INSTITUTIONAL CHARACTERISTICS INFLUENCING NONBLACK ENROLLMENT AT HISTORICALLY BLACK COLLEGES AND UNIVERSITIES IN THE TWENTY -FIRST CENTURY By Charles Robert Shorette II A DISSERTATION Submitted to Michigan State University in partial fulfillmen t of the requirements for the degree of Higher, Adult, and Lifelong Education Ð Doctor of Philosophy 2015 !!ABSTRACT INSTITUTIONAL CHARACTERISTICS INFLUENCING NONBLACK ENROLLMENT AT HISTORICALLY BLACK COLLEGES AND UNIVERSITIES IN THE TWENTY -FIRST CENTURY By Charles Robert Shorette II In order to better understand diversity on HBCU campuses and broaden the scope of scholarly work on HBCUs, this study attempts to provide a nuanced view of nonblack enrollment at HBCUs from an organizational perspect ive by pursuing the primary research question: Are there institutional characteristics of HBCUs that influence nonblack enrollment patterns between the years 2000 -2010? PernaÕs (2006) conceptual model of college choice, specifically the higher education co ntext, and the education production function frameworks serve as lens es through which this issue will be viewed. This investigation uses data from the Integrated Postsecondary Education Data System (IPEDS) and employs a fixed effects panel data regression analysis to examine enrollment patterns in hopes of better understand ing an element of college choice. The findings from this study suggest that institutional characteristics influence nonblack enrollment differently depending on institutional control. For public HBCUs, spending on academic support per FTE student was found to be a strongly positive predictor of nonblack enrollment, whereas graduation rates had statistically significant negative effects on nonblack enrollment at private HBCUs. Consequently, implications for policy, practice, and future research are discussed . !!iii I dedicate this dissertation to my father, Charles Robert Shorette, and my mother, Karen Rae Shorette, who truly e mbody selflessness and unconditional love, and wh o made countless sacrifices with my future in mind so that I may pursue my passions freely. !!iv ACKNOWLEDGEMENTS Throughout my academic journey, I have learned a lot. Mostly, I have learned that I do not know much. But there is one thing that I k now for sure: I could not have done this alone and owe my success to a support system that is second to none. First and foremost, my family has been the constant source of stability and love whenever I was in need. My dad was always happy to hear of the pr ogress I was making and was overjoyed at the thought of his son receiving a doctoral degree. My mom has essentially read and provided editorial feedback on every paper I have written since my days as an English Education major at Florida A&M University, so , in my eyes, she has earned honorary bachelorÕs, masterÕs, and doctoral degrees. On top of her skills as a copy editor, even though she could not always relate to exactly what I was going through, she always knew what I needed to hear to get me through tough times. My brother has always set an example for me to follow and paved the way as a model student, husband, father, and brother. And my wife, April, while simultaneously overcoming her own challenges in becoming a doctor (the Juris Doctor kind), still dedicated every additional moment not being spent on her own pursuits to supporting mine. To my family, I do not have words to express how much I love you and how grateful I am to have you in my life. I am so lucky to have friends who have such confidence in me and truly believe that I will succeed. Each one served a special role in my life as I have grown intellectually; but more importantly, they have stuck with me as I have attempted to be as good of a friend to them as they are to me. If you are reading this, then you know who you are and I donÕt have to name each and every one of you. I feel so fortunate to include in that group of friends my 2010 Michigan !!v State University HALE cohort, as well as my HALEmates who came before and after us. I constantly s ing the praises of my classmates (including those from other programs within the College of Education), for they are diverse, passionate, compassionate, friendly, inspiring, and supremely intelligent. I was never disappointed by the fact that I was never t he smartest one in the room. In fact, I took pride in being included in such an exceptional group of thinkers and doers, and always say that I learned as much from my classmates as I learned from my professors Ñand that is saying a lot when your professors are the faculty for a top nationally ranked graduate program. From advice on how to navigate the doctoral program to intellectual chats over coffee to the nights at RenoÕs when we just needed to unwind, I could always count on my classmates to be there. Along with my classmates at Michigan State, I am so grateful for the relationships I developed with my faculty members in the Higher, Adult, & Lifelong Education program and across the College of Education. The instruction, guidance, and mentorship I receiv ed during my time at Michigan State were invaluable. My faculty members were concerned about my development as a scholar and practitioner, but they were just as concerned about my development as a human being. The Michigan State College of Education truly exemplifies what it means to be a Òcommunity of scholars.Ó I know that my doctoral journey would not have been the same without the Michigan State community. I am particularly thankful for my committee. Each of them has contributed to my development in uni que ways and served as a model advisor, scholar, and person for me to emulate. Thank you to my chair, Brendan Cantwell, for not only taking me on as one of his first advisees, but for demonstrating the persistence and patience that it takes to handle advis ing me. The same can certainly be said of the rest of my committee, !!vi Marilyn Amey, Kris Renn, and Ken Frank, for their dedication to my personal, professional, and academic development. Outside of my immediate Michigan State community, I was fortunate to h ave mentors and colleagues who were always willing and ready to support me in my professional and academic journeys. I do not think I would have been at Michigan State without Billy Molasso, a HALE alum who was my faculty advisor at GW, who suggested that I look into the PhD program at Michigan State, and who did everything he could to set me up for success upon graduating from my masterÕs program. When I first committed to Michigan State, I was excited at the opportunity to work with James Minor. Sadly (fo r me), James accepted a professional opportunity in Atlanta and I thought I was out of luck. However, James did not abandon me. Even from afar, he provided guidance and mentorship, and provided me an opportunity to work with him at the Southern Education F oundation as a doctoral intern in the summer of 2012. Similarly, when I left Washington, D.C., to pursue my doctoral studies in Michigan, I knew I was leaving behind good colleagues and friends like Brian Sponsler. Much like James, though, Brian did not al low distance to affect the support I received from him. He always looked out for me and I was lucky to have worked under his guidance at the Institute for Higher Education Policy during the summer of 2011. In addition to those I worked with in formal capa cities, I also had so many people who saw something in me and agreed to collaborate with me on various projects. To Marybeth Gasman, thank you for taking a chance on me, supporting another HBCU advocate, and introducing me to writing for non -academic audie nces. To Rob Palmer, thank you for inviting me to work with you on projects that proved to be very important in my development as a researcher. To Andrew Arroyo, thank you for not only being quite possibly the most flexible and !!vii understanding person I have worked with and making research fun, but for also being someone to bounce ideas off of during an important stage of my academic journey. To Derek Greenfield, thank you for being an honest mentor, a caring friend, and a great thinker who has opened my mind to new ways of thinking about issues I care deeply about. To Mike Broda, thanks for being a great co -worker, classmate, and a source of quantitative comfort when I really needed it. Finally, I want to acknowledge the very important role my undergraduate an d graduate institutions played in my life. Florida A&M University is where I found my passion for education and social justice. The George Washington University is where I discovered my love for higher education and developed as a practitioner. Michigan St ate University is where I refined my focus and learned how to apply my passions in my research and professional endeavors. I will forever be grateful for the investment those institutions made in me. All I hope is that those around me will receive the retu rn from that investment. !!viii TABLE OF CONTENTS LIST OF TABLES ......................................................................................................................... x LIST OF FIGURES ...................................................................................................................... xi Chapter 1 Introduction ................................................................................................................... 1 Background .............................................................................................................................. 1 Description of Study ................................................................................................................ 9 Chapter 2 Literature Review ........................................................................................................ 14 Minority -Serving Institutions (MSIs) .................................................................................... 14 Historically Black Colleges and U niversities (HBCUs) ........................................................ 16 Racial Diversity at HBCUs .................................................................................................... 24 College Choice ....................................................................................................................... 28 Chapter 3 Research Design and Methodology ............................................................................. 36 Conceptual Framework .......................................................................................................... 36 College Choice ................................................................................................................. 36 Education Production Function ........................................................................................ 37 Conclusion ....................................................................................................................... 39 Data Source ............................................................................................................................ 40 Measures ................................................................................................................................ 41 Sample .................................................................................................................................... 44 Research Method ................................................................................................................... 48 Limitations ............................................................................................................................. 51 Conclusion ............................................................................................................................. 54 Chapter 4 Findings ....................................................................................................................... 55 Descriptive Statistics .............................................................................................................. 55 Results .................................................................................................................................... 57 Public HBCUs .................................................................................................................. 61 Interpretation ............................................................................................................... 61 Private HBCUs ................................................................................................................. 62 Interpretation ............................................................................................................... 63 Impact of Confounding Variable Threshold .................................................................... 64 Lagged V ariable Analysis ................................................................................................ 65 Conclusion ............................................................................................................................. 69 Chapter 5 Discussion and Implications ........................................................................................ 71 Summary of Study ................................................................................................................. 71 Discussion of Findings ........................................................................................................... 75 Implications of Findings ........................................................................................................ 85 Future Research ............................................................................................................... 85 !!ix Policy and Practice ........................................................................................................... 89 Conclusion ............................................................................................................................. 94 APPENDIX .................................................................................................................................. 96 BIBLIOGRAPHY ........................................................................................................................ 99 !!x LIST O F TABLES Table 1 . Description of Predictor Variables Included in S tudy ................................................... 44 Table 2 . Description of Institutions Included in S tudy ................................................................ 45 Table 3. Descriptive Statistics for Institutional Characteristics of Four -Year HBCUs Between 2000-2010 .................................................................................................................................. 57 Table 4 . Regre ssion Analysis for the Effects of Institutional Characteristics o n Nonblack Enrollment at HBCUs B etween 2000 -2010 (No Lags) ............................................................... 60 Table 5 . Regression Analysis for the Effects of Institutional Characteristics o n Nonblack Enrollment at HBCUs B etween 2000 -2010 with One - and Two -Year Lags ............................... 67 Table 6 . All Models with Emphasis on Variables Found to be Statistically Significant in Two Out of the Three Models Between 2003 -2010 ............................................................................. 70 Table 7. IPEDS Definitions for Variables Included in Study ...................................................... 97 !!xi LIST OF FIGURES Figure 1 . Total N onblack Enrollment at All Four -year HBCUs (2000 -2010) ............................... 3 Figure 2 . Percentage of Total Enrollment at Four -year HBCUs (2000 -2010) .............................. 4 Figure 3 . Percent of Undergraduate Enrollment at Four -year Public and Private N ot-for -profit HBCUs in 2010 ............................................................................................................................ 28 Figure 4 . Laura PernaÕs Conceptual Model of College Choice ................................................... 34 Figure 5 . Conceptual Model of Nonblack HBCU Student Choice .............................................. 40 Figure 6 . Influential Institutional Characteristics for Nonblack Students Attending HBCUs (Public vs. Private) ....................................................................................................................... 78 !!1 Chapter 1 Introduction Background For more than a century, historically black colleges and universities (HBCUs) have been at the center of black intellectual development in the U.S. (Gasman, 2008b ; Exkano, 2012 ). Decades of research has demonstrated that HBCUs have pr ovided opportunities for black students to attend college that would not have otherwise existed but for HBCUs, graduated black students at higher rates than their predominantly white institution (P WI) counterparts, and produced a disproportionately large s hare of black graduates at both the baccalaureate and post -baccalaureate level (Bennett & Xie, 2003; Ehrenberg & Rothstein, 1993; Fleming, 1984; Kim & Conrad, 2006; Redd, 1998). Beyond their academic achievements, HBCUs have also played a major role in adv ancing civil rights and serving their communities (Gasman, 2008a; Davis, 2012). Despite the academic and social accomplishments of HBCUs, their place in the U.S. higher education system is often challenged. It is not uncommon to come across an article, op -ed, or radio segment (e.g., on National Public Radio) asking the question: ÒDo we still need HBCUs?Ó (Martin, 2013). These discussions are typically initiated by people who are concerned that HBCUs encourage racial segregation, no longer serve a purpose , and hinder broader diversity efforts taking place in PWIs (Connerly, 2003; Vedder, 2010; Martin, 2013). As a result of the little knowledge most people possess about HBCUs, many have formed misconceptions about these institutions. As well as the challenge HBCUs face educating the general public about their role in society, Sims (1994) suggests that HBCUs also Òhave continued to maintain their !!2 segregated campuses not necessarily out of choice, but because of their inability to attract white studentsÓ and oth er nonblack students (Sims, 1994, p. ix). Considering the relatively low nonblack enrollment at HBCUs, some might presume that contemporary HBCUs resemble the HBCUs of old. However, the enrollment patterns of these institutions have undergone major change s since their beginnings. In 1954, Brown v. Board of Education ended the legal practice of segregation in public education. This decision was the impetus for a dramatic migration of black students away from HBCUs to PWIs . In fact, it took only 20 years for the shift from black students being almost exclusively enrolled in HBCUs to over three -quarters of all black students in the U.S. hig her education system attending P WIs to occur (Allen, 1992). This trend has continued in recent years (Mercer & Stedman, 20 08). Between 2000 and 2010, although total black enrollment in all higher education institutions increased by nearly 70% (over 1,000,000 additional black students), the majority of b lack students chose institutions other than HBCUs ( Delta Cost Project, 201 4). This continued migration of black students to non -HBCU institutions is evident when considering that the share of African -American students enrolled in P WIs increased from approximately 87% in 2000 to 91% in 2010 ( Delta Cost Project, 2014 ). Figure 1 a nd Figure 2 show that, in the aggregate, four -year HBCUs have also experienced stagnant nonblack enrollment over the last decade in terms of total number of nonblack students and a slight decline in percentage of nonblack enrollment at HBCUs (Gasman, 2008b; Delta Cost Project, 2014 ; Shorette & Arroyo, 2015 ). The stagnant levels of nonblack enrollment are cause for concern for two reasons. First, not only are HBCUs not keeping pace with non -HBCU higher education institutions when it comes to general enrollme nt, but they are also losing significant ground when it comes to white and Latino student enrollment. !!3 Considering that Hispanic/Latino students now represent over 11% of the total college -going population nationally (an approximately 3% increase since 2000 ) and with nearly 70% of 2012 Hispanic/Latino high school graduates attending college, the .5% increase in the share of Hispanic/Latino enrollment at four -year HBCUs between 2000 and 2010 seems to demonstrate a disconnect between HBCU enrollment patterns a nd the national demographic shifts that are occurring in the U.S. (Delta Cost Project, 2010; Santiago & Reindl, 2009; Roach, 2013). Furthermore, many HBCUs have become highly tuition -dependent due to historical underfunding by state governments and low lev els of alumni giving, so even the slightest decline in enrollment has a significant impact on operating budgets (Hernandez, 2010). Figure 1. Total N onblack Enrollment at All Four -year HBCUs (2000 -2010) 22,000 22,500 23,000 23,500 24,000 24,500 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 !!4 Figure 2. Per centage of Total Enrollment at Fo ur-year HBCUs (2000 -2010) Few institutions of higher education have been immune to the general decline in state support. In fact, many public research institutions experienced an increase in enrollment while receiving significantly less per -student fundi ng from the state (National Science Board, 2012). However, the struggle for HBCUs to secure equitable funding has been an ongoing uphill battle that extends beyond the regular cyclical challenges faced by non -HBCU institutions. Samuels (2004) noted that in 1933, instead of HBCU land -grant institutions receiving the $2,293,573 they deserved according to the federal funding formula, they received Òa mere $77,99 5Ó (p. 37). HBCUs in the post -civil r ights era have seen little improvement. Federal funding for al l institutions increased 40 percent between the years of 1993 to 2002, while federal funding for HBCUs increased only 24 percent (Minor, 2008a; White House Report, 2005). Boland and Gasman (2014) recently updated MinorÕs (2008a) work documenting contempora ry inequities at the state level. In North Carolina, for example, Minor (2008a) reported that two predominantly 83.9 84.9 10.1 8.8 3.5 6.3 2000 2010 Black enrollment Nonblack enrollment Unknown and nonresident alien !!5 white universities each were receiving approximately $15,700 per student in state allocations while two HBCUs were each receiving $7,800 per stu dent. Boland and Gasman (2014) found that the trend in inequitab le funding in North Carolina is still present, noting that even Òthe highest per FTE HBCU (Winston -Salem University at $10,618 in 2011), is still nearly half that of UNC Chapel Hill ($17,992) and North Carolina State University ($15,558)Ó (p. 8). Considering the cumulative effect of the enrollment and funding challenges presented above, some assert that these institutions can no longer afford to be perceived as places exclusively for black st udents. Gasman (2012) is explicit in the reason why she believes HBCUs must approach fulfilling their institutional missions differently: I no longer think that HBCUs have a choice as to whether or not they should actively reach out to students of all raci al and ethnic backgrounds. The numbers are clear. African -Americans have options, and most are choosing to attend majority institutions. If HBCUs are to survive and thrive in the 21st century, they must fully embrace others. (paragraph 7) GasmanÕs (2012) a ssertion underscores how essential it is that HBCUs un derstand their relationship to nonblack populations and how to utilize that understanding to position themselves for a sustainable future in the U.S. higher education system. Financial stress associate d with low enrollment at HBCUs, particularly any stress directly associated with nonblack enrol lment, is not the only cause for concern, however. Nonblack enrollment that contributes to the diversity of HBCUs is also often directly tied to their educationa l missions; and although much attention has been placed on diversity in higher education, the conversation has been fairly limited to P WIs. The fact that the attentio n has been placed primarily on P WIs is not surprising when considering the disproportionat e percentage of !!6 black students enrolled at P WIs. Research on diversity in higher education has ranged from campus climate assessments for underrepresented students to measures of the educational benefits of diversity programs for all students to challenges students of color face accessing higher education, just to name a few examples (Gurin, Dey, Hurtado, & Gurin, 2002; Gurin, Nagda, & Lopez, 2004; Rankin & Reason, 2005; Roderick, Coca, & Nagoaka, 2011). A common thread among most of the research is that it is focused on these issues of diver sity at P WIs. Jewell (2002) suggests that HBCUsÕ absence from the diversity discussion is ironic considering their noble history in providing an environment conducive to cross -cultural understanding. Dr. Charlie Nelms (2 010), chancellor of North Carolina Cen tral University, similarly notes , ÒHBCUs have never discriminated and are poised to become the best hope for access and success for all disenfranchised populations, irrespective of race and ethnicity.Ó Arguing for HBCU sÕ rightful place in the diversity con versation, Jewell (2002) offers the following sentiments: By rights, they [HBCUs] should occupy a leading position in such discussions, offering the insight that they have gained from their past and reliving those les sons as they contemplate their present and future Ñone in which HBCUs should consider themselves uniquely called upon to provide leadership and to make important contributions in the ongoing quest for a truly inclusive society. (p. 8) Within the scholarly discussion on HBCUs, few have attempted to address issues of diversity or the status of nonblack students. More recently, leaders in the HBCU community have brought issues of diversity at HBCUs to the forefront. Some HBCU presidents, such as former preside nt of Alcorn State University in Mississippi M. Christopher Brown II and !!7 President Michael Sorrell at Paul Quinn College in Texas, have been outspoken in their support of HBCUs embracing diversity. Regarding their own institutions, President Sorrell argues , ÒIf all you see when you look at us is a school for black people, you miss what makes us specialÓ (Sorrell, 2012, paragraph 3) and M. Christopher Brown II expresses his discontent with the diversity efforts at his institution when he says he is not convi nced that Alcorn State has Òaddressed all of the issues of underrepresented groups on campus, not only that those groups begin to appear but that they feel they have community and a voice on our campusÓ (Stewart, 2012, paragraph 3). Scholars have also urge d HBCU leaders, and society more broadly, to view HBCUs as exemplary institutions of higher learning for students from all backgrounds due to their history of successfully providing racially inclusive environments (Gasman, 2012; Gasman & Shorette, 2012). Exploring why nonblack students enroll at HBCUs is critically important to better understanding the college -choice process for all students. A limited number of peer -reviewed studies that are now outdated found that white students attending HBCUs are genera lly influenced by unique academic program offerings, student financial aid/lower cost of attendance, and location (Brown & Stein, 1972; Conrad, Brier, & Braxton, 1997; Nixon & Henry, 1990). Aside from the aforementioned studies, little research has been co nducted examining the college choice of nonblack students at HBCUs. The literature is particularly scarce as it relates to the higher education institutionÕs role in influencing nonblack enrollment patterns (Daniels, 2008). And because more comprehensive and contextual conceptual models of college choice did not exist until more recently, what we do know about the enrollment of nonblack HBCU students is not expressed in terms that are aligned with more contemporary knowledge of the college -choice process. !!8 In addition to the limited understanding of college choice for nonblack students at HBCUs, little is known about organizational behavior that may influence the college -choice decisions for students at HBCUs. In fact, although Minor (2008b) is referencing governance at HBCUs when he calls governance at HBCUs an Òenigma,Ó his comment is indicative of the general lack of understanding about the operation of these institutions and scant research within the field. Bastedo (2012) suggests that while a plethora o f higher education research more broadly has focused on organizational issues such as governance and elite leaders, major issues such as understanding who will attend college and why, from an organizational perspective, have been neglected. Additionally, p revious research and the frames that we use to view organizational issues have not been geared toward understanding diverse populations or organizations (e.g., HBCUs) (Bastedo, 2012, Minor, 2005). Moreover, there is other research that compels researchers to understand differences in organizational behavior and structure within the context of their public and private control. These differences are created by varying forces, including resource dependence, institutional mission, culture, program offerings, governance and management, and the composition of the student body, just to name a few (Altbach, 1999; Goldrick -Rab, Kelchen, & Houle, 2014; Harris, 2013; Tolbert, 1985). For the reasons mentioned above regarding institutional and organizational uniqueness, this study attempts to employ critical quantitative research strategies that aim to use data Òto conduct culturally relevant research by studying institutions and people in context Ó (Wells & Stage, 2015, p. 104). Consequently, enhancing our understanding o f the relationship between institutional characteristics and nonblack enrollment at HBCUs is critical for the following reasons: 1. Our current knowledge about HBCUs from an organizational perspective is limited. !!9 2. Our current knowledge about nonblack students at HBCUs is limited. 