Home Articles Reader Opinion Editorial Book Reviews Discussion Writers Guide About TCRecord
transparent 13
Topics
Discussion
Announcements
 

The Impact of Community College Baccalaureate Adoption on Associate Degree Production


by Justin C. Ortagus, Dennis A. Kramer, Manuel S. González Canché & Frank Fernandez - 2020

Background/Context: As of 2018, a total of 19 states allow at least one community college to offer baccalaureate degrees. Previous researchers have suggested that community college baccalaureate (CCB) adoption will lead to a host of unintended consequences, including decreases in associate degree production.

Purpose/Objective/Research Question/Focus of Study: This study empirically examines the impact of CCB adoption on associate degree production and adds to conversations surrounding the consequences of CCB adoption.

Research Design: We use a quantitative quasi-experimental research design to examine the effect of CCB adoption on associate degree production.

Findings/Results: When comparing adopting and non-adopting community colleges within the state of Florida, the authors find that the adoption of CCB degree programs has a positive impact on overall associate degree production, but this impact varies considerably according to the type of academic degree program.

Conclusions/Recommendations: Opponents of CCB legislation have argued that giving community colleges the authority to confer baccalaureate degrees will detract from the sub-baccalaureate institutional mission of community colleges, but our results suggest that the adoption of a CCB degree program is associated with an overall increase in associate degree production. Findings from this work should be an important consideration for policymakers seeking to increase baccalaureate degree production in addition to—not at the expense of—associate degree programs.



INTRODUCTION


Many policymakers and policy-minded groups have outlined the need to increase the percentage of Americans who earn postsecondary credentials to support the nation’s global competitiveness in the 21st century knowledge economy. For example, the Lumina Foundation advocates for a variety of policies and practices designed to increase the proportion of Americans who hold a postsecondary credential to 60% by 2025 (Lumina Foundation, 2017). As broad-access institutions serving a high number of historically underrepresented students, community colleges will be instrumental to achieving the national completion agenda (Stevens & Kirst, 2015). Community colleges represent a large and important segment of American higher education and award more than 800,000 associate degrees per academic year (American Association of Community Colleges [AACC], 2018).


Previous researchers noted the considerable role and influence of the community college sector in improving the economic and social mobility of its graduates. Specifically, some economists have demonstrated that students who earn associate degrees have better, higher-paying jobs than students with high school diplomas (Belfield & Bailey, 2017; Minaya & Scott-Clayton, 2017). On average, students who earn associate degrees are paid between $4,640 and $7,160 more per year when compared to individuals who start but do not complete college (Belfield & Bailey, 2017). Community college alumni also tend to have stable employment and experience larger growth in earnings over time relative to students who earned a high school degree or short-term vocational certificates (Minaya & Scott-Clayton, 2017). In addition, one can argue that many of these community college students with improved labor market outcomes are traditionally disadvantaged, as over half of all Hispanic students and more than 40% of Black students enrolled in higher education attend community colleges (AACC, 2018).


For most of their history, public community colleges were restrained by state laws from offering credentials higher than the associate degree. Unlike early teacher’s colleges or state normal schools, community colleges were not allowed to emulate state flagship universities by expanding their institutional missions to offer baccalaureate degree programs (Labaree, 2010, 2017). The first community college baccalaureate (CCB) program was established in 1989 (Fulton, 2015). As of 2018, a total of 19 states have changed their policies to allow at least one community college to offer baccalaureate degrees (Povich, 2018). Notably, the number of bachelor’s degrees conferred at community colleges increased from 1,690 in 2000 to 17,035 in 2014 (authors’ calculations using IPEDS data).


Recent scholarship has examined which states adopted CCB degree programs (Henderson, 2014), how CCB degree programs were implemented (Essink, 2013), and whether they increased overall degree production in nursing fields (Daun-Barnett, 2011). A host of other scholars have questioned whether CCB degree programs may detract from the traditional institutional mission of community colleges (e.g., Floyd, Falconetti, & Hrabak 2009; Floyd & Walker, 2009; Russell, 2010; Walker, 2005). However, extant literature has yet to empirically examine the unintended consequences of adopting CCB degree programs, such as detracting from degree production in academic areas that do not offer baccalaureate degree programs.


This study adds to conversations surrounding the consequences of CCB adoption and should be an important consideration for policymakers seeking to increase baccalaureate degree production in addition to—not at the expense of—associate degree programs, particularly given that sub-baccalaureate degree programs have been identified as an important element of national goals for increasing the college completion rate and maintaining national competitiveness in the 21st century knowledge economy (e.g., Lumina Foundation, 2017). To examine this issue, we address the following overarching research question:


1.

Are there changes in associate degree production after adopting CCB degree programs?


Given that the effect of CCB degree programs may be heterogeneous across certain academic programs, and potentially subgroups of students, we also pose the following question:


2.

Do the changes in associate degree production following CCB adoption disproportionately affect certain types of students or academic program areas?


Finally, given evidence of outcome differentials related to time of adoption, we also ask the following question:


3.

Do early adopters of CCBs (between 2001 and 2007) produce fewer associate degrees than their late-adopting peer institutions (between 2008 and 2014)?


This paper presents a quantitative quasi-experimental case study of the unintended consequences of the adoption of CCB degree programs in Florida. Numerous states, including Florida, have allowed community colleges to offer four-year degrees in order to address local labor market needs and increase degree production in high-demand fields, such as nursing and teacher education (Florida Department of Education [FLDOE], 2005). Among the states that have allowed community colleges to award bachelor’s degrees, Florida represents a leader in the CCB movement due to the large number of community colleges in the state that award bachelor’s degrees (Fulton, 2015). As of 2017, 26 of Florida’s 28 community colleges offer at least one baccalaureate degree program, with a total of 187 CCB degree programs across the state (Florida College System, 2017).


