42-59. Dowd, AC. (2004). Income and financial aid effects on persistence and degree attainment in public colleges. Education Policy Analysis Archives, 12(21). Retrieved on February 14, 2009 flom http://epaa.asu.edu/epaa/v12n21/. Erzberger, C., & Prein, G. (1997). Triangulation: Validity and empirically based hypothesis construction. Quality & Quantity, 2, 141-154. Ewell, P. T., & Boeke, M. (2007). Critical connections: Linking states ’ unit record systems to track students ’ progress. New agenda series, Lumina foundation. Indianapolis, IN: Lumina Foundation. Feldman, M. J. (1993). Factors associated with one-year retention in community college. Review of Higher Education, 34(4), 504. Grosset, J. M. (1993). A profile of community college stop-outs. Community College Review, 20(4), 51-58. Grubb, N. (1999). Honored but invisible. New York: Routledge. Hagadorn, L. S., Siadat, M.V., Fogel, S.F., Nora, A. & Pascarella, ET. (1999). Success in college mathematics: Comparisons between remedial and non-remedial first-year college students. Research in Higher Education, 40(3), 261-285. Hawley, T.H., & Harris, T.A. (2005/2006). Student characteristics related to persistence f0 first-year community college students. Journal of College Student Retention, 7(1-2), 117-142. Heam, J.C., McLendon, M.K., & Mokher, CG. (2008). Accounting for student success: An empirical analysis of the origins and spread of state student unit-record systems. Research in Higher Education, 49, 665-683. 206 Herzog, S. (2004). Measuring determinants of student return vs. dropout/stopout vs. transfer: A first-to-second year analysis of new fleshmen. Paper presented at the Annual Forum of the Association for Institutional Research (AIR) in Boston, MA. Heumann, J .E., Warlick, K., & Richards, C. (2000). A guide to an individualized education plan. J essup, MD: Office of Special Education and Rehabilitative Servies, US. Department of Education. Hoachlander, G., Sikora, A. C., Horn, L. (2003). Community college students: Goals, academic preparation, and outcomes (NCES Rep. No. 2003-164). US. Department of Education, National Center of Education Statistics, Washington, D.C.: U.S. Printing Office. Horn, L. J. (1996). Nontraditional undergraduates (NCES Rep. No. 97-578). US. Department of Education, National Center of Education Statistics, Washington, D.C.: U.S. Printing Office. Horn, L .J. (1998). Stopouts or stayouts? Undergraduates who leave college in their first year (NCES Rep. No. 90-087). US. Department of Education, National Center of Education Statistics, Washington, D.C.: U.S. Printing Office. Horn, L. J ., & Nevill, S. (2006). Profile of undergraduates in US. postsecondary education 2003-04: With a special analysis of community college students (NCES Rep. No. 2006-184). US. Department of Education, National Center of Education Statistics, Washington, D.C.: U.S. Printing Office. Horn, L., & Premo, M. (1993). Profile of undergraduates in US. postsecondary education institutions: 1992 —93. Washington, D.C.: US Department of Education, National Center for Education Statistics. Hosmer, D.W., & Lemeshow, s. (2000). Applied logistic regression (2“ ed. ). New York: Wiley-Interscience. Hoyt, J .E., & Winn, BA. (2004). Understanding retention and college student bodies: Differences between drop-outs, stop-outs, opt-outs, and transfer-outs. NASPA Journal, 41(3), 395-417. Jacobson, J., Olsen, C., King Rice, J., Sweetland, S., & Ralph, J. (2001). Educational achievement and black-white inequality ((NCES 2001-061). Washington, D.C.: US. Department of Education, National Center for Education Statistics. Johnson, I. Y. (2006). Analysis of stopout behavior at a public research university: The multi-spell discrete-time approach. Research in Higher Education, 4 7(8), 905- 934. 