3. The current challenges HBCUs are facing (particularly enrollment) demand new approaches and new ways of thinking about recruitment and enrollment. 4. Evidence suggests that HBCUs are not successfully promoting the value of attending their institutions to nonblack populations, resulting in the perpetuation of misconceptions and negative stereotypes among nonblack populations. 5. HBCUs have the ability to make significant contributions to the discussion on diversity in higher education, but can only do so if they understand the experiences of all students attending their institutions Ñincluding nonblack students. 6. Researchers and practitioners alike are calling for increased attention to issues of diversity at HBCUs. Over 15 years ago, McDonough, Antonio, and Trent (1997) stated their aspiration to Òoffer insight into the current policy dilemma of attracting other -race students while still ensuring that HBCUs remain attractive to Black studentsÓ in their work on HBCU college choice. Similarly, it is my hope that understanding the enrollment patterns of nonblack students will inform policy makers and institutional actors of current trends that will allow them to implement strategies that maintain the integrity of HBCU culture and success, while adap ting to contemporary challenges and pursuing a sustainable model for the future. Description of Study As stated previously, some leaders within the HBCU community are calling for increased attention to diversity at HBCUs; however, what exactly that call for diversity means when it comes to institutional and scholarly efforts remains somewhat ambiguous. In the context of racial !!10 demographics at HBCUs, only recently have scholars begun to examine the implications of more diverse groups being represented on H BCU campuses ( Ozuna, 2012 ; Palmer et al., 2015; Palmer & Maramba, 2015 ; Strayhorn, 2010) and no known research has approached the topic from an organizational perspective or using quantitative research methods. In order to better understand diversity on H BCU campuses and broaden the scope of scholarly work on HBCUs, this study attempts to provide a nuanced view of nonblack enrollment at HBCUs from an organizational perspective. To support and extend these lines of inquiry on organizational behavior and its relationship to diversity at HBCUs, the proposed study seeks to understand which institutional characteristics of HBCUs are most closely associated with different levels of nonblack student enrollment. PernaÕs (2006) conceptual model of college choice, sp ecifically the higher education context, will serve as a lens through which this issue will be viewed. Although this particular study is discussed in the context of college choice, it should be noted that individual students are not the unit of analysis. The higher education context of PernaÕs (2006) college -choice model provides a platform from which the discussion begins; however, the study is more situated within the college -choice framework as opposed to being a study on individual student college -choi ce decisions. It would be problematic, though, to only view this as a quantitative issue of enrollment patterns and not to acknowledge the ways in which the contextual factors of the model interact. In other words, this study is interested in how instituti ons may be influencing college -choice decisions; but, conversely, it would be irresponsible to dismiss how a studentÕs race, cultural capital, or social capital may also influence the way he/she interprets the characteristics of an institution in which thi s study is interested. !!11 For the purpose of this study, however, the focus is shifted from the individual studentÕs decision -making process to the ways in which the higher education context may influence enrollment decisions for different populations of stu dents. Ultimately, this investigation uses enrollment patterns to better understand an element of college choice. Specifically, this study looks at organizatio nal factors associated with non black enrollment at HBCUs. Organizational factors are part of the institutional context identified by Perna in level three of her proposed conceptual model, which also includes the geographic location and marketing and recruitment efforts of higher education institutions. Perna cites previous research that suggests certa in institutional characteristics do influence student college choice, particularly characteristics that signal to students that they will be attending an institution aligned with their personal and social identities (Nora, 2004). However, Òinstitutional c haracteristicsÓ are not well defined in the proposed conceptual model and, thus, will be further investigated and defined in this study. Therefore, the primary research question that will be explored in this study is: ¥ Are there institutional characterist ics of HBCUs that influence d nonblack enrollment patterns between the years 2000 -2010? o What organizational characteristics, if any, have the strongest relationship to nonblack enrollment at HBCUs (e.g. tuition, institutional spending, graduation rates, re tention rates, etc.)? o Furthermore, what, if anything, do those organizational characteristics suggest about how institutional context shapes college choice? Having established the primary focus of the study, which essentially boils down to determining what would influence a nonblack person to value and participate in a predominantly black space, it is important to acknowledge the inherent interest convergence argument !!12 embedded in the very nature of this study. Interest convergence is a basic tenet of critic al race theory, which is not just a theoretical framework that studies the relationship among race, racism, and power, but is a movement that seeks to challenge the very foundation s of liberalism (Delgado & Stefancic, 2001). Other basic tenets of critical race theory include the ordinariness/permanence of racism, interest convergence (which will be explained further below), intersectionality of oppressed identities/anti -essentialism, combating revisionist history, and highlighting the unique voices of color (Delgado & Stefancic, 2001; McCoy & Rodricks, 2015). In general, applying this critical framework to the analysis of this study of racial diversity at HBCUs discourages readers from applying Eurocentric standards in assigning value to lower or higher leve ls of nonblack student enrollment (Shorette & Arroyo, 2015). More specifically as it relates to this study, interest convergence posits that whites will only advance the interests and issues of people of color if there is an incentive to do so (Delgado & Stefancic, 2001). In other words, whitesÕ self -interests guide their decision -making when it comes to racial justice, not altruism (Delgado & Stefancic, 2001). And in the case of this study, the question of what characteristics of an HBCU are most highly v alued by nonblack students is as much about the general attitudes and beliefs of nonblack students regarding historically black institutions as it is about the perceived incentives (economic and otherwise) for a nonblack student to choose what mainstream ( i.e., white) society has deemed an inferior education of an HBCU over the ÒbetterÓ education offered by predominantly white institutions. Understanding the issue of nonblack enrollment at HBCUs from this perspective adds more depth to the analysis , acknowl edges the historical challenges HBCUs have faced politically, and f rames the issue properly in the context of race and racism in the U.S. Therefore, it is with this perspective !!13 in mind that I encourage the reader to view this study through a critical lens and consider its findings in the context of interest convergence. !!14 Chapter 2 Literature Review Due to the complexities involved with the way historically black colleges and universities (HBCUs) in the U.S. have come to be what they are today, it is essent ial to have a foundational understanding of the history of these institutions, as well other minority -serving institutions, in order to appropriately assess their current state within the higher education community. With that in mind, this section will provide a general overview of minority -serving institutions , establish the historical context of HBCUs, and discuss racial diversity at HBCUs. Additionally, this chapter will synthesize the literature for the conceptual framework of this study: college choice . Minority -Serving Institutions (MSIs) It was not until relatively recently that the higher education community started referring to a collection of specialized institutions as Òminority -serving institutions,Ó or MSIs. Other than HBCUs, which will be dis cussed in depth in the next section, included in this assortment of colleges and universities are Hispanic -serving i nstitutions (HSIs), Asian American and Native American Pacific Isla nder -serving institutions (AANAPISIs), and tribal colleges and universiti es (TCUs). Each designation of institutions has its own unique history and role within the U.S. higher education system. Allen (2008) points out that MSIs are a Òuniquely American creationÓ and suggests that these institutions have made considerable contri butions in expanding access to higher education for underrepresented racial ethnic minorities, especially considering that they sit at the Òintersection where the American Dream of unbridled possibilities meets the American Nightmare of persistent racial -ethnic subordinationÓ (p. xv -xvi). Aside from HBCUs, t ribal colleges and universities have the most formal and oldest connection to the federal government. In response to the civil rights m ovement and the !!15 commitment from the federal government through the passage of the Tribally Controlled Community College Act of 1978, the number of TCUs grew from single digits in the 1970s to 24 by the 1990s (Guillory & Ward, 2008). Currently, 37 tribally controlled colleges and three federally chartered tribal colleges exist, consisting primarily of two -year institutions (Gasman, 2008; Griffin & Hurtado, 2011). The presidents of the first six TCUs came together to create the American Indian Higher Education Consortium (AIHEC) in 1972, which thereafter has served as the primary professional association, fundraising entity, and advocacy organization for TCUs (Guillory & Ward, 2008). Hispanic -serving institutions (HSIs) were next in the chronological order of federal recognition of MSIs. The federal government officially r ecognized HSIs in 1992 and allowed these institutions to apply for federal appropriations geared toward supporting the educational success of their Hispanic students (Griffin & Hurtado, 2011). Unlike TCUs and HBCUs, which maintain contemporary missions tha t still reflect the historical purpose of educat ing Native American and African -American students, Gasman (2008) notes that HSIs are more so Hispanic ÒenrollingÓ institutions than Hispanic ÒservingÓ due to the fact that they are simply defined as Òaccredit ed degree -granting colleges and universities with Hispanic students accounting for 25 percent or more of the undergraduate enrollmentÓ (p. 23). Rather, the majority of HSIs began as predominantly white institutions and circumstantially became HSIs when His panic populations grew substantially in parts of the U.S. Although most HSIs were not created to serve Hispanic students specifically, the approximately 409 institutions that meet the criteria to be an HSI now enroll over half of the entire Hispanic colleg e student population (HACU, 201 5). The most recently designated category of MSIs is the collection of institutions referred to as Asian American and Native American Pacific Islander Serving Institutions (AANAPISIs). !!16 Like HSIs, AANAPISIs are not instituti ons created to serve Asian American and Pacific Islander students, but are institutions that have Òat least a 10% Asian American and Pacific Islander student population as well as a significant percentage of low income studentsÓ (AAPIHERC, 2013). Six insti tutions were designated to serve as pilots for the AANAPISI program through a component of the Higher Education Opportunity Act in 2008 and as of 2011 there were 11 institutions that met the criteria and were participating in the program (AAPIHERC, 2013). Historically Black Colleges and Universities (HBCUs) According to the White House Initiative on Historically Black Colleges and Universities, there are 102 accredited HBCUs, which includes both two -year and four -year institutions (White House, 201 5). In total, 52 are public institutions and 50 are private institutions, all of which are located in 19 states primarily in the South and along the East Coast, as well as in the District of Columbia and U.S. Virgin Islands. Of the 52 public institutions, 41 are f our-year and 11 are two -year; and of the private institutions, 49 are four -year and 1 is two -year. As of 2011, total enrollment for all HBCUs was 346,338 (Gasman, 2013). Many HBCUs, which will also be referred to as b lack colleges throughout this secti on, emerged in the 1800s mostly throug h the philanthropic efforts of w hite missionaries as unofficial (little to no government support) entities to educate freed slaves (Allen & Jewell, 2002; Allen, Jewel, Griffin, & Wolf, 2007). However, the federal gover nment eventually established an official role for HBCUs through the Second Morrill Act of 1890, which financially supported the establishment and operations of these institutions. After passage of the Second Morrill Act, it became almost the sole responsib ility of HBCUs to educate African Americans, as evidenced by the almost exclusive enrollment of African Americans in HBCUs prior to the 1950s (Allen et al., 2007). !!17 Due to the Òseparate but equalÓ standard of the pre -civil r ights era, HBCUs were expected to gear their institutions toward accommodating the needs of African Americans. In order to distance themselves as far from slavery as possible and remove themselves from the subordination to which they had been subjected for too long, HBCUs set out to pro vide opportunities for intellectual advancement within the African -American community (Allen & Jewell, 2002). Although the primary focus was to contribute to the improvemen t of the b lack community through pr eparing the next generation of b lack leaders, HBC Us also acknowledged that they had been charged with the lofty task of making the world in which they operate a better place to live and modeling what is best in America when it came to inclusivity and civility (Allen et al., 2007). For any organization, the aspirational goals of contributing positively to the change of a society rife with government -mandated discriminatory practices and a culture of intolerance would be difficult to achieve; but HBCUs would find that their path to achieving these goals would be riddled with obstacles above and beyond what the typical U.S. higher education institution faced at that time. Primarily, financial limitations have plagued HBCUs from their very inception. Although Brown v Board of Education and other products o f the civil r ights movement were ostensible wins for people of color in the U.S., historical discrimination in the form of significant underfunding by the government, both before and after Brown, put HBCUs at an almost insurmountable disadvantage (Allen et al., 2007). In every way, HBCUs have been educating the most financially and academically needy students . Currently, more than half of all students at HBCUs receive Pell grants and enroll in developmental courses (Gasman & McMickens, 2010; Parker, 2012). Despite the overwhelming proportion of needy students and the end of official !!18 discriminatory practices, recent data still reveal disparities in funding between HBCUs and PWIs . Minor (2008a) and Boland and Gasman (2014) found that although HBCUs in four sta tes enrolled significantly more African Americans than their PWI counterparts, state funding still favored the flagship PWIs by 2 -to-1 in per -student spending and hundreds of millions of dollars in overall funding per institution. Furthermore, Brown v Bo ard of Education had other unintended consequences for HBCUs. As mentioned before, HBCUs were at one time educating the majority of African -American students who enrolled in college. Desegregation, however, removed barriers for African -American students d esirous of attending PWIs for their prestige and plentiful resources. As a result, enrollment of African -American students shifted dramatically from HBCUs to PWIs. Now, HBCUs educate less than a quarter of the African -American students enrolled in postseco ndary institutions (Allen et al., 2007). Contemporary challenges abound for HBCUs, Òwith inadequate resources by far the greatestÓ (Bridges, Cambridge, Kuh, & Leegwater, 2005, p. 29). Essentially, HBCUs were not designed with future success in mind; rath er, they were simply designed to satisfy the demands of African Americans to have access to higher education while ensuring their continued exclusion from PWIs (Abelman & Dalessandro, 2007). These challenges are underscored by historical and contemporary a ccounts of support (or the lack thereof) for public HBCUs, which reveal inequitable funding practices and attempts to weaken the operational effectiveness and competitiveness of HBCUs. The inequitable funding practices include examples such as public HBCUs receiving only 3 percent of the nearly $2.3 million they deserved according to the federal funding formula in 1933 and states underfunding public HBCUs by $57 million between 2010-2012 by not satisfying the one -to-one matching funding requirement ( Lee & K eys, 2013; !!19 Samuels, 2004). On top of the funding issues, HBCUs have also been forced to defend themselves from circumstances that threaten their very survival and funnel students to non -HBCUs. Contemporary examples of these circumstances include the legal fight by public HBCUs in the state of Maryland regarding duplicate academic programs in high -demand fields at public PWIs that fail to dismantle Òseparate but equalÓ segregation -era practices or the efforts of Florida politicians to end a valuable and succ essful engineering cooperative program between Florida A&M University and Florida State University through backdoor political amendments guaranteeing additional funds to Florida State but not Florida A&M ( Hatter, 2014; Wells, 2013). The historical mission of HBCUs has proved to be both challenging and beneficial. The fact that HBCUs are, and have always been, committed to providing access to a quality higher education for African -American students, many of whom are academically underprepared, has created a dditional challenges in a contemporary context. To accommodate the needs of their students, HBCUs have remained true to their mission by keeping tuition at affordable levels and dedicating significant institutional resources to providing developmental educ ation opportunities. However, keeping average in -state tuition lower than the average tuition at all institutions has hindered the ability of HBCUs to generate additional revenue and attract and retain talented faculty (Bridges et al., 2005). Additionally, many states have adopted policies that restrict four -year institutions from offering remedial courses or even allowing students who need developmental education to be admitted to four -year institutions, thereby prohibiting HBCUs from serving the students for which they have geared their institutional practices and policies (Parker, 2012). Abelman and Dalessandro (2007) argue that the desire to stay true to their institutional mission also inadvertently discouraged leaders from thinking strategically abou t their future !!20 place in society. Findings of their study suggest that many HBCUs lack the vision necessary to compete in todayÕs challenging economic and social climate (Abelman & Dalessandro, 2007). For the HBCUs that did have vision statements, Abelman a nd Dalessandro (2007) found that they lacked many of the essential components of a strong vision, such as being easily shared among stakeholders and possessing compelling and innovative ideas. This is particularly problematic considering that HBCUs are hea vily dependent on new -found visions and innovative strategies to tell the stories of their successes ( Gasman, 2012 ; Minor, 2005). Despite the difficult conditions in which HBCUs have been operating, they continue to excel in many areas. A perfect example of their exceptional performance is the fact that although HBCUs represent only 3% of all higher education institutions, they award approximately 16% of all bachelorÕs degrees obtained by African -American students (NCES, 2014). To go along with the dispro portionate percentage of bachelorÕs degrees that they award considering they enroll less than 10% of all African -American students, they also graduate African -American students at a higher rate than their PWI counterparts, regardless of at -risk factors (Be nnet & Xie, 2003, p. 569). Bennett and Xie (2003) also argue that witho ut HBCUs, there would be a net b lack disadvantage, especially for low -income students. In other words, low -income African -American students are afforded and take advantage of opportunit ies to pursue higher education at higher rates than they would otherwise because HBCUs provide opportunities and create environments conducive to those studentsÕ success that are not being provided by PWIs (Bennett & Xie, 2003). Over time, HBCUs have ben efited from positive portrayals of black c olleges through TV shows such as A Different World and The Cosby Show, and movies such as School Daze, Stomp the Yard , and Drumline (Allen et al., 2007). The combination of HBCUs being portrayed as ÒcoolÓ in popula r culture and their reputation among the black community that they are !!21 supportive environments for students of color led to sustained enrollment of African Americans throughout the 1990s and an increase in total enrollment during the 2000s. In fact, HBCUs are now once again competing with PWIs for the most talented students the b lack community has to offer (Allen et al., 2007). Another unique feature of HBCUs is their mission. The missions of these institutions provide the platform from which institutiona l policies and practices are shaped. Of particular interest is the focus of these institutions on serving the African -American population. Because of the historical role of HBCUs, their missions go beyond the typical generic concepts found in many institu tional mission statements. Due to their commitment to educating underrepresented and marginalized students, HBCUs play important social justice roles and it is reflected in the way they operate and promote themselves (Strayhorn & Hirt, 2008). This social justice focus may be observed in statements such as Winston Salem State UniversityÕs motto, ÒEnter to learn, depart to serveÓ (Bridges, Kinzie, Laird, & Kuh, 2008). It is more than a statement of purpose Ðit is an expectation that HBCUs have for their stude nts to serve the broader community. HBCUs have made it a point to not only promote the intellectual development of their students, but to encourage students to take a community -oriented approach to their education and understand how their knowledge is vita l to the advancement of their people (Gasman & McMickens, 2010). Throughout their history, HBCUs have always played major roles in civil rights and social justice movements, as exemplified through North Carolina A&T University and Florida A&M University st udents initiating bus boycotts, Howard University students creating a YouTube video to protest the actions of George Zimmerman and inaction of the Sanford (Florida) police department in the death of Trayvon Martin, and too many others to name. It is clear that HBCUs continue to place major priority on empowering their !!22 students to pursue social justice by providing spaces for students to express themselves and encouraging students to follow their passions and participate in important movements (Allen et al., 2007). Faculty members are at the very core of carrying out the mission of these institutions and creating the welcoming environment that HBCUs are known for. Research has found that many HBCU faculty view their primary responsibilities as ensuring thei r studentsÕ success, which is achieved through Ògetting to know students well, interacting with them inside and outside of the classroom, [and] participating in campus and community eventsÓ (Beach, Dawkins, Rozman, & Grant, 2008). Studies have shown that f aculty at HBCUs have a much stronger connection to the African -American community they are serving and, therefore, place an emphasis on teaching and learning through community service (Beach et al., 2008). Even beyond the undergraduate experience, HBCUs co ntinue their commitment to social justice through their graduate programs and law schools ( Oguntoyinbo, 2012 ). A major component of the HBCU curriculum that sets it apart from the majority of institutions is the special attention that is paid to African -American culture. The missions of the institutions clearly support the advancement of African -American people , and the curriculum is a testament to this commitment. Black culture is embedded in the curriculum regardless of the academic major or specialty ar ea (Bennett & Xie, 2003). This has major implications for historical and cultural understanding across and within racial groups considering that many students will leave their K -12 experience without sufficient knowledge of African -American history and cul ture. Ultimately, the unique mission of HBCUs, the specialized focus of the HBCU curriculum, and the importance faculty place on student success together create a powerful and !!23 nurturing learning environment (Bridges et al., 2008). HBCUs provide a safe hav en from racial discrimination, and students consistently report that their interaction with faculty, their opportunities for student involvement, and a curriculum that affirms their identity as African Americans were all very important in establishing thei r self -worth ( Bennett & Xie, 2003; Bridges et al., 2005; Cuyjet, 2006). Others have found that HBCUsÕ fundamental belief in human potential, ability to remove stigmas, and holistic support of students produce larger gains in intellectual and personal deve lopment ( Bridges et al., 2008 ; Parker, 2012). In spite o f all the positive contributions HBCUs have made and of all the research that has found HBCUs to be supportive and nurturing environments, particularly for black students, it is necessary to ackno wledge what some consider areas for growth and improvement for HBCUs. For example, HBCUs have been known to promote very conservative and exclusionary forms of masculinity (Harper & Gasman, 2008; Patton, 2011). An unfortunate reality that Kimbrough & Harpe r (2006) remind us of is that there are too few mentors within the HBCU community who encourage positive forms of masculinity. Harris III, Palmer, & Struve (2011), therefore, believe that Òthe institutional response should be to challenge and support men i n expressing themselves in more appropriate, positive, and less -destructive ways and to address the larger campus and environmental issuesÓ that encourage men to rely on these more productive strategies as their first response (p. 57). Furthermore, Wa lters and Hayes (1998) describe the intolerance for LGBT people at HBCUs as Òinstitutional homophobia.Ó Researchers suggest that homophobia is embedded in many aspects of HBCU institutional culture. Harp er and Gasman (2008) highlight the ÒConsequences of Conse rvatismÓ and institutional homophob ia at HBCUs when they identify specific written sexual misconduct policies that explicitly state that Òsodomy and homosexual !!24 actsÓ are strictly forbidden (p. 343). Building upon this idea of institutional homophobia, many scholars have highlighted other overt messages sent to the LGBT community at HBCUs, such as the near absence of any Òstudent organizations, offices, centers, or other resources devoted to LGBT concernsÓ (Patton, 2011), the lack of resources on gay and les bian studies in libraries (Willis, 2004), and the public efforts made by administration to prevent LGBT issues from penetrating the walls of their institutions (Harper & Gasman, 2008; Patton, 2011). Racial Diversity at HBCUs Many scholars have concluded t hat diversity in educational settings produces desirable outcomes for all students, such as higher levels of motivation to understand the perspective of others, less likelihood to view diversity as a divisive issue, enjoyment in learning about others, and higher levels of intellectual development and self -assessed academic skills, among others (Gurin, Dey, Hurtado, & Gurin, 2002; Gurin, Nagda, & Lopez, 2004). In rulings during the 2000s, courts also affirmed the importance of diversity in higher education a s a compelling governmental interest (e.g. Grutter v. Bollinger, 2003). Alluded to in the introduction of this paper was the belief by some within the HBCU community that HBCUs have been modeling diversity and inclusion from their inception. It should be noted, however, that not everyone agrees that HBCUs are the place to address diversity in higher education. In fact, much of the discussion around diversity at HBCUs is approached from a deficit perspective. Some have expressed concern about the perceive d divisiveness that HBCUs cause and the perceived low quality of the educational experience provide d there. Vedder (2010) expresses his belief that the idea of race -based institutions that Òcelebrate homogeneityÓ is ÒdisturbingÓ and is an embarrassment to our nation. Often, low graduation rates and low national rankings are used as evidence of HBCUsÕ p oor quality. Vedder !!25 (2010) notes that Òthe 95 or so four -year domestic HBCUs have typical six -year graduate rates around one -third, compared with well over 50 percent for the general population of schoolsÓ and that U.S. News & World Report rankings do not recognize any HBCU as Òa very fine school of the highest distinctionÓ (paragraph 3). Others also agree that the entire purpose of HBCUs, from their understand ing, is to not be diverse and that maintaining support (financial and otherwise) is in direct conflict with the governmentÕs interest in promoting diversity (Seymore, 2006). Connerly (2003), for example, argues that Òit is hypocritical to support the publi c funding of HBCUs and then turn around and criticize a Ôlack of diversityÕ at other public colleges and universities, since HBCUs, by their very nature, draw away many black students who would otherwise attend racially mixed schools and affect their Ôdive rsityÕÓ (paragraph 3). Essentially, opponents of HBCUs believe that HBCUs and broad diversity efforts are mutually exclusive and cannot coexist in any form. Scholars have suggested that the aforementioned criticisms fail to acknowledge institutional conte xt. Low graduation rates at HBCUs, for example, must be placed in the proper context. The historical mission of HBCUs has influenced their admissions standards and the academic profiles of the students who enroll at these institutions. Many HBCUs are essen tially Òopen admissionsÓ institutions with a focus on educati ng African -American students and have much higher percentages of first -generation college students, students from lower -resource high schools, academically underprepared students, federal -aid eli gible students, and Pell grant recipients than P WI institutions ( Allen, Jewell, Griffin, & Wolf, 2007; Harmon, 2012 ; IHEP, 2004). Decades of research has found that HBCUs provide opportunities for black students to enroll in college that would otherwise no t exist if HBCUs were absent, HBCUs graduate black stude nts at higher rates than their P WI counterparts, and HBCUs produce disproportionate !!26 amounts of black graduates (baccalaureate and post -baccalaureate degrees) considering they make up only 3% of the en tire U.S. higher education system (Bennett & Xie, 2003 ; Ehrenberg & Rothstein, 1993 ; Kim & Conrad, 2006). Furthermore, critics often point to HBCUsÕ low rankings as reason why students should not attend these institutions. However, some within the educatio n community believe the U.S. News & World Report rankingsÕ weaknesses overshadow their strengths (IHEP, 2007). Scholars suggest that the search for legitimacy, as defined by many administrators as higher rankings, forces institutions to become more like e ach other (Toma, 2012). Another unintended result of pursuing higher rankings is that it sometimes causes Òcollege or university personnel to work against their own missionsÓ (IHEP, p. 2, 2007). The consequences of the institutional pursuit of prestige are more selective admissions processes and increases in merit aid, which ultimately produce negative effects for low -income and minority students Ñthe very students that HBCUs have proven to be particularly good at educating (Clarke, 2007). HBCUs are inherent ly at a disadvantage when considering that ÒtraditionalÓ measures used to rank institutions (academic reputation as reported by Òpeers,Ó student selectivity, graduation rates, etc.) are not contextualized and do not accurately represent the successes of th ese institutions. Finally, evidence suggests that the beliefs that Òself segregationÓ in the form of HBCUs produces obstacles to diversity efforts in higher education and that HBCUs are in their very nature not diverse are not supported by scholarly rese arch. From a policy perspective, HBCUs have never had discriminatory policies or excluded anyone from attending, which cannot be said for most P WIs. Since the overwhelming majority Ñover 85%Ñof black students attend P WIs, the claim that HBCUs are diverting ÒmanyÓ black students away from P WIs and hindering their diversity efforts is not aligned with the available scholarship (Connerly, 2003; Gasman, 2008b). !!27 Griffin and Hurtado (2011) suggest that HBCUs, their contribution to the heterogeneity of the U.S. hig her education system, and the unique ways in which they carry out their mission of educating our young people should be applauded, not criticized . While the relatively low levels of racial diversity at HBCUs are framed negatively by some as an unwillingnes s of blacks to integrate into PWIs, others view it positively as the willingness and ability of HBCUs to provide nurturing environments for students of color when PWIs fail to do so (Shorette, 2015). To demonstrate this, some would point to the fact that d espite being historically and persistently underfunded and educating the students that most PWIs are neglecting (i.e., large percentages of black and low -income students), HBCUs continue to pull their disproportionate weight and remain the top producers of black graduates in many disciplines (Lee & Keys, 2013; Shorette, 2015). Racial diversity has been evolving on the campuses of HBCUs. In 2008, Gasman (2008b) recognized that white enrollment had declined over the last two decades; however, in a recently released report, Gasman (2013) suggests that HBCUs in the aggregate have experienced a significant demographic shift, with nonblack enrollment now making up a much greater share of the total HBCU population than it had previously at almost 25 percent of to tal enrollment. In contrast to the statistics presented in GasmanÕs (2013) comprehensive report of all HBCUs, when excluding extreme statistical outliers such as West Virginia State University and Bluefield State University, two -year institutions, and grad uate medical schools such as Morehouse School of Medicine and Meharry Medical College, data for four -year HBCUs paint a very different picture, one which still portrays black enrollment as nearly 90% of total enrollment, with nonblack enrollment at around 8 percent and other (Nonresident alien, race/ethnicity unknown, and two or more races) not far behind at 5 percent (Figure 3 ). Interestingly, Gasman (2013) and Ozuna !!28 (2012) note that Hispanic/Latino student enrollment at HBCUs has increased over the past 3 0 years, particularly in states like Texas; however, considering that Hispanic/Latino students now represent over 11% of the total college -going population and with nearly 70% of 2012 Hispanic/Latino high -school graduates attending college, the modest .5% increase in the share of Hispanic/Latino enrollment at four -year HBCUs between 2000 and 2010 seems to demonstrate a disconnect between HBCU enrollment patterns and the national demographic shifts that are occurring in the U.S. (Delta Cost Project, 2009; Roach, 2013 ; Santiago & Reindl, 2009). Figure 3. Percent of Undergraduate Enrollment at Four -year Public and Private N ot-for -profit HBCUs in 2010 College C hoice Broader enrollment trends in higher education are the cumulative product of individual student choices. Therefore, in order to understand higher education enrollment issues, it is critical to understand how students mak e the decision to attend a particular institution. Choosing 87% 8% 5% Black Nonblack (White, Hispanic, Asian, Native Hawaiian/Other Pacific Islander, and American Indian/Alaska Native) Other (Nonresident alien, Race/ Ethnicity Unknown, and Two or more races) !!29 a college is a multifaceted process shaped by a variety of factors. Alt hough scholars have dedicated significant effort to understanding the college -choice process for students over the last 30 years, questions still remain and some areas have been less explored than others. For example, much of the research focused on the in dividual studentÕs ability to navigate the college -choice process in the initial models. As a result, our understanding about how broader contextual factors such as state policy or institutional characteristics influence college choice is less developed. Similarly, fewer developments have been made in our understanding of how college choice differs for students of color and even less is understood about the college -choice process for students attending minority -serving institutions (MSIs). Generally, the c ollege -choice process has been broken down into three stages: predisposition, search, and choice (Hossler & Gallagher, 1987). Predisposition is defined as the stage in which students consider pursuing education beyond high school; search involves identifyi ng institutions possessing the studentÕs desired attributes; and choice is the final stage of choosing the school to attend (Hossler & Gallagher, 1987). Although the model places more of an emphasis on the individual, Hossler and Gallagher (1987) suggest that their model is interactive and accounts for influential aspects of state policy, higher education organizations, etc., at each phase of the college -choice process. However, they are explicit in their belief that those external factors pose a ÒmodestÓ l evel of influence at best and have Òlittle direct impact on student college choiceÓ (Hossler & Gallagher, 1987, p. 209). Bergerson (2009) notes that a major critique of Hossler and GallagherÕs model has been its inability to account for variations in the college -choice process across different populations, especially for traditionally underrepresented students. Findings from a semin al college -choice study affirm critiques to Hossler and Gallagher (McDonough, 1997). McDonough (1997) states , !!30 ÒNot all college -bound students face equal choice if they start out with different family and school resources that enable or constrain their educational and occupational mobility possibilitiesÓ (p. 150). Additionally, McDonough (1997) submits that the college -choice pro cess is complicated and not aligned with models that assume rationality, perfect information, and informed consumers. Furthermore, revelations from large -scale studies involving Chicago Public School students continue to underscore the complexities of the college -choice process and the need for increased attention on contextual factors affecting college -choice outcomes (Roderick et al., 2011). Due to the perceived lack of generalizability of the Hossler and Gallagher (1987) college -choice model to underre presented populations, many scholars set out to address the gaps left by previous studies. Consequently, researchers found that college choice differed significantly for white and black students. Major findings that emerged included the idea that black stu dents perceived the college -choice process differently, their choices were more strongly influenced by socioeconomic factors, and the school and community context mattered greatly (Freeman, 2005; McDonough, 1997; McDonough, Antonio, & Trent, 1997; Mickelso n, 1990; Pitre, 2006). Researchers posit that there is a clear difference between the individual habitus and cultural capital of white students and students of color; that is, students of color generally have less favorable views of the Òfit between a stud entÕs psychosocial needs and the perception that they can be met on a specific campusÓ and possess lower levels of Òsupport and encouragement from family and community upon which a student could draw to influence his or her desire to attend college and to formulate a support systemÓ (Nora, 2004, p. 182). Ultimately, Freeman (2005) suggests that Òstudents are influenced by their perceptions which shape their realities,Ó and that discovery has contributed to the harsh truth that ÒAfrican American students can aspire to !!31 participate in higher education but can believe that actually doing so might not be economically viableÓ (p. 5). Interestingly, and of particular importance to this study, researchers have also discovered differences in the reasons students cho ose to attend HBCUs. Free man (2005) learned that African -American students in her study were more likely to consider HBCUs if they had attended predominantly white high schools or had opportunities to interact with white students more regularly; conversely , Òstudents who had virtually no contact with other races (cultural isolation) Ñthat is, students attending predominantly African American schools Ñexpressed the need to share their cultureÓ and were more likely to pursue higher education at P WIs. Additional ly, McDonough, Antonio, and Trent (1997) found that African -American students choosing HBCUs differed from all students regardless of race, including black students attending PWIs, in their reasons for choosing an HBCU. Particularly signif icant was the fac t that African -American students attending HBCUs believed more strongly than any other group that choosing a college that had a strong academic reputation, produced graduates in high demand by employers, and provided substantial financial aid were the prim ary reasons one should attend college (McDonough, Antonio, & Trent, 1997). In the aggregate, some institutional attributes have been found to factor into college -choice decisions for many students. Nora (2004) notes that, over time, certain institutional characteristics have been found to be more influential than others, including: Ò(a) specific academic programs, (b) affordable tuition costs, (c) financial aid availability, (d) general academic reputation/general quality, (e) location (distance from home) , (f) size, and (g) social atmosphereÓ (p. 181). Conrad et al. (1997) discovered, however, that for white students attending HBCUs, their college -choice criteria were significantly different. For example, four of !!32 the five factors that Conrad et al. (1997) identify in their study as important to wh ite students choosing HBCUs are not consistent with previous research, such as: Òavailability of programs in high -demand fields, availability of academic programs unique to the public HBCU Õs geographic location, a vailability of graduate (masterÕs) programs in high -demand fields, and the offering of programs through alternative delivery systemsÓ (p. 56). The variation in preferences for different populations of students pursuing different types of higher education i nstitutions adds to the complexities of the college -choice process. From an organizational perspective, much of the literature about organizations and their role in influencing enrollment focuses on the effect of tuition and financial aid offers on student sÕ choices. For example, Van Der Klaauw (2002) found that financial aid offers significantly influence the college -going decisions of students, particularly for low -income students. Not only have studies shown that students from lower -income families are m ore susceptible to financial aid offers in the college -choice process ( McPherson and Schapiro, 1991 ; St. John, 1990), but Linsenmeier, Rosen, & Rouse (2006) found that replacing loans with grants at one university also increased the likelihood of low -incom e minority student matriculation by 8 to 10 percentage points. Aside from financial aid, findings from a study of UK students suggest that the quality of facilities on college campuses played a significant role in the college -choice process (Price, Matzdor f, Smith, & Agahi, 2003). Paulsen (1990) suggested that the study of college choice could be viewed in two ways: from macro -level and micro -level perspectives. According to his definition, this macro -level insight of college choice provides an indication of Òhow changes in environmental and institutional characteristics affect an institutionÕs total enrollment,Ó whereas the micro -level speaks to the behaviors of individual students provide an indication of not only how the !!33 environmental and institutional c haracteristics factor into the equation, but also how student characteristics Òaffect a studentÕs choices about whether or not to attend college and which college to attendÓ (Paulsen, 1990, p. 5 -6). One of the more recent and noteworthy developments in the literature that includes PaulsenÕs (1990) micro - and macro -level concepts and addresses the holistic influence that contextual factors exert during the entire college -choice process is Laura PernaÕs (2006) conceptual model of college choice. Having synthe sized the scholarly contributions from the past two decades, Perna (2006) concludes that instead of navigating their way through an uninhibited linear college -choice process, students must interact with a variety of layers that ultimately influence the out come of attending college. The layers of PernaÕs conceptual model, from broad external factors to the more narrow individual factors, include: the social, economic, and policy context; the higher education context; the school and community context; and the habitus at the individual level (see Figure 4). !!34 Figure 4. * Laura PernaÕs Conceptual Model of College Choice !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!* Perna, L. (2006). Studying college access and choice: A proposed conceptual model. In J. C. Smith (Ed.) Higher Education: Handbook of Theory and Research , Volume 21 (pp. 99 -157). Netherlands: Springer. !!"!It is not to say that these findings cannot inform future research a bout private HBCU nonblack !!!!35 Within each layer of the conceptual model are specific factors that comprise the layer and influence the college -choice process for students. T he social, economic, and policy context includes demographic, economic, and public policy characteristics; the higher education context includes marketing and recruitment, location, and institutional characteristics; the school and community context includ es the avai lability and types of resources and structural supports and barriers; and the individual studentÕs habitus layer includes demographic characteristics (race/ethnicity and gender), cultural capital (cultural knowledge and value of college attainme nt), and social capital (information about college and assistance with college processes) (Perna, 2006). All of the aforementioned layers ultimately influence the most central factors of a studentÕs college -choice process, which are the demand for higher e ducation (academic preparation and academic achievement), supply of resources (family income and financial aid), expected benefits (monetary and non -monetary), and expected costs (college costs and foregone earnings) (Perna, 2006). This integrated model ad dresses the reality that Òwhen considered separately, neither rational human capital investment models nor sociological approaches are sufficient for understanding differences across groups in student college choiceÓ (Perna, 2006, p. 114). !!!!!!!36 Chapter 3 Research Design and Methodology The following chapter will introduce the conceptual framework, the data source, and quantitative methods that will be used to answer the research questions of interest: Are there institutional characteristics of HBCUs that influence d nonblack enrollment patterns between the years 2000 -2010? Furthermore, what organizational characteristics, if any, have the strongest relationship to nonblack enrollment at HBCUs during that time period (e.g. tuition, institutional spending, gr aduation rates, retention rates, etc.)? Conceptual Framework College choice. It is important to note again that this study, although discussed in the context of college choice, is not strictly an individual college -choice study. Drawing from Paulsen (1990 ) and Perna (2006), this study is viewing the enrollment behavior of students in groups as a macro -level consideration in the overall college -choice process. The higher education context of PernaÕs (2006) college -choice model will serve as the foundation f or the discussion. Although the nature of the data and methodological approach of this study will not afford the researcher insight into individual studentsÕ college -choice decisions, the investigation will use enrollment patterns to better understand an element of college choice Ñthe institutional element Ñthat interacts with student -level factors. For this study, the focus will be placed on the higher education context at public and private, not -for -profit, four -year HBCUs. As stated previously, included i n this context, according to Perna (2006), are institutional characteristics, location, and marketing and recruitment. Perna (2006) suggests that higher education institutions play a significant role in influencing the college -choice process by actively an d passively conveying certain messages !!37 about their institution. From a studentÕs point of view, their choice is largely determined by their belief that the higher education institution will be comfortable, accepting, and a good fit (Nora, 2004). Findings from Hazzard (1996) and Conrad et al. (1997) suggest that for white students at HBCUs, the influential factors for choosing to attend an HBCU could be placed into four general categories: academic program offerings, student financial support, location, an d institutional characteristics. Most relevant to this study, Conrad et al. (1997) advanced some policies and practices they felt could enhance institutional efforts to attract white students to HBCUs. First, they posit that states should provide adequate funding to allow HBCUs to ensure there are meaningful numbers of high -demand academic programs and ample financial support for students. Second, Conrad et al. (1997) urge HBCUs to focus their attention on three major strategies at the institutional level: Òprogram distinctiveness and enhancement, student financial support and recruitment, and institutional enhancementÓ (p. 57). In short, these strategies essentially translate into developing and maintaining unique programs in high -demand fields Ñparticularl y at the graduate level; strengthening scholarships awarded to white students; and increasing institutional quality and reputation through faculty, facilities, and diversity initiatives (Conrad et al., 1997). Education production function. Many of the org anizational characteristics being examined in this study that fall within the higher education context of PernaÕs (2006) conceptual model manifest themselves through the allocation of resources, such as spending on instructional activities or spending on a dministrative costs. Investigations into the effect of school resources on educational outcomes have a long history in the K -12 scholarship (Hanushek, 1997); however, little research has been conducted on the relationship between institutional !!38 expenditures and student outcomes in postsecondary educational contexts (Ryan, 2005). Recent contributions to the postsecondary education production function research have begun to shed some light on the issue. Although the Delta Cost Project concluded that, in the ag gregate, there is no clear correlation between general funding levels and degree production, other scholars have found strong relationships between institutional expenditures and student outcomes (Desrochers, Lenihan, & Wellman, 2010). For example, regardi ng degree completion, one study found that student services expenditures influence persistence and graduation rates (Webber & Ehrenberg, 2010), while another study suggested that instructional expenditures and various institutional characteristics proved t o be positively associated with degree completion (Hamrick, Schuh, & Shelley, 2004). Still another found that tuition revenue and educational and general expenditures were positively associated with degree completion (Titus, 2006). In studies focusing on s tudent outcomes other than degree completion, administrative spending was found to have a negative effect on student engagement (Ryan, 2005) and Toutkoushian and Smart (2001) suggest that targeted institutional expenditures do affect student gains in areas such as interpersonal skills, learning/knowledge, tolerance/awareness, graduate/professional school preparation, and communication skills. Some of the studies previously mentioned state explicitly that HBCUs are statistically significant outliers in comp rehensive models of all postsecondary institutions (Hamrick, Schuh, & Shelley, 2004; Ryan, 2004). As a result, the authors emphasized repeatedly that these institutions introduced much more complexity into the study than was expected and suggested that tho se findings clearly warranted a Òneed to conduct more in -depth studies of these [HBCUs] and other institutions with distinctive histories, missions, and identitiesÓ (Hamrick, Schuh, & Shelley, 2004; Ryan, 2004, p. 109). In making that suggestion, Ryan (200 4), for example, !!39 recognized that a culturally sensitive approach might be more appropriate when examining the effects of institutional expenditures on degree attainment at HBCUs. RyanÕs (2004) observation is directly aligned with MinorÕs (2008) and critica l quantitative scholarsÕ (Wells & Stage, 2015) belief that researchers must conduct studies and view issues at HBCUs in