207 Kim, D.H., & Schneider, B. (2005). Social capital in action: Alignment of parental support in adolescents’ transition to postsecondary education. Social Forces, 84(2), 1181-1206. Kerpelman. J. L., Shoffner, M. F., & Ross-Griffin, S. (2002). African American mothers’ and daughters’ beliefs about possible selves and their strategies for reaching the adolescents’ future academic and career goals. Journal of Youth and Adolescence, 31(4), 289-302. KewalRamani, A., Gilbertson, 8., Fox, M. A., & Provasnik, S. (2007). Status and trends in the education of racial and ethnic minorities (N CES Rep. No. 2007-039). Washington, D.C.: US. Department of Education, National Center for Education Statistics. Lee, M. M. (1996). Student retention survey: Why students did not return, Spring semester 1994. Valhalla, NY: Westchester Community College, Office of Institutional Research (ERIC No. ED391565). Lee, S. J ., & Oyerman, D. (2007). Reaching the future: The education-focused possible selves of low-income mothers. New Directions for Adult and Continuing Education No. 114 (pp. 39-50). Washington D.C.: Association for the Study of Higher Education. Leondari, A. (2007). Future time perspective, possible selves, and academic achievement. New Directions for Adult and Continuing Education No. 114 (pp. 17-26). Washington D.C.: Association for the Study of Higher Education. Leondari, A., Syngollitou, E., & Kiosseoglou, G. (1998). Academic achievement, motivation and future selves. Education Studies, 24(2), 153-163. Light, A. (1993). Hazard model estimates of the decision to reenroll in school. Labour Economics, 2, 381-406. Lincoln, Y., & Guba, E.G. (1985). Naturalistic inquiry. Thousand Oaks, CA: Sage Publications. Lumina Foundation. (2008). Memo to the Presidential Transition Team December 15, 2008. Retrieved on 3/17/08 from http://www.luminafoundation.org/our_work/memo to_the_presidential_transition team.html. Markus, H., & Nurius, P. (1986). Possible Selves. American Psychologist, 41(9), 954- 969. McDonough, P. (1997). Choosing colleges: How social class and schools structure opportunity. Albany, NY: SUNY Press. 208 McNeal, R.B., Jr. (1999). Parental involvement as social capital: Differential effectiveness on science achievement, truancy, and dropping out. Social Forces, 78, 117-44. Menard, S. (2002). Applied logistic regression analysis (2"d ed.). In M.S. Lewis-Beck (Series Ed.), Quantitative applications in the social sciences. Thousand Oaks, CA: Sage Publicaitons. Merriam, SB. (1998). Qualitative research and case study applications in education. San Francisco: Jossey-Bass Publishers. Miles, M.B., & Huberman, A.M. (1994). Qualitative data analysis (2”! ed. ). Thousand Oaks, CA: Sage Publications. Mohammadi, J. (1994). Exploring retention and attrition in a two-year public community college. Martinsville, VA: Patrick Henry Community College. Morse, J. M. (2003). Principles of mixed methods and multimethod research design. In A. Tashakkori & C. Teddlie (eds.), Handbook of Mixed Methods in Social & Behavioral Research (pp. 189-208). Thousand Oak, CA: Sage Publications. Napoli, A. R., & Wortman, P. M. (1996). A meta-analysis of the impact of academic and social integration of persistence community college students. Journal of Applied Research in the Community College, 4(1), 5-21. Napoli, A. R., & Wortman, P. M. (1998). Psychosocial factors related to retention and early departure of two-year community college stopouts. Research in Higher Education, 39(4), 419—455. Nora, A., Attinasi, In, L. C., & Matonak, A. (1990). Testing qualitative indicators of precollege factors in Tinto’s attrition model: A community college student population. The Review of Higher Education, 13(3), 337-356. Norman, C.C., & Aron, A. (2003). Aspects of possible self that predict motivation to achieve or avoid it. Journal of Experimental Social Psychology, 39, 500-507. Obama, B. (2009). Remarks of President Barack Obama — As Prepared for Delivery Address to Joint Session of Congress. Retrieved on 3/17/09 from http://www.whitehouse.gov/the_press_office/rema_r_‘l_