their proper cultural and historical context. Conclusion. Unfortunately, no known research exists that has extended RyanÕs (2004) line of inquiry and examined these particular issues at HBCUs, and previous college -choice studies involving institutional characteristics have not focused on institutional expenditures. Since there is evidence that institutional expenditures may be related to student outcomes and that institutional characteristics have been found to be inf luential factors in the college -choice process, this study attempts to fill the current void in our understanding of how institutional expenditures and other institutional cha racteristics may influence nonblack enrollment at HBCUs. More specifically, because previous studies of nonblack students attending HBCUs reveal differences in the reasons they chose to attend an HBCU (Conrad et al., 1997; Daniels, 2008; Hazzard, 1996 ), th is study is interested in understanding how institutional characteristics in the form of institutional expenditures and various indicators of institutional performance and reputation (graduation rates, retention rates, selectivity, and 75 th percentile ACT scores) may influence the college -choice decisions of nonblack students who may be perceiving investments in different areas of institutional operations as contributing to their success differently than black students. Below is a visual representation of how various concepts of college choice, HBCUs, and education production function will be incorporated to form the conceptual model of this study. !!40 Figure 5. Conceptual model of nonblack HBCU student choice ! Data S ource To answer the research questions described previously, the researcher will use data from the Delta Cost Project at American Institutes for Research (AIR) database, which is drawn directly from the Integrated Postsecondary Education Data System (IPEDS) from the years 1987 -2010. IPEDS is a comprehensive database of postsecondary education data maintained by the U.S. Department of EducationÕs National Center for Education Statistics (NCES) and collected through a series of surveys. As stipulated in Title IV of the Higher Education Act, all i nstitutions receiving federal financial aid are required to complete the surveys. The data collected through the surveys include information such as basic institutional characteristics (e.g., public or private control ) and level of degrees offered; enrollm ent; tuition and fees; student financial aid (federal, Institutional characteristics of HBCUs (layer 3 - higher education context) Spending on institutional grant aid, instruction, academic support, student services, administration, and auxiliary services Graduation rates and selectivity Nonblack HBCU student (layer 1 - habitus) Race HBCU choice !!41 state, and institutional); student persistence and success (retention and graduation rates); and institutional resources (human resources and institutional finances). The Delta Cost Project organizes IPEDS data collected by NCES and presents aggregate measures of institutional education and related costs, which are defined by the Delta Cost Project as Òall spending for instruction and student services, plus a portion of spending on academic and institu tional support and for operations and maintenance of buildingsÓ (Delta Cost Project, 2012). All the financial data is adjusted for inflation according to 2010 dollars. The IPEDS data made available through the Delta Cost Project is particularly useful fo r the purposes of this study for three primary reasons: First, the dataset provides measures that serve as proxies for Òinstitutional characteristicsÓ (i.e. , the services on which an institution chooses to direct resources, graduation rates, etc.) which ar e of particular interest for the purpose of this study, as opposed to focusing on individual student perceptions or outcomes; second, the available data allow for analysis over time rather than limiting the researcher to a cross -sectional snapshot of one y ear; and third, the dataset includes the entire population of HBCUs (rather than a sample of a population), which strengthens the inferences being made. Measures PernaÕs (2006) conceptual model of college choice, specifically the higher education context, the education production function framework, and previous research on institutional characteristics influencing college choice (Conrad et al., 1997; Nora, 2004) guided the selection of independent variables used in this study. Data collected through the IPEDS surveys include measures that serve as proxies for general academic reputation and quality, such as retention rates, graduation rates, 75 th percentile composite ACT scores, and acceptance rates. Unfortunately, many of these variables were poorly !!42 repo rted, particularly by private HBCUs and at all institutions before the 2003 -2004 academic year, and therefore limit the analysis. Of the aforementioned measures, graduation rates was the most consistently reported measure, with less than two percent missin g after the 2002 -2003 academic year. The 75th percentile composite ACT scores were approximately 40 percent missing among all HBCUs after the 2002 -2003 academic year. Because there was a theoretical interest in including a measure of selectivity in the stu dy, an attempt was made to identify a selectivity measure using sources outside of the IPEDS database, including acceptance rates found in the U.S. News & World Report rankings and acceptance rates and Admissions Competitiveness Index found in BarronÕs Pro files of American Colleges . Unfortunately, the U.S. News and World Report rankings did not provide the additional data that was needed to fill in the gaps in the IPEDS data and the BarronÕs Profiles of American Colleges Admissions Competitiveness Index, al though providing an index for nearly every HBCU in the study, was a static variable over time and was therefore omitted in the regression analysis due to colinearity. Before excluding the selectivity variable entirely due to missing data , a new binary var iable was created to capture whether an institution reported 75 th percentile composite ACT scores (coded as Ò0Õ) or whether the data was missing (coded as Ò1Ó )Ñthe assumption being that institutions may be more likely to report this data if they are a more selective/competitive institution and less likely to report this data if they are less selective/non -competitive. However, a correlation analysis was performed to test this hypothesis by correlating the BarronÕs Admissions Competitiveness Index (ranging f rom 0 for noncompetitive to 4 for very competitive) and missing ACT scores. The results to this correlation analysis suggested that there was no significant correlation between the level selectivity and whether or not an institution reported ACT scores. Additionally, in response to the aforementioned issues in the data, separate !!43 models were constructed to account for the inconsistently reported or missing data and to establish the strength of the models with and without the fully reported variables: one mod el including all years between 2000 -2010 and one model using only the years after 2003. Other variables, such as spending on instruction, student services, academic support, administration, institutional financial aid, and auxiliary services, serve as indi cators of how the institution prioritizes the allocation of resources in its effort to promote student success and strengthen the institutionÕs performance, reputation, and appearance. To account and control for the fact that varying levels of total enroll ment directly affect total operating budgets, spending variables will be expressed as spending per full -time equivalent (FTE) student rather than total amount spent per category. Ryan (2004) notes that viewing the institutional expenditures as per FTE, as opposed to as a percentage of total expenditures, permits the testing of the unique effect of actual expenditures in the categories included in the study by controlling for the expenditure categories excluded from the study. Similarly, to control for vary ing levels of the raw number of nonblack students enrolled at each institution, nonblack enrollment will be expressed as a percentage or share of total enrollment. !!44 Table 1. Description of Predictor Variables Included in Study Variable Theoretical Proxy Code Graduation Rate Institutional quality/reputation 0-100% Tuition and fees (sticker price) Institutional commitment to affordability Thousands of dollars Institutional Financial Aid Institutional commitment to affordability Thousands of dollar s per full -time equivalent student Spending on Instruction Investment in quality classroom experiences Thousands of dollars per full -time equivalent student Spending on Academic Support Investment in quality academic personnel and services Thousands of d ollars per full -time equivalent student Spending on Student Services Investment in quality extracurricular activities and support Thousands of dollars per full -time equivalent student Spending on Administration Investment in quality institutional managem ent Thousands of dollars per full -time equivalent student Spending on Auxiliary Services Investment in self -supporting enterprises such as residence halls, intercollegiate athletics, and food services Thousands of dollars per full -time equivalent student Sample This study used the following selection criteria to determine an HBCUÕs eligibility for inclusion in the sample: 1. Must be an HBCU between the years of 2000 -2010 (IPEDS variable code - hbcu) 2. Must be a Òfour -year or higherÓ institution (IPEDS variabl e code - iclevel) 3. Must be a public or private not -for -profit institution (IPEDS variable code - control) 4. Must not be a primarily professional/graduate school (such as medical schools) !!45 According to the IPEDS database and the criteria for the sample used in this study, 85 institutions met the criteria of being either a public or private, not -for -profit , four -year HBCU between the years of 2000 -2010. As stated previously, the criteria for determining HBCU status has been established by the U.S. Department of Education in the H igher Education Act of 1965 as a school of higher learning whose principal mission was and is the education of African Americans and was accredi ted and established before 1964 (UNCF, 2013). Of those 85 HBCUs, 45 (53%) of them are private, 40 (47%) are public, and all of them are concentrated in 19 states, as well as the District of Columbia and the U.S. Virgin Islands. Table 2 presents a general snapshot for each of the 85 HBCUs in this study, including the name, location, institutional control, total enrollment as of 2010, and the varying degrees of nonblack enrollment between 2000 and 2010. Table 2. Description of Institutions Included in Study Name of Institution State Control Total Enrollment 2009-2010* Share of nonblack enrollment 2000-2001 Share of nonblack enrollment 2009-2010 Change in share of nonblack enrollment 2000-2010 Alabama A & M University AL Public 5,327 12.86% 5.40% -7.46% Alabama State University AL Public 5,564 9.39% 3.29% -6.10% Albany State University GA Public 4,473 7.87% 6.79% -1.08% Alcorn State University MS Public 3,334 5.38% 6.57% +1.19% Allen University SC Private 827 0.30% 0.60% +0.30% Arkansas Baptist College AR Private 640 1.27% 6.56% +5.29% Benedict College SC Private 2,983 0.11% 0.70% +0.59% Bennet t College for Women NC Private 766 0.0% 0.91% +0.91% Bethune -Cookman University FL Private 3,637 2.27% 3.51% +1.24% Bluefield State College WV Public 1,989 91.06% 83.76% -7.30% Bowie State University MD Public 5,617 20.00% 7.99% -12.01% Central State U niversity OH Public 2,436 5.31% 2.67% -2.64% Cheyney University of Pennsylvania PA Public 1,488 9.28% 3.09% -6.19% Claflin University SC Private 1,860 0.45% 2.04% +1.59% Clark Atlanta University GA Private 3,873 1.37% 0.82% -0.55% Concordia College -Sel ma AL Private 568 8.76% 2.64% -6.12% !!46 Table 2 (contÕd) Coppin State University MD Public 3,801 5.25% 2.05% -3.20% Delaware State University DE Public 3,609 21.37% 14.41% -6.96% Dillard University LA Private 1011 0% 1.68% +1.68% Edward Waters Coll ege FL Private 831 2.21% 3.85% +1.64% Elizabeth City State University NC Public 3,264 24.31% 16.91% -7.40% Fayetteville State University NC Public 6,283 28.12% 22.47% -5.65% Fisk University TN Private 650 0.22% 1.54% +1.32% Florida Agricultural and Mec hanical University FL Public 12,274 7.34% 8.37% +1.03% Florida Memorial University FL Private 1,923 4.77% 3.95% -0.82% Fort Valley State University GA Public 3,553 5.94% 3.85% -2.09% Grambling State University LA Public 4,992 3.9% 3.68% -0.22% Hampton University VA Private 5,402 12.09% 8.33% -3.76% Harris -Stowe State University MO Public 1,886 21.52% 8.16% -13.36% Howard University DC Private 10,573 4.34% 3.85% -0.49% Huston -Tillotson University TX Private 882 17.73% 19.73% +2.0% Interdenominational Theological Center GA Private 421 2.98% 1.66% -1.32% Jackson State University MS Public 8,783 3.37% 6.85% +3.48% Jarvis Christian College TX Private 628 1.35% 6.05% +4.70% Johnson C Smith University NC Private 1,466 0.84% 0.75% -0.09% Kentucky State U niversity KY Public 2,834 38.49% 26.60% -11.89% Lane College TN Private 2,146 0.0% 0.19% +0.19% Langston University OK Public 2,749 41.02% 14.26% -26.76% Le Moyne -Owen College TN Private 890 0.0% 0.45% +0.45% Lincoln University MO Public 3,314 64.71% 57.33% -7.38% Lincoln University of Pennsylvania PA Public 2,649 4.93% 2.87% -2.06% Livingstone College NC Private 1,082 1.83% 1.29% -0.54% Miles College AL Private 1,791 .20% 1.73% +1.53% Mississippi Valley State University MS Public 2,850 4.9% 4.77% -0.13% Morehouse College GA Private 2,689 0% 0.85% +0.85% Morgan State University MD Public 7,226 2.8% 4.88% +2.08% Morris College SC Private 966 0.22% 0.31% +0.09% Norfolk State University VA Public 6,993 12.02% 7.22% -4.80% North Carolina A & T State University NC Public 10,614 10.17% 10.08% -0.09% North Carolina Central University NC Public 8,587 16.25% 14.05% -2.20% Oakwood University AL Private 1,916 2.65% 1.93% -0.72% Paine College GA Private 907 0% 1.87% +1.87% Paul Quinn College TX Private 171 8.36% 1.75% -6.61% Philander Smith College AR Private 668 0% 1.04% +1.04% Prairie View A & M University TX Public 8,608 9.84% 10.79% +0.95% !!47 Table 2 (contÕd) Rust College MS Private 1,072 .35% 1.40% +1.05% Saint AugustineÕs College NC Private 1,529 1.21% 2.68% +1.47% Saint PaulÕs College VA Private 584 2.80% 2.74% -0.06% Savannah State University GA Public 3,820 8.64% 4.63% -4.01% Shaw University NC Private 2,538 3.26% 2.52% -0.74% South Carolina State University SC Public 4,538 5.47% 4.01% -1.46% Southern University and A & M College LA Public 8,218 4.27% 8.07% +3.80% Southern University at New Orleans LA Public 3,141 4.64% 3.53% -1.11% Southwestern Christian College TX Private 201 0.0% 2.98% +2.98% Spelman College GA Private 2,229 0% 0.31% +0.31% Stillman College AL Private 1,041 .48% 2.98% +2.50% Talladega College AL Private 700 2.42% 5.71% +3.29% Tennessee State University TN Public 8,824 22.36% 23.42% +1.06% Texas College TX Private 964 2.38% 15.04% +12.66% Texas Southern Univer sity TX Public 9,394 11.36% 11.80% +0.44% Tougaloo College MS Private 939 0.00% 1.60% +1.60% Tuskegee University AL Private 2,931 1.33% 2.86% +1.53% University of Arkansas at Pine Bluff AR Public 3,792 6.08% 4.03% -2.05% University of Maryland Eastern Shore MD Public 4,433 20.93% 16.35% -4.58% University of the District of Columbia DC Public 5,253 13.30% 16.26% +2.96% University of the Virgin Islands VI Public 2,602 7.51% 11.45% +3.94% Virginia State University VA Public 5,366 8.13% 4.40% -3.73% Vir ginia Union University VA Private 1,691 1.66% 1.95% +0.29% Virginia University of Lynchburg VA Private 327 0.00% 1.22% +1.22% Voorhees College SC Private 701 4.62% 1.28% -3.34% West Virginia State University WV Public 6,229 86.59% 69.72% -16.87% Wilber force University OH Private 710 5.70% 2.96% -2.74% Wiley College TX Private 1,237 3.11% 5.50% +2.39% Winston -Salem State University NC Public 6,427 18.29% 16.99% -1.30% Xavier University of Louisiana LA Private 3,338 7.7% 17.14% +9.44% * Total enrollme nt is defined by NCES, IPEDS for the academic year 2009 -2010 as the number of students enrolled in the fall at postsecondary institutions. Students reported are those enrolled in courses creditable toward a degree or other formal award; students enrolled i n courses that are part of a vocational or occupational program, including those enrolled in off -campus centers; and high school students taking regular college courses for credit. Institutions report annually the number of full - and part -time students, by gender, race/ethnicity, and level (undergraduate, graduate, first -professional); the total number of undergraduate entering students (including first -time, transfers -in, part -time students, and non -degree students); and retention rates. It is important t o note that for the purposes of this study, some institutions were excluded from the entire sample of HBCUs. Morehouse School of Medicine and Meharry Medical College !!48 were excluded due to their status as graduate/professional schools. American Baptist Colle ge was also excluded from this study since it was not granted the HBCU designation until 2013, therefore not meeting the criteria of being an HBCU between the years of 2000 -2010. Research Method This research uses IPEDS data to conduct a nonexperimental, observational study using the Ordinary Least Squares (OLS) regression technique. This classic statistical approach is intended to establish a relationship between multiple predictor/independent variables (X) and an outcome/dependent variable (Y) of interes t while minimizing the difference between the actual Y scores and the predicted Y scores using the least squares criterion (Hinkle, Wiersma, & Jurs, 2003). The hope is that through the OLS method, the researcher can determine a Òbest fitÓ model for predict ing the outcome variable of interest .!Basic cross -sectional data analysis allows a researcher to observe differences across units of analysis at one particular point in time . One way to address that inherent limitation of cross -sectional data is to use pan el data, which provides a more comprehensive history and can track changes within an institution. Panel data allows researchers to analyze multiple waves of observed data over a period of time within multiple units, which can be organized such that each unit has multiple records to represent the multiple waves of data collected. In this study, the data are organized so each unit (institution) has records, where t is the number of waves for individual (institution) , so the total number of records in the analysis is !!(Johnson, 1995). The basic equation for the panel data regression is as follows: tiiti!!!49 !!!!!!!!!!!!!!!!!!!!!!!"!!!#!$%&'&( !!!!)!!outcome/dependent variable of interest !"!)!!slope or regression coefficient for the respective predictor/independent variables !#$)!predictor/independent variables !!!)!!resid ual errors of the regression !!This study is less concerned about determining effects over time between institutions; instead, the focus is on aggregate effects within institutions during the observed time period. Focusing on aggregate effects will prove be neficial when determining the practical and theoretical significance of the results since the within unit findings will have broader applications. Therefore, the simplest way to focus on effects within institutions over a specified period of time is to con duct a fixed effects OLS panel data regression analysis, which will remove the institutional effect and Òignore the time dimension of the data structure and treat each observation as independently drawnÓ (Zhang, 2010 , p. 316). Although the analysis will be focused on a specific period of time in which the units were observed (between the years 2000 -2010), no inference will be made regarding the effect of time on the outcomes of interest since it will not be included as a variable in the equation. Ultimately , a fixed effects approach removes the effects of time -invariant characteristics of institutions and shifts the focus to assessing the net effect of the predictor variables. Therefore, the following OLS model will be used to investigate the effects of inst itutional characteristics on nonblack enrollment at HBCUs between 2000 -2010: yit (pct_nonblack) = !0 + !1 (pct_grad_rate )it + !2 (tuitionandfee02 )it* + !3 (tuitionandfee03 )it + !4 (instaid_spend_1000 )it + !5 (instruct_spend_1000 )it + !6 (acadsupp_spend_1000 )it + !7 (studserv_spend_1000 )it + !8 (admin_spend_1000 )it + !9 (aux_spend_1000 )it + !! + "it !!50 Where : ! y represents the percentage of nonblack students ! !o represents the overall intercept ! !1 (pct_grad_rate) it represents the effect of graduation rates for case i in the sample at t time period ! !2 (tuitionandfee02 )it represents the effect of in-state tuition* for case i in the sample at t time period ! !3 (tuitionandfee03 )it represents the effect of out-of-state/private tuition for case i in the sample at t time period ! !4 (instaid_spend_1000 )it represents the effect of institutional grant aid per F TE student for case i in the sample at t time period ! !5 (instruct_spend_1000 )it represents the effect of spending on instruction per FTE student for case i in the sample at t time period ! !6 (acadsupp_spend_1000 )it represents the effect of spending on acade mic support per FTE student for case i in the sample at t time period ! !7 (studserv_spend_1000 )it represents the effect of spending on student services per FTE student for case i in the sample at t time period ! !8 (admin_spend_1000 )it represents the effect o f spending on administration per FTE student for case i in the sample at t time period ! !9 (aux_spend_1000 )it represents the effect of spending on auxiliary services per FTE student for case i in the sample at t time period ! !! represents all time -invar iant variables that could influence nonblack enrollment but are not included in this study ! "it represents residual error for case i in the sample at t time period * The in -state tuition variable was used exclusively in the public HBCU model. Note: All spe nding and tuition variables were transformed by dividing per FTE figures by 1000 for easier interpretation of coefficients. It is possible that the college -choice decisions of students may be influenced by the result of an institutionÕs decisions/behavior /performance in the years leading up to the point at which a student chooses a college to attend. Applying lags to dependent variables is not uncommon in higher education research and can be observed in recent studies ranging from the study of growth in in ternational doctoral education (Taylor & Cantwell, 2014) to the effect of state financial policies on the production of bachelorÕs degrees (Titus, 2009). Therefore, lags will be applied to !!51 the dependent variables in this study and the same fixed effects pa nel data regression approach will be used to conduct two additional analyses for public and private HBCUs between 2003 -2010: one analysis with one -year lags for the predictor variables ( t-1) and one analysis with two -year lags for the predictor variables ( t-2). yit (pct_nonblack) = !0 + !1 (pct_grad_rate )i(t -1) or (t -2) + !2 (tuitionandfee02 )i(t -1) or (t -2)* + !3 (tuitionandfee03 )i(t -1) or (t -2) + !4 (instaid_spend_1000 )i(t -1) or (t -2) + !5 (instruct_spend_1000 )i(t -1) or (t -2) + !6 (acadsupp_spend_1000 )i(t-1) or (t -2) + !7 (studserv_spend_1000 )i(t -1) or (t -2) + !8 (admin_spend_1000 )i(t -1) or (t -2) + !9 (aux_spend_1000 )i(t -1) or (t -2) + !! + "it Finally, in order to gain a deeper understanding of HBCUs in various institutional contexts, two differe nt regression models will be constructed: 1) only public HBCUs and 2 ) only private HBCUs. Distinguishing between public and private HBCUs is a necessary approach due to the significant differences between these institutional types in almost all regards, in cluding total enrollment, institutional missions, institutional histories, program offerings, and more. Considering HBCUs within their specific institutional control will allow for more precise interpretations of the statistical analysis. Limitations Firs t, it is important to acknowledge the limitations of the dataset. Because this is a secondary data analysis, not every theoretical concept was measured directly and formatted according to the objectives of this study. Therefore, certain variables will be s erving as proxies for various theoretical constructs. Additionally, the statistical approach being used presents a challenge. Specifically, because the fixed effects OLS panel regression ignores the dimension of time and pools all observations for all unit s of analysis (institutions) into one model, it produces a greater likelihood of model errors for a given unit being correlated. Researchers may !!52 misinterpret statistically significant results without accounting for within -cluster error correlation. In orde r to address this potential problem, the researcher will use a cluster -robust standard errors approach, which reduces the usually downward bias in standard errors that occurs. By reducing the aforementioned biases, the researcher can control for unobserved heterogeneity (i.e. , omitted information from unobserved variables) and make more conservative assumptions about the findings. Furthermore, conducting a quantitative study restricts the richness of findings due to its focus on aggregate statistical effec ts. Particularly for a concept like college choice, numerical representations do not allow for the personal narratives of nonblack students who could directly describe their perceptions of HBCUs in their college -choice decisions. There is also a possibilit y that findings from a quantitative study may miss relevant and influential factors related to the phenomenon due to a focus on a particular theory or due to the restrictions of the dataset. For example, it is possible that marketing and recruitment could play a role in varying levels of nonblack enrollment at HBCUs; however, the IPEDS dataset does not include information regarding the amount of money spent specifically on marketing and recruitment or how much of that marketing and recruitment is geared tow ard nonblack students. Finally, because this is an observational study rather than what many consider the gold standard in research Ña randomized experiment Ñsome may express concerns regarding the ability of the researcher to infer causality. Scholars have debated this issue for decades and although there is no absolute solution to the problem of causal inference in observational studies, certain statistical methods and theories have been developed to address many of the limitations of observational studies . Angrist and Pischke (2009) suggest that it is Òuseful to think of causal relationships in terms of the potential outcomesÓ that describe what would happen given an !!53 alternative scenario (p. 52). Typically, proving causality is done through conducting rand omized experiments. However, in the social sciences in particular, it is difficult to conduct true randomized experiments. In the case of observational, non -experimental studies, not all conditions can be met in order to perfectly infer causality. For one, the counterfactual cannot be provided; that is, data does not exist that provides the outcome for the participant under both the treatment and control conditions (Morgan & Winship, 2007; Murnane & Willett, 2011). Additionally, most datasets have not obser ved everything , so there is a real possibility that there are unobserved dimensions (confounding influences) of an observational study. In order to address the limitations of observational studies, the researcher has paid careful attention to the method b eing used and the selection of the variables included in the statistical models. Regression techniques have proven to be useful tools in minimizing bias. In fact, Winship and Morgan (2007) argue that regression models can be used as descriptive tools that serve as Òa method for obtaining a best -fitting descriptive model under entailed linearity constraintsÓ (p. 123). Regression techniques are certainly not the solution to solving all causality issues, but by taking steps to reduce observed bias and address confoundedness, one can draw reasonable inferences from an observational study (Murnane & Willett, 2011). It is possible that critical readers may suggest that the researcher did not control for some unobserved variable, such as individual studentsÕ proxim ity to the institutions included in the study or the socioeconomic status of students. Frank (2000; 2014) and others (Frank et al., 2013; Pan & Frank, 2003) offer a strategy for researchers to determine the strength of their findings from an observational study when all influences cannot be controlled for: the percent of an estimate that must be due to bias to invalidate an inference . The calculation allows the researcher to address the concern of not having controlled for every potential influence by dete rmining how !!54 much bias would have to be due to omitted variables such that it would make the predictor of interest insignificant (Frank, 2000; Frank, 2014; Frank et al., 2013; Pan & Frank, 2003). This technique will be used to test the robustness of the fin dings from this study. Conclusion Ultimately, by viewing the issue of nonblack enrollment at HBCUs through the lens of the higher education context of PernaÕs (2006) conceptual model of college choice and employing a panel data OLS regression analysis, thi s study attempts to fill the current gap in understanding about how institutional characteristics Ñas manifested through signals of institutional priorities that fall within the institutionÕs realm of control , such as graduation rates, spending on various a spects of institutional operations, tuition, and student financial aid Ñmay have influenced nonblack enrollment at HBCUs between the years 2000 -2010. Stata © Version 13.1 statistical software was used to conduct the analysis. !!55 Chapter 4 Findings The preceding chapters established the context and methodological approach for this study, which examines the effect of institutional characteristics on nonblack enrollment at HBCUs between the years 2000 -2010. This chapter will present descriptive statistics for the variables in the study, report the results of the fixed -effects panel OLS regression analyses conducted for the two separate models, and provide an interpretation of the results. Descriptive Statistics Table 3 provides a statistical de scription of all variables over the ten -year period included in the quantitative analysis in this study. The statistics are presented according to institutional control (i.e., public versus private) . Although there are many similarities between the institu tional categories, there are some important distinctions to be made. To start, some of the non -spending characteristics differed significantly between public and private HBCUs. The average percentage of nonblack enrollment, the dependent variable in this study, varied greatly between the different institutional groupings, with public HBCUs averaging 15.52 percent nonblack enrollment and 19.20 percent variation within the sample, compared to 2.91 percent average nonblack enrollment and 3.77 percent variatio n within the sample of private HBCUs. Although the mean six -year graduation rates were similar between the samples (public HBCUs = 33.11 percent; private HBCUs = 37.73 percent), the level of variation within each sample differed significantly, with a high level of variation at private HBCUs (20.91 percent) and a relatively moderate level of variation at public HBCUs (10.41 percent). The non -spending characteristics highlight and provide a preview of some of the key differences between public and private HBC Us. !!56 When it comes to spending, similar and sometimes even more significant differences can be observed. For example, the average amount of institutional grant aid per FTE student at public HBCUs was less than half that of privates (public HBCUs = $978.88; private HBCUs = $2,144.84), and the variation was far lower at public HBCUs (public HBCUs = $703.85; private HBCUs = $1,555.79). Average in -state tuition at public HBCUs was $3,850.70 and varied $1,388.66 per standard deviation. No distinction was made in the reporting of in -state and out -of-state tuition by private HBCUs, so only out -of-state tuition was included in the model, in which the average was $9,753.41 with a variance of $3,382.04. Of all the spending categories, only two were higher at public HBC Us: spending on instruction and spending on academic support. Public HBCUs spent $5,950.60 per FTE student on instruction compared to $5,363.94 at private HBCUs, and $1,721.88 on academic support spending per FTE student at public HBCUs compared to $1,546. 53 at private HBCUs. Spending on student services per FTE student was lower at public HBCUs compared to private HBCUs, at $1,476.58 and $2,026.92 respectively. Average spending on administration per FTE student at private HBCUs ( $5,361.69) was nearly double that of public HBCUs ( $3,024.65) and had almost three times the variation per standard deviation (private HBCUs = $3,880.19; public HBCUs = $1,363.93). Public and private HBCUs spent similarly on auxiliary services, with public HBCUs spending $2,611.25 per FTE student and private HBCUs spending $2,838.63 per FTE student. It is important to emphasize that almost every variable for private HBCUs had median values considerably lower than the mean and large standard deviation values, suggesting that there ar e likely a handful of private HBCUs exerting disproportionate influence on the mean values. In contrast, most of the variables for public HBCUs had similar mean and median values with far lower standard deviation values. !!57 Ultimately, the descriptive statis tics for the ten years of data highlight significant variations between public and private HBCUs, as well as within each sample of institutions, that demonstrate the appropriateness of their inclusion in the model beyond their theoretical significance. Fur thermore, the significant variations underscore the importance of considering these institutions in the context of their public or private control. Table 3 . Descriptive Statistics for Institutional Characteristics of Four -Year HBCUs Between 2000-2010 Pub lic HBCUs Private HBCUs Variable Mean Median Standard Deviation Mean Median Standard Deviation Percentage of nonblack students 15.52% 8.74% 19.20% 2.91% 1.79% 3.77% Six -year graduation rate 33.11% 34.21% 10.41% 37.73% 32.26% 20.91% In-state tuition $3,850.70 $3,595.00 $1,388.66 -- -- -- Out -of-state/private tuition $9,927.32 $9,466.00 $3,082.19 $9,753.41 $9,580.00 $3,382.04 Average institutional grant aid per FTE $978.88 $832.03 $703.85 $2,144.84 $1,794.28 $1,555.79 Spending on instruction per FTE $5,950.60 $5,799.17 $1,609.46 $5,363.94 $4,766.22 $3,119.87 Spending on academic support per FTE $1,721.88 $1,586.70 $782.18 $1,546.63 $1,264.52 $1,206.08 Spending on student services per FTE $1,476.58 $1,290.16 $755.01 $2,026.92 $1,850.89 $1,059.95 Spending on administration per FTE $3,024.65 $2,715.74 $1,363.93 $5,361.69 $4,699.64 $3,880.19 Spending on auxiliary services per FTE $2,611.25 $2,441.27 $1,323.58 $2,838.63 $2,584.85 $1,363.44 Results This section will present the results of the statisti cal analysis conducted in this study. In order to examine the influence of institutional characteristics on nonblack enrollment at HBCUs during the years 2000 Ð2010, a fixed -effects panel data regression approach was employed. This !!58 approach was used to cont rol for time -invariant characteristics and to focus on the effect of the institutional characteristics of HBCUs on nonblack enrollment during that ten -year period more broadly as opposed to the effect of those variables between specific institutions. Addit ionally, due to the significant variations between public and private HBCUs presented in the previous section and the fact that institutional control is a fixed characteristic, the results will reflect the outcomes of two different statistical runs: one fo r public HBCUs and one for private HBCUs. Therefore, the following OLS model was used to investigate the effects of institutional characteristics on nonblack enrollment at HBCUs between 2000 -2010: yit (pct_nonblack) = !0 + !1 (pct_grad_rate )it + !2 (tuitionandfee02 )it* + !3 (tuitionandfee03 )it + !4 (instaid_spend_1000 )it + !5 (instruct_spend_1000 )it + !6 (acadsupp_spend_1000 )it + !7 (studserv_spend_1000 )it + !8 (admin_spend_1000 )it + !9 (aux_spend_1000 )it + !! + "it Where : ! y represents the percentage of nonblack students ! !o represents the overall intercept ! !1 (pct_grad_rate) it represents the effect of graduation rates for case i in the sample at t time period ! !2 (tuitionandfee02 )it represents the effect of in-state tuition* for case i in the sample at t time period ! !3 (tuitionandfee03 )it represents the effect of out-of-state/private tuition for case i in the sample at t time period ! !4 (instaid_spend_1000 )it represents the effect of institutional grant aid per F TE student for case i in the sample at t time period ! !5 (instruct_spend_1000 )it represents the effect of spending on instruction per FTE student for case i in the sample at t time period ! !6 (acadsupp_spend_1000 )it represents the effect of spending on acade mic support per FTE student for case i in the sample at t time period ! !7 (studserv_spend_1000 )it represents the effect of spending on student services per FTE student for case i in the sample at t time period ! !8 (admin_spend_1000 )it represents the effect o f spending on administration per FTE student for case i in the sample at t time period !!59 ! !9 (aux_spend_1000 )it represents the effect of spending on auxiliary services per FTE student for case i in the sample at t time period ! !! represents all time -invar iant variables that could influence nonblack enrollment but are not included in this study ! "it represents residual error for case i in the sample at t time period * The in -state tuition variable was used exclusively in the public HBCU model. Note: All spe nding and tuition variables were transformed by dividing per FTE figures by 1000 for easier interpretation of coefficients. It is important to note that certain variables were excluded from the final model. An attempt was made to include measures of select ivity, such as acceptance rates and 75 th percentile composite ACT scores; however, data for these measures were inconsistently reported. Additionally, a variable for average Pell grant per FTE student was included initially as an indicator for the average income level of students, but its inclusion in the model possessed no statistical or explanatory value. Finally, due to missing data for graduation rates prior to the 2003-2004 academic year, two different models for each sample of institutions were produc ed to test the strength of the models: 1) includes all observations between 2000 -2010 and 2) restricts observations to 2003 -2010. Additionally, because a fixed -effects approach was used on panel data, there is potential for within -cluster error correlation . Therefore, clustered robust standard errors were used to address the potential issue of correlated standard errors and to produce more conservative estimates. Table 4 represents the results of the OLS regression analysis using a fixed -effects approach wi th clustered robust standard errors. !!60 Table 4. Regression Analysis for the Effects of Institutional Characteristics on Nonblack Enrollment at HBCUs Between 2000 -2010 (No Lags) Variable Public HBCUs 2000-2010 Public HBCUs 2003-2010 Private HBCUs 2000-2010 Private HBCUs 2003-2010 Six -year graduation rate -.0181997 (.0151393) -.0252166 (.0134727) -.0210625* (.0079537) -.0178574* (.0061545) In-state tuition and fees .0461497 (.5375114) .2093815 (.4882733) -- -- Out -of-state/private tuition and fees -.2128044 (.1140625) -.1425392 (.0857878) .0954461 (.1066235) .1190529 (.128104) Average institutional grant aid per FTE .1044036 (.0985309) .1221302 (.0922915) -.0144039 (.0429366) -.0400106 (.0455493) Spending on instruction per FTE -.2168081 (.1 803143) -.2791163 (.1629154) .0142608 (.0741923) -.0096151 (.0914913) Spending on academic support per FTE 1.968452** (.4836124) 1.938156** (.4307524) -.1890025 (.1587041) -.1623406 (.1504499) Spending on student services per FTE -.1628289 (.3937928) .1112896 (.414478) .1010161 (.1184704) -.0612087 (.1196233) Spending on administration per FTE -.1192686 (.2304753) -.083315 (.2464594) -.0247102 (.029452) -.0429588 (.0245745) Spending on auxiliary services per FTE .3021178 (.2117668) .3076137 (.1 845381) -.0673361 (.1420998) -.2399925 (.1250437) Constant 16.21421** (1.836291) 15.14593** (1.750284) 3.171402* (.9877662) 3.686735* (1.217239) Observations 353 (40/40 institutions) 315 (40/40 institutions) 378 (43/45 institutions) 338 (43/45 instit utions) R-squared (within) .2910 .3002 .1095 .1237 F-statistic 5.26** 6.95** 1.84 2.09* Clustered robust standard errors in parentheses. All spending and tuition variables were transformed by dividing per FTE figures by 1000 for easier interpretation of coefficients. * significant at p<.05 level; ** significant at p<.001 level Although the number of observations and R -squared values were slightly higher for the models inc luding all years, the models restricted to the years between 2003 -2010 had higher F -statistics with more statistically significant p -values (both for the F -statistics of the entire model and among the predictors being tested), suggesting that there is stro nger evidence that the !!61 restricted models are better fits. As a result, all of the findings in the following sections will be discussed based on the products of the 2003 -2010 restricted models. Public HBCUs (2003-2010). For public HBCUs, the analysis indica tes that the only statistically significant predictor of nonblack enrollment is spending on academic support per FTE student. The interpretation of the coefficient in this case suggests that every $1,000 increase in spending on academic support per FTE stu dent corresponds with an almost two percent increase in nonblack enrollment (t = 4.50, p < .001). Although academic support spending was the only variable that reached statistical significance at 95 percent confidence or above, two variables fell between t he 90 Ð 95 percent confidence interval level (i.e., .10 > p > .05): both six -year graduation rates ( t = -1.87) and spending on instruction ( t = -1.71) were associated with decreases in nonblack enrollment. However, neither have much tangible value consider ing that a $1,000 increase in spending on instruction per FTE translated to only a .28 percent decrease in nonblack enrollment and an increase in one percent increase in graduation rates would produce only a .02 percent decrease in nonblack enrollment. The R-squared value for the model suggests that 30 percent of the variance in nonblack enrollment can be explained by the variables included in this model. Interpretation. Statistically, the strongest model of the two separate models was the public HBCU model . Not only was the R -squared value almost three times higher than the other models (explaining 30 percent of the variance in nonblack enrollment), but the F -statistic was also larger (6.95) with a highly statistically significant p -value ( p < .001). All of these statistics taken together suggest that the public HBCU model possesses significant explanatory power. To go along with the strength of the overall model, the public HBCU model also produced the most statistically significant result when it comes t o the effect of an institutional !!62 characteristic on nonblack enrollment. The results from the public HBCU analysis reveal that spending on academic support per FTE student possessed both statistical significance and explanatory relevance. In fact, the analy sis suggests that every $1,000 increase in spending on academic support per FTE student translates to nearly a two percent increase in nonblack enrollment. With higher numbers of students enrolled on average at public HBCUs, a two percent increase translat es to a greater impact on the total number of nonblack students enrolled. For example, using the average enrollment figures for public HBCUs (approximately 4,900 students), a two percent increase in nonblack enrollment would equal nearly 100 additional nonblack students. Or at an institution like Florida A&M University, whose enrollment is closer to 13,000, a two percent increase would equal approximately 260 nonblack students. In both of the aforementioned examples, even a two percent increase in nonblack enrollment could have a significant impact on the demographic makeup and, potentially, the campus climate of a public HBCU. Private HBCUs (2003-2010). The analysis indicates that graduation rates is the only significant predictor of nonblack enrollment at private HBCUs. The interpretation of the coefficients for that measure suggests that every one percent increase in graduation rates corresponds with an approximately .018 percent decrease in nonblack enrollment (t = -2.90, p < .006). In addition to the v ariable that was found significant at the 95 percent confidence interval level, two variables fell between the 90 Ð 95 percent confidence interval level: spending on administration and spending on auxiliary services. Both variables were found to be negativ ely associated with nonblack enrollment, with every $1,000 spent on administration corresponding to a -.04 (t = -1.75) percent decrease in nonblack enrollment and every $1,000 increase in !!63 spending on auxiliary services per FTE student corresponding with an approximately .24 percent decrease in nonblack enrollment (t = -1.92). The R -squared value for the model suggests that 12 percent of the variance in nonblack enrollment can be explained by the variables included in this model. Interpretation. Although the private HBCU model (R -squared = .12) does not possess the same strength as the public HBCU model, it does contain explanatory value in its finding that higher graduation rates predict lower levels of nonblack enrollment. The coefficient for the statistically significant variable in this model suggest that every one percent increase in graduation rates corresponds with a .018 percent decrease in nonblack enrollment. This small decreases in nonblack enrollment may seem inconsequential, but when considered in the context of the extremely low levels and very little variation of nonblack enrollment at private HBCUs, these findings contain more explanatory relevance than one might initially expect. In more tangible terms, it would take a large change in six -year graduation rates to have an effect on nonblack enrollment, with a 20 percent increase in the graduation rates of private HBCUs translating to only an approximately .35 percent decrease in nonblack enrollment. In the case of the average private HBCU, that . 3 percent decrease would translate to a difference of approximately one student. Although the effect of six -year graduation rates has little chance of influencing nonblack enrollment in any meaningful and observable way at private HBCUs, the fact that it i s statistically significant at all is in itself significant and will be explained in the following chapter. The variables that fell between the 90 -95 percent confidence interval level have limited tangible value when it comes to their effect on nonblack e nrollment. The coefficient for spending on auxiliary services suggests that it would take an additional $4,000 spent per FTE student to !!64 change nonblack enrollment by one percent. Considering the median value for academic support spending is approximately $2,500 with a standard deviation of approximately $1,300, it is unlikely that the average private HBCU would be able to invest the necessary resources to produce any tangible change in nonblack enrollment. The same could be said for spending on administrati on, which would require an additional $5,000 spent per FTE student to change nonblack enrollment by a quarter of one percent. Although the $5,000 figure i s close to the median value for this spending category (median ! $4,700), it would take a private HBCU doubling its spending on administration to achieve that quarter of one percent change. Again, although these variables may not seem to have tangible infl uence over the percentage of nonblack students enrolled at private HBCUs, the fact that they are statistically significant at all is important to acknowledge and discuss further. Impact of confounding variable threshold (ICVT). As m entioned previously, be cause this study is not a randomized experiment, a conscious effort must be made to controlling for unobserved influences when inferring causality. Accordingly, efforts were taken to test the robustness of the findings using methods developed by Frank (200 0) and Frank et al. (2013) which quantify how much bias there must be in order to invalidate an inference. Stated in statistical terms, FrankÕs (2000) method answers the question: Given that a variable ÒxÓ is statistically significant, how much of an estim ate must be due to bias to invalidate an inference? Specifically, FrankÕs (2014) KonFound -it!© program, which allows a researcher to input specific data (estimated effect of variable, observed t -critical value, number of covariates, sample size, etc.) from a regression output to determine the percent of cases from a sample that would need to be replaced to invalidate an inference, was used to test the robustness of the findings from the regression analysis for public and private HBCUs between the years 2003 -2010. !!65 According to FrankÕs (2014) KonFound -it!© results, for the finding in the public HBCU model that academic support spending per FTE student (! = 1.94, SE = .431 ) was statistically significant, over 55 percent of the cases from the sample would need to be replaced in which there is an effect of zero to make academic support spending irrelevant. Similarly, the KonFound -it!© results suggested that approximately 32 percent of the cases would have to be replaced to invalidate the finding that graduation rates at private HBCUs (! = -.018, SE = .006) are a statistically significant negative influence on nonblack enrollment. The results of these tests a re significant for one important reason: In most studies, a sample is drawn from a larger population to test a hypothesis. However, this study did not have to use a sample because every institution in the population was represented in the analysis. In othe r words, there are no additional sample cases (i.e., institutions) that could replace the cases in this study and, therefore, the robustness of the findings that academic support spending influences nonblack enrollment at public HBCUs is strengthened and t he ability to generalize to the population of all public HBCUs is greatly improved. Lagged variable analysis. The primary model used in this study analyzed the effects of predictor variables that were drawn from the same year as the dependent variable (e. g., the influence of spending variables from 2002 on nonblack enrollment from 2002). However, it is possible that the college -choice decisions of students may be influenced by the result of an institutionÕs decisions/behavior/performance in the years leadi ng up to the point at which a student chooses a college to attend. Therefore, using the same fixed effects panel data regression approach, two additional analyses were conducted for public and private HBCUs between 2003 -2010: one analysis with one -year lag s for the predictor variables ( t-1) and one analysis with two -year lags for the predictor variables ( t-2). !!66 yit (pct_nonblack) = !0 + !1 (pct_grad_rate )i(t -1) or (t -2) + !2 (tuitionandfee02 )i(t -1) or (t -2)* + !3 (tuitionandfee03 )i(t -1) or (t -2) + !4 (instaid_spend_1000 )i(t -1) or (t -2) + !5 (instruct_spend_1000 )i(t -1) or (t -2) + !6 (acadsupp_spend_1000 )i(t -1) or (t -2) + !7 (stud serv_spend_1000 )i(t -1) or (t -2) + !8 (admin_spend_1000 )i(t -1) or (t -2) + !9 (aux_spend_1000 )i(t -1) or (t -2) + !! + "it !!67 Table 5. Regression Analysis for the Effects of Institutional Characteristics on Nonblack Enrollment at HBCUs Between 2003 -2010 with One - and Two -Year Lags Variable Public HBCUs 2003-2010 w/ One -Year Lag Public HBCUs 2003-2010 w/ Two -Year Lag Private HBCUs 2003-2010 w/ One -Year Lag Private HBCUs 2003-2010 w/ Two -Year Lag Six -year graduation rate -.0068955 (.0145413) .0005725 (.0136481) -.010433 (.0094767) -.0150541* (.0071501) In-state tuition and fees .0951888 (.5886867) .1192733 (.515942) -- -- Out -of-state/private tuition and fees -.2589376** (.1095921) -.1728239 (.0889665) .0684721 (.1110521) -.003454 (.0959336) Average institutional grant aid per FTE .0932757 (.100242) .0240676 (.080946) -.0309119 (.042183) -.003557 (.0450396) Spending on instruction per FTE -.0846652 (.1827749) -.0839849 (.2098588) -.0115489 (.0764463) .0443794 (.11 11161) Spending on academic support per FTE 1.731752** (.4700118) 1.176629* (.4396413) -.1228936 (.1429446) -.2842294* (.1329587) Spending on student services per FTE .1536694 (.4795277) .4970107 (.5076051) .147386 (.1166998) .114481 (.1294597) Spen ding on administration per FTE -.0198428 (.1990045) .0687136 (.1556758) .006413 (.0169039) .0211105 (.0191144) Spending on auxiliary services per FTE .3358522* (.1115772) .3180176* (.1360052) .0422984 (.0869959) -.0187938 (.1041414) Constant 14.8 1916** (1.837867) 13.92161** (1.999925) 2.482165* (.8716942) 3.337966** (.7547756) Observations 312 (40/40 institutions) 272 (40/40 institutions) 336 (43/45 institutions) 294 (43/45 institutions) R-squared (within) 0.2828 .2141 .0679 .0884 F-statist ic 4.36** 2.99* 1.19 1.95* Clustered robust standard errors in parentheses. All spending and tuition variables were transformed by dividing per FTE figures by 1000 for easier interpretation of coefficients. * significant at p<.05 level; ** sign ificant at p<.001 level The regression analyses using lagged predictor variables did not possess the same strength as the initial models as measured by the R -squared values, bu t did produce statistically significant findings in both the public and private HBCU models. Most notably, the strongest predictors of nonblack enrollment from the initial models remained significant in the models !!68 with lagged variables (public HBCUs = acad emic support spending per FTE student; private HBCUs = graduation rates). For public HBCUs, academic support spending continued to exert the strongest influence on nonblack enrollment, although its level of influence and statistical significance decreased as the amount of lag increased from one ( ! = 1.731752, p < .001) to two years ( ! = 1.176629, p < .05). In more tangible terms, these coefficients translate to an approximately 1.7 and 1.2 percent increase in nonblack enrollment with every $1,000 increase in academic support spending per FTE student. For private HBCUs, the graduation rates variable lost its statistical significance in the model with a one -year lag, but regained its significance in the model with a two -year lag ( ! = -.0150541, p < .05). In the two -year lag model, the coefficient translates to an approximately .015 percent decrease in nonblack enrollment for every one percent increase in graduation rates at private HBCUs. Since the coefficients for these variables are nearly the same as the initial model, the explanatory significance is therefore similar to the one provided for both of these variables in the discussion of the initial model. There were also some variables that emerged as statistically significant in the models with lagged var iables that were not statistically significant in the initial models. For public HBCUs, two variables were statistically significant: out -of-state tuition and auxiliary spending per FTE student. In both the one - and two -year lag models, auxiliary spending was found to be a positively statistically significant variable ( ! = .3358522 and .3180176 respectively, p < .05), which translates to an approximately .3 percent increase in nonblack enrollment with every $1,000 increase in auxiliary spending per FTE student. Out -of-state tuition was also an extremely statistically si gnificant negative influence in the one -year lag model ( ! = -.2589376, p < .001), but did not maintain its significance in the two -year lag model (.10 < p < .05). For every !!69 additional $1,000 in out -of-state tuition, nonblack enrollment decreases by approx imately a quarter of one percent. Conclusion Interestingly, each model produced findings unique from one another Ñwhich supports the need for separate public and private models. However, there were also statistically significant variables common to all mod els. For public HBCUs, spending on academic support per FTE student was found to be a strongly positive predictor of nonblack enrollment in models with and without the lag, while graduation rates had statistically significant negative effects on nonblack enrollment at private HBCUs with and without the lag. Other variables that emerged as statistically significant in the models with a one - and two -year lag include out -of-state tuition (negative) and auxiliary spending per FTE student (positive) for public H BCUs, and spending on academic support per FTE student (negative) at private HBCUs. Furthermore, a test of the robustness of the findings suggest that over 55 percent of the cases would have to be replaced in which the effect is zero in order for the infer ence to be invalidated for the finding that academic support spending influences nonblack enrollment at public HBCUs, and approximately 32 percent of the cases would have to be replaced to invalidate the finding that graduation rates at private HBCUs are a statistically significant negative influence on nonblack enrollment. For the purposes of explaining the results and discussing the implications for research, policy, and practice, the final chapter will focus only on the variables that were statistically significant in two out of three of the models, which were academic support spending and auxiliary services spending for public HBCUs, and graduation rates for private HBCUs. The variables that will be discussed have been highlighted in Table 6, which juxt aposes the models being used for the final analysis. !!70 Table 6. All Models with Emphasis on Variables Found to be Statistically Significant in Two Out of the Three Models Between 2003 -2010 Variable Public HBCUs 2003-2010 w/No Lag Public HBCUs 2003-2010 w/ One-Year Lag Public HBCUs 2003-2010 w/ Two -Year Lag Private HBCUs 2003-2010 w/No Lag Private HBCUs 2003-2010 w/ One -Year Lag Private HBCUs 2003-2010 w/ Two -Year Lag Six -year graduation rate -.0252166 (.0134727) -.0068955 (.0145413) .0005725 (.0136481) -.0178574* (.0061545) -.010433 (.0094767) -.0150541* (.0071501) In-state tuition and fees .2093815 (.4882733) .0951888 (.5886867) .1192733 (.515942) -- -- -- Out -of-state/private tuition and fees -.1425392 (.0857878) -.2589376** (.1095921) -.1728239 (.0889665) .1190529 (.128104) .0684721 (.1110521) -.003454 (.0959336) Average institutional grant aid per FTE .1221302 (.0922915) .0932757 (.100242) .0240676 (.080946) -.0400106 (.0455493) -.0309119 (.042183) -.003557 (.0450396) Spending on inst ruction per FTE -.2791163 (.1629154) -.0846652 (.1827749) -.0839849 (.2098588) -.0096151 (.0914913) -.0115489 (.0764463) .0443794 (.1111161) Spending on academic support per FTE 1.938156** (.4307524) 1.731752** (.4700118) 1.176629* (.4396413) -.162340 6 (.1504499) -.1228936 (.1429446) -.2842294* (.1329587) Spending on student services per FTE .1112896 (.414478) .1536694 (.4795277) .4970107 (.5076051) -.0612087 (.1196233) .147386 (.1166998) .114481 (.1294597) Spending on administration per FTE -.083315 (.2464594) -.0198428 (.1990045) .0687136 (.1556758) -.0429588 (.0245745) .006413 (.0169039) .0211105 (.0191144) Spending on auxiliary services per FTE .3076137 (.1845381) .3358522* (.1115772) .3180176* (.1360052) -.2399925 (.1250437) .0422 984 (.0869959) -.0187938 (.1041414) Constant 15.14593** (1.750284) 14.81916** (1.837867) 13.92161** (1.999925) 3.686735* (1.217239) 2.482165* (.8716942) 3.337966** (.7547756) Observations 315 (40/40 institutions) 312 (40/40 institutions) 272 (40/40 institutions) 338 (43/45 institutions ) 336 (43/45 institutions) 294 (43/45 institutions) R-squared (within) .3002 0.2828 .2141 .1237 .0679 .0884 F-statistic 6.95** 4.36** 2.99* 2.09* 1.19 1.95* Clustered robust standard errors in parentheses. All s pending and tuition variables were transformed by dividing per FTE figures by 1000 for easier interpretation of coefficients. * significant at p<.05 level; ** significant at p<.001 level !!71 Chapter 5 Discussion and Implications The previous chapter presented the results from the quantitative analysis of the effect of institutional characteristics on nonblack enrollment at HBCUs between the years 2000 -2010. Using a fixed -effects pan el data regression analysis, two different models were used to test the effect of institutional characteristics on nonblack enrollment: 1) only public HBCUs and 2) only private HBCUs. The two separate regression models produced unique sets of findings. Add itionally, further analyses were conducted for public and private HBCUs using one - and two -year lags for the dependent variables. These models produced findings similar to those of the initial models, but also introduced new influential factors. The follow ing sections will summarize the study, connect the findings to previous research, and discuss the implications of the findings for policy, practice, and future research. Summary of Study Determining why nonblack students enroll at HBCUs is critically impo rtant to better understanding the college -choice process for all students. Aside from a limited number of studies, little research has been conducted examining the college choice of nonblack students at HBCUs. The literature is particularly scarce as it re lates to the higher education institutionÕs role in influencing nonblack enrollment patterns (Daniels, 2008). And because more comprehensive and contextual conceptual models of college choice did not exist until more recently, what we do know about the en rollment of nonblack HBCU students is not expressed in terms that are aligned with more contemporary knowledge of the college -choice process. In addition to the limited understanding of college choice for nonblack students at HBCUs, little is known about organizational behavior that may influence the college -choice !!72 decisions for students at HBCUs . In fact, a lthough Minor (2008b) was referencing governance at HBCUs when he went so far as to call governance at HBCUs an Òenigma,Ó his comment is indicative of the general lack of understanding about the operation of these institutions and scant research within the field. Bastedo (2012) suggests that while a plethora of higher education research more broadly has focused on organizational issues such as governance and elite leaders, major issues such as understanding who will attend college and why, from an organizational perspective, have been neglected. Additionally, previous research and the frames that we use to view organizational issues have not been geared t oward understanding diverse populations or organizations (e.g., HBCUs) (Bastedo, 2012, Minor, 2005). In order to better understand racial diversity on HBCU campuses and broaden the scope of scholarly wo rk on HBCUs, this study attempts to provide a nuanced view of nonblack enrollment at HBCUs from an organizational perspective. To support and extend these lines of inquiry on organizational behavior and its relationship to diversity at HBCUs, the study seeks to understand which institutional characteristics of HBCUs are most closely associated with different levels of nonblack student enrollment. PernaÕs (2006) conceptual model of college choice, specificall y the higher education context, serve d as a lens through which this issue was viewed. For the pur pose of this study, the focus shifted from the individual studentÕs decision -making process to the ways in which the higher education context may influence enrollment decisions for different populations of students. Ult imately, this investigation used enrollmen t patterns to better understand an element of college choice . Specifically, this study looked at organizational f actors associated with non black enrollment at HBCUs. Organizational factors are part of the higher education context identified by Perna in lev el three of her proposed conceptual !!73 model, which also includes the geographic location and marketing and recruitment efforts of higher education institutions. The primary research questions explored in this study were : ¥ Are there institutional characteris tics of HBCUs that influence d nonblack enrollment patterns between the years 2000 -2010? o What organizational characteristics, if any, have the strongest relationship to nonblack enrollment at HBCUs (e.g. tuition, institutional spending, graduation rates, r etention rates, etc.)? o Furthermore, what, if anything, do those organizational characteristics suggest about how institutional context shapes college choice? In order to examine the influence of institutional characteristics on nonblack enrollment at HBCUs during the years 2000 Ð2010, a fixed -effects panel data regression approach was employed and clustered robust standard errors were used to address the potential issue of correlated standard errors and to produce more conservative estimates. Due to missing data that weakened the statistical significance of models including all years between 2000 -2010, the focus of the analysis was restricted to the years 2003 -2010. Furthermore, due to the significant variations in the predictor variables between public and p rivate HBCUs, two different statistical models were run: one for public HBCUs and one for private HBCUs. Each model produced distinctly different from one another: For public HBCUs, spending on academic support per FTE student was found to be a strongly p ositive predictor of nonblack enrollment, whereas graduation rates had a statistically significant negative effect on nonblack enrollment at private HBCUs. The public HBCU model produced the most statistically significant result when it comes to the effe ct of an institutional characteristic on nonblack enrollment. The results of the analysis !!74 suggest that every $1,000 increase in spending on academic support per FTE student translates to a nearly two percent increase in nonblack enrollment. Using the avera ge enrollment figures for public HBCUs (approximately 4,900 students), a two percent increase in nonblack enrollment would equal nearly 100 additional nonblack students. Or at an institution like Florida A&M University, whose enrollment is closer to 13,000 , a two percent increase would equal approximately 260 nonblack students. The coefficient for the statistically significant variable in the private HBCU model suggests that every one percent increase in graduation rates corresponds with a .018 percent dec rease In more tangible terms, it would take a large change in six -year graduation rates to have an effect on nonblack enrollment, with a 20 percent increase in the graduation rates of private HBCUs translating to only an approximately .35 percent decrease in nonblack enrollment. In the case of the average private HBCU, that .3 percent decrease would translate to a difference of approximately one student. Furthermore, a test of the robustness of the findings suggest that over 55 percent of the cases would h ave to be replaced in which the effect is zero in order for the inference to be invalidated for the finding that academic support spending influences nonblack enrollment at public HBCUs, and approximately 32 percent of the cases would have to be replaced t o invalidate the finding that graduation rates at private HBCUs are a statistically significant negative influence on nonblack enrollment. Because it is possible that the college -choice decisions of students may be influenced by the result of an institutio nÕs decisions/behavior/performance in the years leading up to the point at which a student chooses a college to attend, two additional analyses were conducted for public and private HBCUs between 2003 -2010 using the same fixed effects panel data regression !!75 approach: one analysis with one -year lags for the predictor variables ( t-1) and one analysis with two-year lags for the predictor variables ( t-2). Although t he regression analyses using lagged predictor variables did not possess the same strength as the i nitial models as measured by the R -squared values, they did produce statistically significant findings in both the public and private HBCU models. Most notably, the strongest predictors of nonblack enrollment from the initial models remained significant in the models with lagged variables (public HBCUs = academic support spending per FTE student; private HBCUs = graduation rates). Other variables emerged as statistically significant in the models with a one - and two -year lag include out -of-state tuition (negative) and auxiliary spending per FTE student (positive) for public HBCUs, and spending on academic support per FTE student (negative) at private HBCUs. However, for the purposes of explaining the results and discussing the implications for research, poli cy, and practice, the final chapter will focus only on the variables that were statistically significant in two out of three of the models, which were academic support spending and auxiliary services spending for public HBCUs, and graduation rates for priv ate HBCUs. The variables that will be discussed were highlighted in Table 6, which juxtaposes the models being used for the final analysis. Discussion of Findings There seems to be one overarching and conclusive implication effecting the interpretation of all of the findings from this study: institutional context matters. Just as scholars have stressed the importance of acknowledging the unique characteristics of HBCUs that distinguish them from the broader higher education community (Griffin & Hurtado, 201 1; Minor, 2008; Ryan, 2004), the results of this study also reveal the importance of taking the public -private context into consideration when studying HBCUs. Specifically as it relates to this study, distinguishing !!76 between public and private HBCUs seems t o be a critical factor in nonblack students choosing to attend an HBCU. From an organizational theory perspective, this is an important distinction to be made when conducting research on public and private institutions since considering them in the aggrega te can lead to overgeneralizations and ignore their varying interests, accessibility, and institutional missions (Griffin & Hurtado, 2011; Perry & Rainey, 1988). From a college choice perspective, students would be unable match their educational and social needs to an institution, as Nora (2004) suggests occurs in the college -choice process, without sufficient institutional diversity in program, residential, and other offerings (Harris, 2013). Viewing the effects of institutional characteristics on nonblac k enrollment at HBCUs with institutional context in mind is essential. Lee (2015), in his chapter titled ÒMoving Beyond Racial and Ethnic Diversity at HBCUs,Ó presents compelling data to underscore the necessity of not viewing HBCUs as a monolithic group, but rather a heterogeneous group of institutions with some shared histories and missions. As opposed to the broad stroke that is often used to portray HBCUs, Lee (2015) points to some of the following variations to lay the foundation for a more nuanced und erstanding of HBCUs: ¥ HBCUs are not just four -year institutions. Approximately 12% of all HBCUs are two -year institutions. ¥ Carnegie classifications vary greatly. 48% baccalaureate universities, 24% masterÕs universities, 12% associates institutions, 10% res earch universities, 4% seminaries, and 2% medical schools. ¥ Percentage of Pell grant recipients differ between institutional type. Private HBCUs have the highest percentage of Pell grant recipients at 77%, while public HBCUs have the lowest percentage at 66 %. !!77 Furthermore, on top of the aforementioned distinctions , the difference s in institutional context among HBCUs is demonstrated in this study by the fact that 1) both the institutional characteristics influencing nonblack enrollment and the direction of th at influence (positive or negative) differ entirely between public and private HBCUs, 2) the overall statistical strength of the public HBCU model is higher, and 3) the statistically significant institutional characteristics have larger coefficients and sm aller p -values at public HBCUs compared to private HBCUs. Juxtaposing the influential institutional characteristics at public and private HBCUs provides some insight into the way nonblack students may be perceiving investments in various aspects of instit utional operations at HBCUs. The results of the analysis revealed that there were three variables that were statistically significant in at least two out of the three models (2003 -2010 with no lag, 2003 -2010 with one -year lag, and 2003 -2010 with two -year l ag), which suggest that nonblack students are more likely to enroll at a public HBCU that spends more per FTE student on academic support and auxiliary spending, while nonblack students are less likely to enroll at a private HBCU that graduates students at a higher rate (see Figure 6). !!78 Figure 6 . Influential Institutional Characteristics for Nonblack Students Attending HBCUs (Public vs. Private) Understanding how each of those spending categories is defined provides additional insight i nto what messages those expenditures may be sending to nonblack students. Academic support spending (variable name: acadsupp01) is defined as Ò expenses of activities and services that support the institution's primary missions of instruction, research, and public service. It includes the retention, preservation, and display of educational materials (for example, libraries, museums, and galleries); organized activities that provide support services to the academic functions of the institution (such as a demo nstration school associated with a college of education or veterinary and dental clinics if their primary purpose is to support the instructional program); media such as audiovisual services; academic administration (including academic deans but not depart ment chairpersons); and formally organized and separately budgeted Public HBCUs Nonblack students are more likely to attend a public HBCU if the institution spends more per FTE student on academic support (all models). Nonblack students are more likely to attend a public HBCU if the institution spends more per FTE student on auxiliary services (in models w/one- and two-year lag). Private HBCUs Nonblack students are less likely to attend a private HBCU if the institution has a higher graduation rate (all models except one-year lag). !!79 academic personnel development and course and curriculum development expensesÓ ( Delta Cost Project, 2014 ). Auxiliary services spending (variable name: auxiliary01) is defined as Ò expenses associated with essentially self -supporting operations of the institution that exist to furnish a service to students, faculty, or staff, and that charge a fee that is directly related to, although not necessarily equal to, the cost of the service. Examples are residence halls, food services, student health services, intercollegiate athletics (only if essentially self -supporting), college unions, college stores, faculty and staff parking, and faculty housingÓ ( Delta Cost Project, 2014 ). After establishing the institutional characteristics with the strongest influence on nonblack enrollment at HBCUs, understanding exactly what is included in these respective spending categories, and considering them in the context of their public and private control, some qu estions still remain unanswered: ¥ How do these findings relate to prior research? ¥ Why might these specific institutional characteristics be important to nonblack students attending HBCUs? o Further, why might the influential institutional characteristics di ffer for nonblack students at public versus private HBCUs? ¥ How do these findings contribute to our understanding of the higher education context in PernaÕs conceptual model of college choice? ¥ What implications do the possible explanations for the influence of these institutional characteristics have for policy, practice, and future research on nonblack students at HBCUs? !!80 Previous research on the effect of institutional expenditures on college choice is extremely limited and the findings primarily revolve a round the general idea that higher education institutions do respond to market forces, such as increased competition for consumers/students (Epple, Romano, & Sieg, 2006; Hoxby, 1997, 2009). Jacob, McCall, and Stange (2013) point out, however, that these s tudies primarily investigated the effect of price, geographic location, and academic aspects of colleges on postsecondary enrollment. In response to the void in research examining the effect of institutional expenditures on college choice, Jacob et al. (20 13) conducted a study in which the focus was determining the effect of amenity/auxiliary versus academic spending in the college -choice process. The findings from their study suggest that college -choice decisions are, in fact, influenced by these two afore mentioned spending categories. Specifically, they found that relatively few students were willing to pay for a postsecondary institution that spends more on instruction, while low -ability, high -income students possessed the greatest willingness to pay for amenities, and high -ability students found greater value in high academic quality institutions. It is important to note, however, that none of the previous research examined these issues in the context of HBCUs. Because student -level data was not used in this study, no comparison can be made regarding the willingness of nonblack students to pay more for HBCUs with particular characteristics based on student ability or demographics. However, the results from this current study do reveal some interesting fin dings that relate to the broader discussion around the importance of amenity spending in the college -choice process. For example, although Jacob et al. (2013) suggest that students were generally more willing to pay for an institution with higher amenity/a uxiliary spending, the current study found that amenity/auxiliary spending had only a slightly positive effect on nonblack enrollment at public HBCUs in the lagged models and no !!81 effect on nonblack enrollment at private HBCUs. The fact that amenity/auxiliar y spending does not exert the same consistent level and direction of influence on the college -choice process for nonblack students at all HBCUs as it does the general population of college -going students suggests that nonblack students at HBCUs likely view the higher education context of the college -choice process significantly different from their non -HBCU peers. Particularly, because nonblack students at public HBCUs seem to be positively influenced by spending on auxiliary services, this may speak to the different expectations nonblack students have of HBCUs depending on the various offerings found at public versus private HBCUs. In addition to the amenity/auxiliary spending being an inconsistent factor in the college -choice process of nonblack students a ttending HBCUs, there is much to be said about the consistent and strong positive influence academic support spending has on nonblack enrollment at public HBCUs. From a statistical perspective, the strength of this spending category is difficult to dismiss in light of the fact that it was significant in every model and a test of robustness suggested that over 55 percent of cases in the study would have to be replaced with cases in which the effect was zero in order to make academic support spending insignif icant. And considering that academic support spending includes institutional functions such as demonstration school s associated with a college of education or veterinary and dental clinics , this finding seems to align with previous findings from Conrad et al. (1997) and Daniels (2008) that suggest white students at public HBCUs are influenced by program offerings in high -demand professions. Therefore, as an example, an institution like Delaware State University Ñthat not only has a veterinary science program , but also offers opportunities to supplement the academic experience with practical applications on their advanced farms and in their research extension programs Ñmay be able to attract more nonblack students who perceive these offerings as unique !!82 opportun ities to develop critical professional skills. This may also explain the insignificance of academic support spending in the private HBCU model, since many private HBCUs are more likely to be smaller institutions with a liberal arts focus (not relating to v ocational or technical education), which means that they are less likely to offer similar professionally focused co -curricular services as public HBCUs with missions and funding sources geared more toward serving the needs of the community/state through ap plied research and vocational preparation. Although graduation rates are not as easily controlled by an institution as are decisions to allocate money in different areas of institutional operations, the statistically significant negative association of gr aduation rates to nonblack enrollment does seem to serve as a reflection of the intentions of private HBCUs. Consider that Fisk University, Howard University, Morehouse College, Spelman College, and Claflin University dominate the top of the list when it c omes to graduation rates among all HBCUs during the years being observed in this study. Then consider that these same institutions are all privately controlled, are liberal arts institutions, have the highest composite ACT scores for institutions that repo rted this data, and all have some of the lowest nonblack enrollment Ñsome barely registering one percent nonblack enrollment. Essentially, the combination of these forces leads to the following result: the graduation rates variable is serving as a proxy for ÒprestigeÓ (i.e., more selective, black liberal arts colleges). Beyond the statistical significance, though, this finding may suggest that three things could be happening: 1. private HBCUs are doing little to attract and recruit nonblack students (whether i n order to adhere to their historical mission or otherwise) , and/or !!83 2. considering that these institutions are often affectionately referred to as meccas for black education or as the ÒBlack Ivy League,Ó nonblack students may be receiving a message that these institutions are Òelite,Ó exclusively black, and not for them. Furthermore, it is worthy to mention and consider other variables that emerged as statistically significant in the models with one - and two -year lags but did not meet the criteria of being si gnificant in two of the three models. For example, spending on academic support had inverse influences on nonblack enrollment depending on institutional control in the models with lags. As far as academic support spending is concerned, its opposite and neg ative relationship to nonblack enrollment at private HBCUs may suggest that the products of this spending category differ from the products at public HBCUs that are attracting nonblack students. And regarding auxiliary services, the positive relationship t o nonblack enrollment at public HBCUs may suggest that nonblack students may be more interested in the amenities this spending category produces than nonblack students at private HBCUs. The negative influence of higher out -of-state tuition almost seems int uitive, but itÕs worth pointing out that this also alludes to previous research that found low cost and proximity to home to be a key factor in white students choosing to attend an HBCU (Brown & Stein, 1972; Conrad et al., 1997). Regardless of the potentia l reasons for the statistical significance, the fact that these variables are influential in some of the models suggest that it is worthy of further investigation in future studies. The findings from this study also seem to suggest that institutional chara cteristics such as institutional expenditures did in fact exert significant influence in the college -choice process for nonblack students between the years 2003 -2010 and should be considered when examining the higher education context of PernaÕs (2006) con ceptual model of college choice. In alignment !!84 with PernaÕs assertions, the results of this study point to the preference of students (nonblack in this case) to attend higher education institutions (HBCUs in this case) with certain characteristics. The find ings also suggest that the lower levels of the model interact with the higher education context, such that the individual habitus influences how messages from the higher education institution are received by the student who is trying to determine how a par ticular institution can address his/her concerns regarding expected costs and benefits. In the context of HBCUs, it seems evident that studentsÕ race (nonblack in this case) and cultural and social capital do affect the ways they determine the expected be nefits of attending an HBCU and if they perceive an HBCU as a good fit socially and academically. For example, considering that academic support spending emerged as a significant factor to nonblack students attending public HBCUs, it is possible that the i nteractions nonblack students have with individuals in their own social networks (family, high school teachers/counselors, friends, etc.) are influencing them to value certain institutional characteristics (such as vocationally oriented program offerings) over others and, thus, making public HBCUs an option in their college -choice process. On the other hand, there is an implicit assumption that nonblack students were, in fact, considering multiple institutions and choosing to attend an HBCU, as opposed to p otentially lacking cultural capital or awareness altogether and attending as a product of circumstance or convenience. Taken a step further, the findings from this study bring into question how college choice for students considering HBCUs may or may not fit into a one -size-fits -all model such as PernaÕs. At the very least, this study makes a case for layers of PernaÕs model, such as layer 3 (higher education context), being more nuanced and acknowledging how different variables may be more influential tha n others depending on the type of institutions students are considering. In !!85 other words, PernaÕs model may operate very differently and include different variables for students considering HBCUs and other MSIs versus PWIs in their set of schools, or for st udents primarily considering private institutions versus public institutions. For example, although this study did not investigate layer 4 (the social, economic, and policy context) of PernaÕs model, federal and state policies that affect HBCUs differently than PWIs (i.e., the Parent PLUS loan crisis from the early 2010s) support the idea of HBCU college -choice existing in a far different dimension than non -HBCUs and, thus, deserving of special consideration. Implications of Findings Future research. Becau se this study was narrowly tailored to quantitatively examine nonblack enrollment at HBCUs during the years 2003 -2010 through the lens of PernaÕs (2006) conceptual model of college choice using IPEDS data exclusively, there are myriad opportunities to exte nd this line of inquiry. From a theoretical perspective, PernaÕs conceptual model of college choice is general and inclusive, which allows researchers to fit more specific theoretical constructs into the more broadly defined categories of the model. The va riables examined in this study are a perfect example: although institutional expenditures were not specifically included in PernaÕs model or discussed in her literature, the more broadly defined category of ÒInstitutional CharacteristicsÓ allowed me, as th e researcher, to justify the inclusion of those variables based on previous scholarship and my interpretation of the category. This is certainly a positive aspect of the model; however, as researchers continue to examine college choice through this framewo rk, it will be important to update the model to include relevant concepts as they emerge. Therefore, the findings of this study, alongside previous research, seem to suggest that two theoretical constructs should be included in any future discussions on co llege choice, particularly within the !!86 higher education context of PernaÕs conceptual model: the important variations between public and private institutions and the influence of institutional expenditures. Specifically related to the study of HBCUs, this study followed the suggestion of scholars to study HBCUs within their unique institutional context that, as described through the literature review, sets them apart from non -HBCU institutions. In doing so, and not surprisingly, this study produced results that lie in contradiction to broader studies (such as Jacob et al., 2013) suggesting that spending on auxiliary services/amenities is influential in the college choice process for (mostly nonblack) students in the aggregate. Furthermore, the quantitative (i.e., amount spent on various institutional functions) and qualitative (i.e., missions and histories) differences between public and private HBCUs supported the importance of conducting organizational studies with public and private control in mind. Taken together, the findings from this study affirm the need to continue researching HBCUs using culturally sensitive approaches that acknowledge the factors that make them distinctive organizations within the U.S. higher education system. Methodologically, di fferent approaches can add depth to our understanding of the influence of institutional characteristics in the college -choice process for nonblack students at HBCUs. Quantitative approaches may start by identifying what is not being captured through the IPEDS dataset and find ways to measure the effect of those variables. For example, spending on marketing and recruitment efforts is not explicitly gathered through IPEDS. However, marketing and recruitment could be a significant factor in the college -choice process for nonblack HBCU students. And r elated to the previous point, it may be necessary to identify additional datasets that will allow researchers access important data and to construct multilevel models that will control for variables not included in this study Ñparticularly, variables present in other levels of !!87 PernaÕs conceptual model, such as student academic background, socioeconomic status, geographic location, cultural and social capital, school and community context, etc. Furthermore, to the ext ent possible considering the relatively low number of nonblack students at HBCUs, additional insight would be gained by disaggregating nonblack students to investigate whether differences in the influence of institutional characteristics exist between vari ous demographic groups included in nonblack. Hidden in the statistics regarding nonblack enrollment at four -year HBCUs is the fact that white students comprise nearly 80 percent of nonblack enrollment during the ten -year period in this study. Therefore, th e aggregate findings for nonblack student enrollment at HBCUs may be hiding important variations in the college -choice process for different demographic groups. Moreover, scholars have not yet sufficiently investigated the implications of the relatively ne w IPEDS category ÒTwo or more racesÓ or the previously included ÒRace/ethnicity unknownÓ variable in the study of race in higher education, especially as it relates to the study of race and nonblack enrollment at HBCUs. One study examined the rise in the p ercentage of students in the ÒUnknownÓ category and found that a sizeable portion of students in this category are white (Smith, Moreno, Clayton -Pedersen, Parker, & Teraguchi, 2005). The ÒTwo or more racesÓ variable was not available for all of the years between 2000 -2010, so it was not necessarily an issue for this study, but the same Smith et al. (2005) study suggested that capturing what is truly reported in ÒTwo or more racesÓ has implications for the accuracy of other racial categories as well. It will be important for researchers using post -2010 data to understand the inclusion of these variables in future studies since recent IPEDS data suggest that, regarding nonblack enrollment at HBCUs, ÒTwo or more racesÓ and ÒRace/ethnicity unknownÓ combined are second only to white student enrollment. !!88 Quantitative studies, however, have inherent limitations by reducing data strictly to numbers. Therefore, future qualitative studies will be necessary to develop a more robust understanding of nonblack HBCU student college -choice. In conducting qualitative studies, researchers must a cknowledge the agency that institutions have in the college -choice process for nonblack stude nts and include this concept when developing future studies on nonblack students attending HB CUs. PernaÕs (2006) model clearly demonstrates that a studentÕs college -choice decision is not made in isolation, but that it is ultimately influenced by many other factors. Because the research, including this study, provides evidence that the higher educ ation context does exert influence in the college -choice process, it is critical that future qualitative studies control for that influence and gather data accordingly. Collecting qualitative data that recognizes the higher education institutionÕs agency in the college -choice process may look different depending on the focus of the line of inquiry. Drawing upon the results of this study, however, institutional expenditures Ñand the subsequent products of those expenditures Ñare worthy of inclusion. Therefore , it may be necessary to c onceptualize institutional expenditure categories so that they can be investigated in qualitative ways. For example, researchers may ask participants directly how manifestations of academic support spending, such as various co -cur ricular services tied directly to academic programs like veterinary clinics or developmental research schools (K -12 schools), influenced their decision to attend an HBCU . Furthermore, the results of this study suggest that perceptions may be different for nonblack students attending public versus private HBCUs; therefore, it is essential that future researchers attempt to capture any variation that may exist both between and within these contexts. One way to address some of these suggestions in a more holis tic way would be to conduct in -depth case studies using HBCUs with some of the most dramatic changes in nonblack !!89 enrollment over the past decade. HBCUs that have experienced major shifts in nonblack enrollment may provide compelling insight into the issues presented in this study. Methodologically, using a case study method would allow researchers to collect additional data at the institutional level, further test hypotheses surrounding the connection between various institutional characteristics and nonbla ck enrollment, and interpret findings in their proper institutional context while still creating opportunities for the generalization of commonalities between institutions (Baxter & Jack, 2008). In addition to investigating the influence of the higher edu cation context on the individual studentÕs college -choice process, it is important that researchers develop studies that include the perspectives of those who control the higher education context: administrators. The results of this study suggest that non black students are influenced by institutional expenditures, but these results do not necessarily establish a link between varying levels of institutional expenditures and the conscious decisions of administrators to allocate resources in those areas for t he purpose of attracting nonblack students. As a result, future researchers should inquire into how HBCU administrators understand the relationship between institutional spending and diversity efforts. Do administrators make conscious decisions to allocate resources with racial diversity outcomes in mind? Does the desire to maintain a culture supportive of the institutionÕs unique mission to educate African -American students influence their decisions when making spending decisions? Do the perspectives of ad ministrators regarding nonblack enrollment and institutional expenditures vary according to their functional area? These are just a few overarching questions to guide future inquiry in this context. Policy and practice. With the exception of a select group of public HBCUs (e.g., in Mississippi or Tennessee), the majority of HBCUs have not experienced pressure from the state !!90 to explicitly address nonblack enrollment. What is true regardless of state policies regarding nonblack enrollment, however, is that th e pressure to maintain and increase enrollment in response to declining state support and the continued migration of black students to non -HBCU institutions make nonblack enrollment a critical piece of the puzzle to a sustainable future for HBCUs. At the s ame time, HBCU leaders must deal with the paradox of balancing increases in nonblack enrollment with the desire to maintain the traditional, and still very relevant, mission of educating African -American students. This paradox is the context within which H BCU leaders must make policy decisions about nonblack enrollment, and it is important to acknowledge this paradox as this section attempts to link the findings of the study to policy and practice at HBCUs. One implication emerges that is more broadly appl icable for the state and federal policy context, which is the fact that HBCUs are distinctly different from their PWI counterparts and, therefore, deserve to be given attention in the policy development process accordingly. Goldrick -Rab, Kelchen, and Houle (2014), in a study on federal financial aid, point to myriad differences between HBCUs and non -HBCUs (i.e. higher percentage of students receiving Pell grants, smaller endowments, etc.) that provide justification for federal policymakers more carefully considering HBCUs in broader policy efforts Ñwhich certainly aligns with previous research (Bastedo, 2012; Boland & Gasman, 2014 ; Minor, 2008; Ryan, 2005 ). This special consideration for HBCUs is particularly important in the context of the growing movement t o adopt performance -based funding formulas at the state level. Although this study spoke specifically to racial diversity, which is currently not included as an outcome measure in the performance -funding formula for most participating states, the finding t hat HBCUs are unique regarding the effect of institutional expenditures in college -choice decisions and other educational outcomes !!91 supports and reinforces the idea that state policies developed as one -size-fits -all may not only prove ineffective in produci ng certain outcomes at HBCUs, but may be detrimental to HBCUs (Friedel, Thornton, DÕAmico, & Katsinas, 2013). Certainly, one way to address this issue is to make genuine efforts to include and value the perspective of HBCU leaders and scholars in the polic y development process. Although there are some implications for state and federal policy, I believe the primary benefits of these findings related to policy and practice lie in the institutional context. Furthermore, because of the statistical and explanat ory strength of the model, the implications of the findings primarily apply to public HBCUs €. Specifically, I believe the findings of this study have the potential to inform public HBCU leaders in the strategic planning process, particularly members of boa rds of trustees and regents, and executive -level HBCU administrators. For HBCU leaders and policymakers who seek to address institutional needs and priorities, it is critically important that their decisions are data -driven and focused on investing in area s of the institution that will produce those desired outcomes. For example, in states with growing populations of Hispanic students (i.e., Texas or North Carolina) who are increasingly pursuing higher education, HBCUs may request additional funding or poli cymakers make focus additional allocations specifically on recruiting and retaining these nonblack students, using the findings of this study as data -driven justification for that decision. As mentioned previously, institutional context matters and these findings should be considered in the context of an institutionÕs mission and strategic plan. Institutional leaders should view the findings from this study as an impetus to examine their own institutionÕs !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!"!It is not to say that these findings cannot inform future research a bout private HBCU nonblack enrollment, but more so that the findings suggest that there are likely different factors that influence nonblack enrollment at private HBCUs not captured in this study. !!!92 expenditures and nonblack enrollment. HBCUs are not homogenous and do not have identical institutional missions and aspirations, especially as it relates to racial diversity (Lee, 2015). As a result, these findings may inform policy and practice in various ways. For example, a public HBCU that does want to increase nonblack enrollment may look deeper to determine its level of academic support spending and what the products of that spending may mean for attracting nonblack students Ñwhom the findings of this study suggest are influenced by academic support functions. Certainly, considering what is involved in academic support spending, any additional funding in support of these functions would benefit all students and would not compromise the integrity of the HBCU mission. Just like any organizational change i nitiative, should institutions decide that additional funding or a reallocation of existing resources in any expenditure category is the proper course of action, the decision should be guided by the stated mission and objectives of the organization (Barr, 2002). Aside from focusing efforts and relying solely upon institutional characteristics that do influence nonblack enrollment, another approach may involve focusing on better promoting the value of other institutional characteristics that appear to be le ss influential in the college -choice process for nonblack students. It is possible that nonblack students are not necessarily aware of the benefits of attending an HBCU beyond the academic support functions that seem to primarily influence their decision t o attend a public HBCU. Therefore, future marketing and recruiting campaigns may focus on other unique aspects of the institution that support student success and promote academic and personal development, such as the quality of student -faculty interaction s, community service opportunities, an institutionÕs historical commitment to social justice, or the value of diversity at an HBCU. If HBCUs are to compete with larger, more well -resourced !!93 historically white institutions for nonblack students, they must do so by possessing a more robust value proposition that includes more than just high -quality academic support functions. As important as it is to point out the agency HBCUs have in doing what is necessary to attract students and serve them well once they a re there, it is just as important to acknowledge the responsibility state and federal agencies have to support HBCUs as they attempt to carry out their educational missions. The reality is that many public HBCUs are still inequitably funded and must regula rly fight for their survival, despite their proven ability to remain the top producers of bachelorÕs degree for black students in the U.S. (Shorette, 2015). Politically, HBCUs must constantly justify their very existence and must often defend themselves fr om the attacks carried out by state politicians. For an example of these political attacks, one must look no further than the recent interactions between the state of South Carolina and its only public HBCU, South Carolina State University (SCSU). In 2015, with SCSU facing significant financial challenges attributed to declining enrollment and changes in funding formulas, the state legislature proposed to temporarily close the public HBCU for at least one year in hopes of getting the institution in better f inancial and organizational shape Ña move that many consider to be a certain nail in the coffin of SCSU. Rather than investing in the institution and providing it with the essential financial and human resources necessary to recover from decades of underfun ding at the hands of the state and increase enrollment, the stateÕs solution was to blame SCSU for its unwillingness to cut even more from its budget, deem it inoperable, and end its existence. There is little -to-no evidence to suggest that predominantly w hite public institutions face the same level of scrutiny as HBCUs when it comes to matters of institutional performance and some have been very explicit in pointing out that racism is a driving variable in that !!94 equation, suggesting that Ò white politic ians have viewed S.C. State as ÔtheirÕ school, and if ÔtheyÕ canÕt r un it effectively, then thatÕs ÔtheirÕ problem Ó (Noble, 2015). Although this is just one example, it is reflective of the blatant neglect that HBCUs have received in regard to public policy. I f states are to hold HBCUs to the same standards as other public institutions in the state and expect HBCUs to attract and serve diverse students from their state, they must take into consideration their achievements, unique contributions, and inequitable treatment from the state and invest accordingly. Without an equitable approach from the state, HBCUs will continue to be forced to do more with less and continue to exist in a Òseparate but unequalÓ political context, unable to fully pursue institutional a ctivities that would allow them to thrive. Furthermore, without equitable approaches from policymakers, HBCUs will continue to be forced to deal with issues in the context of interest convergence, whereas only progress will be made if politicians can be co nvinced that investments in HBCUs benefit everyone else (i.e., mainstream society, the state, white people, etc.). Conclusion This study quantitatively examined institutional characteristics Ñas exemplified primarily through institutional expenditures Ñthat influence nonblack enrollment at public and private four -year HBCUs between the years of 2000 -2010. In doing so, the research revealed that nonblack enrollment at public HBCUs is positively influenced by higher levels of academic support spending, while n onblack enrollment at private HBCUs is negatively influenced higher graduation rates. The findings suggest that nonblack students may be attracted to specific academic support functions at public HBCUs, such as demonstration school s associated with a colle ge of education or veterinary and dental clinics . Conversely, the findings suggest that three things may be happening regarding private HBCUs: !!95 1. private HBCUs with the most prestige are doing little to attract and recruit nonblack students, and/or 2. consideri ng that these institutions are often affectionately referred to as meccas for black education or as the ÒBlack Ivy League,Ó nonblack students may be receiving the message that these institutions are Òelite,Ó exclusively black, and not for them. As a schol ar who is critical of race and the role it plays in the U.S. higher education system, as well as someone who is very aware of the contemporary relevance of HBCUs and their mission to educate African -American students, I feel it is important to note that th is study and its findings are by no means a prescription for how to increase nonblack enrollment at HBCUs or a statement regarding the value of nonblack students when it comes to the quality of the educational experience at HBCUs. Rather, the results of th is study are intended to equip institutional actors and policymakers with additional data to inform their approach to managing enrollment as nonblack students become an increasingly important group of students when it comes to the sustainability of HBCUs. Additionally, the findings from this study are intended to equip researchers with a more diverse perspective of nonblack enrollment as they delve deeper into the phenomenon and attempt to develop a more robust understanding of the experiences of nonblack s tudents at HBCUs. !!96 APPENDIX !!97 APPENDIX Table 7. IPEDS Definitions for Variables Included in Study Variable Definition pct_nonblack (Outcome variable of interest) The percent of nonblack enrollment as deriv ed from the sum of total enrollment for Asian, Hispanic, American Indian, & white students divided by total enrollment for the institution. pct_grad_rate Percentage of full -time, first -time, degree/certificate -seeking undergraduate students graduating wit hin 150 percent of normal time. tuitionfee02_tf (in state) tuitionfee03_tf (private and out of state) Average cost of tuition per FTE, defined as the tuition charged by institutions to those full -time undergraduate students who do or do not meet the sta te's or institution's residency requirements. instaid_spend_fte Average amount of institutional grants (scholarships/fellowships) received by full -time, first -time degree/certificate -seeking undergraduate students. Institutional grants - Scholarships and fellowships granted and funded by the institution and/or individual departments within the institution, (i.e., instruction, research, public service) that may contribute indirectly to the enhancement of these programs. Includes scholarships targeted to cer tain individuals (e.g., based on state of residence, major field of study, athletic team participation) for which the institution designates the recipient. instruct_spend_fte Average spending on instruction per FTE* as derived from total spending in dol lars on instruction divided by total FTE count. Instruction is defined as a functional expense category that includes expenses of the colleges, schools, departments, and other instructional divisions of the institution and expenses for departmental researc h and public service that are not separately budgeted. Includes general academic instruction, occupational and vocational instruction, community education, preparatory and adult basic education, and regular, special, and extension sessions. Also includes e xpenses for both credit and non -credit activities. Excludes expenses for academic administration where the primary function is administration (e.g., academic deans). Information technology expenses related to instructional activities are included if the in stitution separately budgets and expenses information technology resources (otherwise these expenses are included in academic support). Operations and maintenance and interest amounts attributed to the instruction function have been subtracted from the to tal instructional expenditure amount at FASB reporting institutions. Operations and maintenance amounts (and interest in the 2009 aligned form) attributed to the instruction function have been subtracted from the total amount at public aligned form reporti ng institutions. acadsupp_spend_fte Average spending on academic support per FTE as derived from total spending on academic support divided by total FTE count. Academic support is defined as a functional expense category that includes expenses of activit ies and services that support the institution's primary missions of instruction, research, and public service. It includes the retention, preservation, and display of educational materials (for example, libraries, museums, and galleries); organized activit ies that provide support services to the academic functions of the institution (such as a demonstration school associated with a college of education or veterinary and dental clinics if their primary purpose is to support the instructional program); media such as audiovisual services; academic administration (including academic deans but not department chairpersons); and formally organized and separately budgeted academic personnel development and course and curriculum development expenses. Also included ar e information technology expenses related to academic support activities; if an institution does not separately budget and expense information technology resources, the costs associated with the three primary programs will be applied to this function and t he remainder to institutional support. Operations and maintenance and interest amounts attributed to the academic support function have been subtracted from the total academic support expenditure amount at FASB reporting institutions. Operations and maint enance amounts (and interest in the 2009 aligned form) attributed to the academic support function have been subtracted from the total academic support expenditure amount at public aligned form reporting institutions. !!98 Table 7 (contÕd) studserv_spend_fte Average amount of spending on student services per FTE as derived from the total spending in dollars on student services divided by total FTE count. Student services is defined as a functional expense category that includes expenses for admissions, regis trar activities, and activities whose primary purpose is to contribute to students emotional and physical well -being and to their intellectual, cultural, and social development outside the context of the formal instructional program. Examples include stude nt activities, cultural events, student newspapers, intramural athletics, student organizations, supplemental instruction outside the normal administration, and student records. Intercollegiate athletics and student health services may also be included exc ept when operated as self - supporting auxiliary enterprises. Also may include information technology expenses related to student service activities if the institution separately budgets and expenses information technology resources(otherwise these expense s are included in institutional support.) Operations and maintenance and interest amounts attributed to the student services function have been subtracted from the total student services expenditure amount at FASB reporting institutions. Operations and mai ntenance (and interest in the 2009 aligned form) amounts attributed to the student services function have been subtracted from the total student services expenditure amount at public aligned form reporting institutions. admin_spend_fte Average spending o n administration per FTE as derived from the total spending in dollars on institutional support divided by total FTE count. Institutional support is defined as a functional expense category that includes expenses for the day -to-day operational support of t he institution. Includes expenses for general administrative services, central executive -level activities concerned with management and long -range planning, legal and fiscal operations, space management, employee personnel and records, logistical services such as purchasing and printing, and public relations and development. Also includes information technology expenses related to institutional support activities. Operations and maintenance and interest amounts attributed to the institutional support functi on have been subtracted from the total institutional support expenditure amount at FASB reporting institutions. Operations and maintenance amounts (and interest in the 2009 aligned form) attributed to the institutional support function have been subtracted from the total institutional support expenditure amount at public aligned form reporting institutions. aux_spend_fte Auxiliary enterprises - total expenses is the sum of all operating expenses associated with essentially self -supporting operations of th e institution that exist to furnish a service to students, faculty, or staff, and that charge a fee that is directly related to, although not necessarily equal to, the cost of the service. Examples are residence halls, food services, student health service s, intercollegiate athletics (only if essentially self -supporting), college unions, college stores, faculty and staff parking, and faculty housing. The amount of interest attributed to the auxiliary enterprise function has been subtracted from the total a uxiliary enterprise expenditure amount at institutions reporting on the FASB or 2010 aligned form. * FTE, or Full -time equivalent enrollments, are derived from the enrollment by race/ethnicity section of the fall enrollment survey. The ful l-time equivale nt of an institution's part -time enrollment is estimated by multiplying part -time enrollment by factors that vary by control and level of institution and level of student; the estimated full -time equivalent of part -time enrollment is then added to the full -time enrollment of the institution. This formula is used by the U.S. Department of Education to produce the full -time equivalent enrollment data published annually in the Digest of Education Statistics. !!99 BIBLIOGRAPHY !!100 BIBLIOGRAPHY Abelman, R. & Dalessandro, A. (2007). The institutional vision of historically Black colleges and universities. Journal of Black Studies, 20 (10), 1-30. Allen, W.R. (1992). The color of success: African American stude nt outcomes at predominantly White and historically Black college and universities. Harvard Education Review, 62 , 26-44. Allen, W. & Jewell, J. (2002). A backward glance forward: Past, present, and future perspectives on Historically Black Colleges and Universities. The Review of Higher Education, 25 (3), 241-261. Allen, W.R., Jewell, J.O., Griffin, K.A., & Wolf, D.S. (2007). Historically Black colleges and universities: Honoring the past, engaging the present, touching the future. Journal of Negro Education, 76 (3), 263-280. Allison, P. D. (1994). Using panel data to estimate the effects of events. Sociological Methods and Research, 23 , 174-199. Altbach, P. (1999). Private higher education: Themes and variations in comparative perspective. Prospects, 29 (3), 310-323. Angrist, J. & Pischke, J. (2009). Mostly harmless econometrics: An empiricistÕs companion. Princeton: Princeton University Press. Asian American and Pacific Islander Higher Education Resource Center (AAPIHERC). (2013). Wha t is an AANAPISI? Retrieved from http://aapiherc.southseattle.edu/en/about -aanapisis/what -is-an-aanapisi Barr, M. (2002). Academic administratorÕs guide to budgets and financial management. San Francisco: Jossey -Bass. Bastedo, M. (2012). The organiz ation of higher education: Managing colleges for a new era. Baltimore: The Johns Hopkins University Press. Baxter, P., & Jack, S. (2008). Qualitative case study methodology: Study d esign and implementation for novice r esearchers . The Qualitative Repo rt, 13(4), 544-559. Retrieved from http://nsuworks.nova.edu/tqr/vol13/iss4/2 Beach, A., Dawkins, P., Rozman, S., & Grant, J. (2008). Faculty development at Historically Black Colleges and Universities. In Gasman, M., Baez, B., & Turner, C.S.V. (Eds), Understanding minority -serving institutions (pp. 156-168). New York: State University of New York Press. !!101 Bennett, P., and Xie, Y. (2003). Revisiting racial differences in college attendance: The role of historically Black colleges and universities. Amer ican Sociological Review 68(4): 567-580. Bergerson, A. (2009). College choice and access to college: Moving policies, research, and practice to the 21 st century. ASHE higher education report, 35(4). Boland, W. & Gasman, M. (2014). AmericaÕs public HBCUs: A four state comparison of institutional capacity and state funding priorities. A report produced by the University of Pennsylvania Center for Minority -Serving Institutions, Philadelphia, PA. Bridges, B., Kinzie, J., Laird, T., & Kuh, G. (2 008). Student engagement and student success at historically Black and Hispanic -serving institutions. In Gasman, M., Baez, B., & Turner, C.S.V. (Eds), Understanding minority -serving institutions (pp. 217-236). New York: State University of New York P ress. Bridges, B., B. Cambridge, G.D. Kuh, and L. H. Leegwater (2005). Student engagement at minority -serving institutions: Emerging lessons from the BEAMS project. New Directions for Institutional Research 125, 25 -43. Brown, C. I., & Stein, P. R. (1 972). The White student in five predominantly Black universities. Negro Educational Review, XXXIII(4), 148-170. Brown, M.C. (2002). Good intentions: Collegiate desegregation and transdemographic enrollments. The Review of Higher Education, 25 (3), 263-280. Carnevale, A. & Strohl, J. (2013). Separate & Unequal: How higher education reinforces the intergenerational reproduction of white racial privilege. A report produced by the Georgetown University Center on Education and the Workforce, Washington, D.C. Clarke, M. (2007). The impact of higher education rankings on student access, choice, and opportunity. In Institute for Higher Education Policy (Ed.) College and university ranking systems: Global perspectives and American challenges (pp. 35-47). Washington, D.C. Connerly, W. (2003). At issue: Are racially identifiable colleges and universities good for the country? CQ RESEARCHER, 13 , 1061. Conrad, C. F., Brier, E. M., & Braxton, J. M. (1997). Factors contributing to the matriculation of White students in public HBCUs. Journal of a Just & Caring Education, 3(1), 37-63. Cuyjet, M. (2006). African American college men: Twenty -first century issues and concerns. In Cuyjet, M. (Ed.), African American Men in College (pp. 3-23). San Francisco: Joss ey- Bass. !!102 Davis, C.H.F. (2012). The story of Trayvon Marvin and resurgent social justice among HBCU college -goers. HBCU Digest . Retrieved from http://hbcudigest.com/the -story -of- trayvon -martin -and -resurgent -social -justice -among -hbcu -college -goers/ Delta Cost Project. (2014 ). Public release IPEDS data for years 1987 -2010. Retrieved from http://nces.ed.gov/ipeds/deltacostproject/ Ehrenberg, R. G., & Rothstein, D. S. (1993). Do Historically Black Institutions of Higher Education Confer Unique Advantag es on Black Students: An Initial Analysis. National Bureau of Economic Research, Cambridge, MA, Unpublished paper. Epple, D., Romano, R., & Sieg, H. (2006). Admission, tuition, and financial aid policies in the market for higher education. Econometrica, 74(4), 885-928. Exkano, J. (2012). Toward an African Cosmology: Reframing How We Think About Historically Black Colleges and Universities . Journal of Black Studies, 44 (1), 63-80. Fleming, J. (1984). Blacks in college: A comparative study of studentsÕ success in Black and in White institutions. San Francisco: Jossey -Bass. Frank, K. (2000). Impact of a confounding variable on a regression coefficient. Sociological Methods and Research, 29 (2), 147-194. Frank, K. (2014). KonFound -it!. Retrieved from https://www.msu.edu/~kenfrank/research.htm#causal Frank, K., Maroulis, S., Duong, M., & Kelcey, B. (2013). What would it take to Change an Inference?: Using RubinÕs Causal Model to Interpret the Robustness of Causal Inferences. Educational Evaluation and Policy Analysis , 35, 437-460. Freeman, K. (2005). African Americans and college choice: The influence of family and school. New York: State University of New York Press. Friedel, J., Thornton, Z., DÕAmico, M., & Katsinas, S. (2013). Performance -based funding: The national landscape. A report produced by the Education Policy Center at the University of Alabama, Tuscaloosa, Alabama. Gasman, M. (2008a). Minority -serving institutions: The pathway to successfully education students of color. A re port produced by Lumina Foundation for Education, Indianapolis, IN. Gasman, M. (2008b). Minority -serving institutions: A historical backdrop. In M. Gasman, B. Baez, & C. S. V. Turner (Eds.), Understanding minority -serving institutions (pp. 18- 27). Albany: State University of New York Press. !!103 Gasman, M. (2012). Historically black colleges and universities must embrace diversity. The Chronicle of Higher Education . Retrieved from http://chronicle.com/blogs/conversation/2012/11/08/historically -black -colleges -and -universities -must -embrace -diversity/ Gasman, M. (2013). The changing face of historically black colleges and universities. A report produced by the University of Pennsylvania Center for Minority -Serving Institutions, Philadelphia, PA. Gasman, M. & McMickens, T.L. (2010). Liberal or professional education? The missions of public Black colleges and universities and their impact on the future of African Americans. SOULS: A Critical Journal of Black Politics, Culture, and Society, 12(3), 286-305. Gasman, M. & Shorette, C.R. (2012). HBCUs, places for all to learn. Diverse Issues in Higher Education . Retrieved from http://diverseeducation.com/article/49854/ Goldrick -Rab, S., Kelchen, R., & Houle, J. (2014). The color of student debt: Impl ications of federal loan program reforms for black students and historically black colleges and universities. A report produced by the Wisconsin HOPE Lab, Madison, Wisconsin. Griffin, K. & Hurtado, S. (2011). Institutional variety in American higher edu cation. In J. Schuh, S. Jones, & S. Harper (Eds.), Student Services: A Handbook for the Profession (pp. 24 - 42). John Wiley & Sons. Gurin, P., Dey, E., Hurtado, S., & Gurin, G. (2002). Diversity and higher education: Theory and impact on educational o utcomes . Harvard Educational Review, 72 (3), 330-366. Gurin, P., Nagda, B., & Lopez, G. (2004). The benefits of diversity in education for democratic citizenship. Journal of Social Issues, 60 (1), 17-34. Hardy, M. & Reynolds, J. (2009). Incorporating cat egorical information into regression models: The utility of dummy variables. In M. Hardy & A. Bryman (Eds.), Handbook of Data Analysis (pp. 209-236). Thousand Oaks: SAGE. Harris, M. (2013). Defining institutional diversity. ASHE higher education repo rt series (AEHE): Understanding institutional diversity in American higher education, 39 (3), 1-15. Hatter, L. (2014). Proposed split of FAMU -FSU engineering shows old wounds run d eep. WFSU . Retrieved from http://news.wfsu.org/post/proposed -split -famu -fsu-engineering - shows -old -wounds -run-deep Hernandez, A. (2010). Independent thinking: HBCUs explore ways to liberate themselves from tuition dependence. Diverse Issues in Higher Education, 27 (5), 12. !!104 Hinkle, D., Wiersma, W., & Jurs, S. (2003). Applied st atistics for the behavioral sciences (Fifth edition). Belmont: Wadsworth. Hispanic Association of Colleges & Universities (HACU). (201 5). Fact sheet: Hispanic higher education and HSIs. Retrieved from http://www.hacu.net/hacu/HSI_Fact_Sheet.asp Hoss ler, D., & Gallagher, K.S. (1987). Studying college choice: A three -phase model and the implications for policy -makers. College and University, 2, 207Ð221. Hoxby, C. (1997). How the changing market structure of U.S. higher education explains college tuition. NBER Working Paper 6323. Hoxby, C. (2009). The changing selectivity of American colleges. Journal of Economic Perspectives, 23 (4), 95-118. Institute for Higher Education Policy (IHEP). (2004). Serving the Nation. A report produced by the Ins titute for Higher Education Policy, Washington, DC. Institute for Higher Education Policy (IHEP). (2007). College and university ranking systems: Global perspectives and American challenges. A report produced by the Institute for Higher Education Policy, Washington, D.C. Jacob, B., McCall, B., & Stange. K. (2013). College as country club: Do colleges cater to students' preferences for c onsumption? NBER Wor king Paper 18745. Jewell, J. (2002). To set an example : The tradition of diversity at Historica lly Black Colleges and Universities. Urban Education, 37 (1), 7-21. Johnson, D. (1995). Alternative methods for the quantitative analysis of panel data in family research: Pooled time -series models. Journal of Marriage and Family, 57 (4), 1065- 1077. Kim, M.M. & Conrad, C. (2006). The impact of Historically Black Colleges and Universities on the academic success of African -American students. Research in Higher Education, 47 (4), 399-427. Lee, J.M. (2015). Moving beyond racial and ethnic diversity at HBCUs. In R. Palmer, C.R. Shorette, and M. Gasman (Eds.), Exploring Diversity at Historically Black Colleges and Universities: Implications for Policy and Practice (pp. 17-35). San Francisco: Jossey - Bass. Lee, J.M. & Keys, S. (2013). Land -grant but un equal: State one -to-one match funding for 1890 land -grant universities. A report produced by the Association for Public and Land -Grant Universities, Washington, D.C. Linsenmeier, D., Rosen, H., & Rouse, C.E. (2006). Financial aid packages and college !!105 enrollment decisions: An econometric case study. The Review of Economics and Statistics, 88(1), 126Ð145. Martin, M. (2013). Do we still need HBCUs? National Public Radio. Retrieved from http://www.npr.org/2013/01/22/169980243/do -we-still-need -hbcus McDonough, P. M. (1997). Choosing colleges: How social class and schools structure opportunity . SUNY Press. McDonough, P., Antonio, A., & Trent, J. (1997). Black students, black colleges: An African American college choice model. Journal for a Just a nd Caring Education, (3) 1, 9-36. McPherson, M. S. & Schapiro, M.O. (1991). Does student aid affect college enrollment? New evidence on a persistent controversy. American Economic Review, 81(1), 309Ð318. Mercer, C.J. & Stedman, J. (2008). Minority -servi ng institutions: Selected institutional and student characteristics. In M. Gasman, B. Baez, & Turner, C. (Eds.) Understanding Minority -Serving Institutions , (pp. 28-42). New York: State University of New York Press. Mickelson, R. A. (1990). The atti tude -achievement paradox among Black adolescents. Sociology of Education, 63 , 44-61. Minor, J.T. (2008a). Contemporary HBCUs: Considering institutional capacity and state priorities. A research report. Michigan State University, College of Education, Department of Educational Administration. East Lansing, MI. Minor, J.T. (2005). Faculty governance at historically Black colleges and universities . Academe, 91 (3), 34-37. Minor, J.T. (2008b). Groundwork for studying governance at Historically Black Colleges and Universities. In Gasman, M., Baez, B., & Turner, C.S.V. (Eds), Understanding minority - serving institutions (pp. 169-182). New York: State University of New York Press. Morgan, S. & Winship, C. (2007). Counterfactuals and causal inference: Methods and principles for social research. New York: Cambridge University Press. Murnane, R. & Willett, J. (2011). Methods matter: Improving causal inference in educational and social science research. New York: Oxford University Press. National C enter for Education Statistics (NCES). (2014). Fast facts: Historically black colleges and universities. Retrieved from http://nces.ed.gov/fastfacts/display.asp?id=667 National Science Board. (2012). Science Board concerned about declines in public resea rch university funding. Retrieved from http://www.nsf.gov/news/news_summ.jsp?cntn_id=125542 !!106 Nelms, C. (2011). Strengthening AmericaÕs Historically Black Colleges and Universities: A call to action. Report: NCCU Office of the Chancelor. Nixon, H. L ., & Henry, W. J. (1990). Factors associated with enrollment decisions of Black students and White students in colleges at which they are in the minority: Thoughts for administrators. National Forum of Educational Administration and Supervision, 7(3), 43-47. Noble, P. (2015). Race a factor in South Carolina State University Ômess.Õ Diverse: Issues in Higher Education . Retrieved from http://diverseeducation.com/article/69851/ Nora, A. (2004). The role of habitus and cultural capital in choosing a col lege, transitioning from high school to higher education, and persisting in college among minority and nonminority students. Journal of Hispanic Higher Education, 3 (2), 180-208. Oguntoyinbo, L. (2012). Historically Black law schools stay the course on social justice mission. Diverse Issues in Higher Education , Retrieved from http://diverseeducation.com/article/16984c4/historically -black -law-schools -stay- the -course -on-social -justice -mission.html Ozuna, T. (2012). Examining the first -year experience s and perceptions of sense of belonging among Mexican American students enrolled in a Texas HBCU. Doctoral dissertation, University of Texas at Austin. Available electronically from http://hdl.handle.net/2152/ETD -UT-2012-08-6002. Palmer, R. & Maramb a, D. (2015). Racial microaggressions among Asian American and Latino/a student at an historically black college and university. Journal of College Student Development . Palmer, R., Maramba, D., Allen, T., & Goings, R. (2015). From Matriculation to Engag ement on Campus: Delineating the Experiences of Latino/a Students at a Public Historically Black University. In R. Palmer, C.R. Shorette, & M. Gasman (Eds.), Exploring diversity at historically black colleges and universities: Implications for policy and practice , New Directions for Higher Education (pp. 67-78). San Francisco: Jossey -Bass. Pan, W. & Frank, K. (2003). A probability index of the robustness of a causal inference. Journal of Educational and Behavioral Statistics, 28 (4), 315-337. Parker, T . (2012). The role of Minority -serving institutions in redefining and improving developmental education. A report published by the Southern Education Foundation, 135 Auburn Avenue NE; Second Floor, Atlanta, GA 30303. Paulsen, M. (1990). College choice : Understanding student behavior. ASHE -ERIC Higher Education Report No. 6 (ISSN -0884-0040). Washington, D.C. !!107 Perna, L. (2006). Studying college access and choice: A proposed conceptual model. In J. C. Smith (Ed.) Higher Education: Handbook of Theory and Research , Volume 21 (pp. 99 - 157). Netherlands: Springer. Perry, J. and Rainey, H. (1988). The public -private distinction in organization theory: A critique and research s trategy . The Academy of Management Review , 13(2), 182-201. Peterson, R. D., and Hamrick, F. A. (2009). White, male, and Òminority:Ó Racial consciousness among white male undergraduates attending a Historically Black University. The Journal of Higher Education , 80(1), 34-58. Pitre, P. (2006). College choice: A study of African American and white student aspirations and perceptions related to college attendance. College Student Journal, 40 (3), 562- 574. Price, I., Matzdorf, F., Smith, L., & Agahi, H. (2003). The impact of f acilities on student choice of university. Facilities, 21 (10), 212-222. Rankin, S. & Reason, R. (2005). Differing perceptions: How students of color and white students perceive campus climate for underrepresented groups. Journal of College Student D evelopment, 46 (1), 43-61. Redd, K. (1998). Historically Black Colleges and Universities: Making a comeback. New Directions for Higher Education, 102 , 33-43. Ricard, R.B. & Brown, M.C. (2008). Ebony towers in higher education: The evolution, mission, and presidency of historically black colleges and universities. Sterling: Stylus. Roach, R. (2013). Latino college enrollment rate surpasses that of whites. Diverse Issues in Higher Education . Retrieved from http://diverseeducation.com/article/53313/ Roderick, M., Coca, V., & Nagoaka, J. (2011). Potholes on the road to college: High school effects in shaping urban studentsÕ participation in college application, four -year college enrollment, and college match. Sociology of Education, 84 (3). Samuel s, A. (2004). Is separate unequal? Black colleges and the challenge to desegregation. Lawrence: University Press of Kansas. Santiago, D. (2010). Ensuring AmericaÕs future: Benchmarking Latino college completion to meet national goals 2010 to 2020. A r eport produced by Excelencia in Education, Washington, D.C. Santiago, D., & Reindl, T. (2009). Taking stock: Higher education and Latinos. A report produced by Excelencia in Education, Washington, D.C. !!108 Sayrs, L.W. (1989). Pooled time series analysis . Newbury Park: Sage. Seymore, S. (2006). IÕm confused: How can the federal government promote diversity in higher education yet continue to strengthen historically black colleges? Washington and Lee Journal of Civil Rights and Social Justice, 12(2), 287-319. Shorette, C.R. (2015). Black colleges matter. Inside Higher Ed . Retrieved from https://www.insidehighered.com/views/2015/01/15/essay -why -historically -black - colleges -matter-because -they -serve -black -students Shorette, C.R. & Arroyo, A.T. (2015). A closer examination of white student enrollment at HBCUs. In R. Palmer, C.R. Shorette, & M. Gasman (Eds.), Exploring diversity at historically black colleges and universities: Implications for policy and practice , New Directions for Higher Educatio n (pp. 49-65). San Francisco: Jossey -Bass. Sims, S. (1994). Diversifying Historically Black Colleges and Universities: A new higher education paradigm. Westport: Greenwood Publishing Group. Smith, D.G., Moreno, J., Clayton -Pedersen, A.R., Parker, S., a nd Teraguchi, D.H. (2005). ÒUnknownÓ students on college campuses: An explanatory analysis. A report produced by the James Irvine Foundation Campus Diversity Initiative Evaluation Project, San Francisco, CA. Sorrell, M. (2012). You can be my kind an d not be my color. The Dallas Morning News . Retrieved from http://www.dallasnews.com/opinion/sunday -commentary/20120824 -michael -sorrell -you-can-be-my-kind -and -not-be-my-color.ece St. John, E. P., (1990). Price response in enrollment decisions: An analysi s of the high school and beyond sophomore cohort. Research in Higher Education, 31(2), 161Ð176. Stewart , P. (2012). Three Mississippi HBCUs finding diversity fuels their mission. Diverse Issues in Higher Education . Retrieved from http://diverseeduc ation.com/article/48872/ Strayhorn, T. (2010). Majority as temporary minority: Examining the influence of faculty - student relationships on satisfaction among white undergraduates at historically black colleges and universities. Journal of College Stud ent Development, 51 (5), 509- 524. Strayhorn, T. & Hirt, J. (2008). Social justice at historically Black and Hispanic -serving institutions: Mission statements and administrative voices. In M. Gasman, B. Baez, & C.S.V. Turner (Eds), Understanding minor ity-serving institutions (pp. 203-216). New York: State University of New York Press. Taylor, B. & Cantwell, B. (2014). Global competition, US research universities, and international doctoral education: Growth and consolidation of an organizational f ield. !!109 Research in Higher Education, 54 , 411-441. Titus, M. (2009). The production of bachelorÕs degrees and financial aspects of state higher education policy: A dynamic a nalysis . The Journal of Higher Education, 80 (4), 439-468. Tolbert, P. (1985). In stitutional environments and resource dependence: Sources of administrative structure in institutions of higher e ducation . Administrative Science Quarterly, 30 (1), 1-13. Toma, J. (2012). Institutional strategy: Positioning for prestige. In M. Bastedo (Ed.) The organization of higher education: Managing colleges for a new era (pp. 118-159). Baltimore: The Johns Hopkins University Press. Toutkoushian, R. & Smart, J. (2001). Do institutional characteristics affect student gains from college? Review of Higher Education, 25(1), 39 -61. Vedder, R. (2010). Why do we have HBCUs? The Chronicle of Higher Education . Retrieved from http://chronicle.com/blogs/innovations/why -do-we-have -hbcus/27506 U.S. Census Bureau. (2012). U.S. Census Bureau projections s how a slower growing, older, more diverse nation a half century from now. Retrieved from http://www.census.gov/newsroom/releases/archives/population/cb12 -243.html Van Der Klaauw, W. (2002). Estimating the effects of financial aid offers on college enrollment: A regression -discontinuity approach. International Economic Review , 43(4), 1249-1287. Wells, C. (2013). Maryland universities unnecessarily duplicated the programs of black colleges, court rules . The Baltimore Sun . Retrieved from http://article s.baltimoresun.com/2013 -10- 08/news/bs -md-black -colleges -rulling -20131007_1_black -institutions -black -colleges - hbcus Wells, R. & Stage, F. (2015). Past, Present, and Future of Critical Quantitative Research in Higher Education . In R.Wells & F. Stage (Ed s.) New Scholarship in Critical Quantitative Research ÑPart 2: New Populations, Approaches, and Challenges , New Directions for Institutional Research (pp. 103-112). San Francisco: Jossey -Bass. White House Initiative on HBCUs Annual Report (2005). Educatio n Publications Center, U.S. Department of Education, P.O. Box 1398, Jessup, MD 20794 -1398 White House Initiative on Historically Black Colleges and Universities. (201 5). What is an HBCU? Where are the HBCUs? Retrieved from http://www.ed.gov/edblogs/whhb cu/one -hundred -and -five -historically -black -colleges -and -universities/ Zhang, L. (2010). The use of panel data models in higher education policy studies. Higher Education: Handbook of Theory and Research, 25, 307-349.