We use institutional theory to inform hypotheses and interpret findings regarding the impact of CCB degree programs in Florida. As detailed in the theoretical framework section below, institutional theory suggests that colleges and universities adhere to isomorphic pressures that reward mission drift (e.g., creating new baccalaureate degree programs) over efficiency (e.g., maintaining and improving associate degree production across fields) (see DiMaggio & Powell, 1983; Meyer & Rowan, 1977; Morphew & Huisman, 2002). As community colleges pursue status by developing new CCB degree programs, they may devote less attention and fewer resources to their existing sub-baccalaureate degree programs. We estimate several regression models using various difference-in-differences approaches to determine whether community colleges awarded fewer associate degrees across fields after establishing CCB degree programs. When comparing adopting and non-adopting community colleges within Florida, we find evidence that the adoption of CCB degree programs has a positive impact on overall associate degree production, but this impact varies considerably according to the type of associate degree program. This study concludes by providing the implications of our findings for practitioners, policymakers, and researchers.


THE POLICY CONTEXT OF CCB DEGREE PROGRAMS


Previous research has shown that the majority of the U.S. population lives within 25 miles of a community college (Rephann, 2007). Based on geographic accessibility alone, many states can improve access to baccalaureate degree programs by expanding offerings at community colleges rather than building new brick-and-mortar four-year colleges and universities in remote areas. In 1989, the West Virginia legislature was the first to authorize the adoption of a CCB degree program. The CCB movement continued to grow over the years, with 12 additional states authorizing community colleges to issue bachelor’s degrees by 2004. As of 2018, 19 state legislatures or governing boards authorize at least one community colleges to offer baccalaureate degrees (Povich, 2018). The growth of CCB degree programs has been attributed largely to the need to improve responsiveness to local and state workforce demands (Russell, 2010; Walker, 2005), but additional reasons for adopting CCB degree programs have been outlined in previous work. Specifically, advocates of CCB degree programs have suggested that offering bachelor’s degrees at community colleges can reduce costs for both students and taxpayers (Bemmel, 2008) and expand access to bachelor’s degrees for traditionally underserved students, such as rural students who may not live near a four-year college or university (Bemmel, Floyd, & Bryan, 2008).


However, the notion of offering bachelor’s degrees at community colleges does not come without its detractors. Opponents of CCB policies have argued that CCB degree programs would duplicate the efforts of four-year colleges and universities and represent a form of mission drift that could move community colleges away from their original mission of open-access education and the provision of sub-baccalaureate credentials (Floyd & Walker, 2009; Russell, 2010; Walker, 2005). Mission drift was previously defined as “a well-known phenomenon in American higher education in which one segment of higher education redefines its mission to include the responsibility already being performed by another” (Kerr, 2001, p. 3). Levin (2004) discussed shifting community college missions as a matter of institutional identity and suggested that any community college offering bachelor’s degrees “possesses an identity that is obviously no longer simply a sub-baccalaureate institution, and possibly no longer a post-secondary institution that serves, as one of its main functions, marginalized and underserved groups as its primary client or customer” (p. 16).


The CCB movement in the state of Florida began in 2001 when state legislators adopted Senate Bill 1162, which officially granted St. Petersburg Junior College the ability to drop “junior” from its name and confer baccalaureate degrees (Florida Senate, 2008). Florida lawmakers approved CCB degree programs to help meet the “critical statewide need for trained teachers, nurses, and information technology employees” (FLDOE, 2005, p. 1). In 2008, Florida expanded the legislative authority of its community colleges to offer CCBs in additional areas of workforce development, focusing specifically on counties with labor market shortages in the fields of nursing and education (FLDOE, 2008). At the time of CCB adoption, Florida ranked in the top quintile in associate degree production due to the strength of its public community colleges (Furlong, 2005), but extant literature has yet to explore the unintended consequences of adopting CCB degree programs, such as whether CCB degree programs come at the expense of associate degree production.


LITERATURE REVIEW


Although there is a limited body of literature pertaining to the effects of CCB degree programs, several scholars have explored the motivational forces and policy antecedents related to CCB degree program adoption (Gonzales, 2005; Hrabak, 2009; McKee; 2001; McKinney, Scicchitano, & Johns, 2013). For example, Petry (2006) administered a survey to 38 administrators from the first five community colleges in Florida to adopt CCB degree programs, finding that colleges offered CCB degree programs to increase access to bachelor’s degrees and improve workforce development in their surrounding communities. Henderson (2014) employed event history analysis to examine which states were most likely to adopt CCB degree programs and found evidence of CCB policy diffusion, as states were more likely to adopt CCB provisions if they bordered a state that previously allowed community colleges to award baccalaureate degrees. Additional research has questioned how community colleges in Florida will continue to balance the addition of CCB degree programs with the traditional community college mission (Floyd, Falconetti, & Hrabak, 2009).


Fewer empirical studies have examined what happens following the implementation of CCB degree programs. Previous work suggested that CCB adoption was positively related to the overall production of nursing degrees (Daun-Barnett, 2011). Park, Tandberg, Shim, Hu, and Herrington (2016) examined CCB adoption and its impact on the number and diversity of students who earn bachelor’s degrees in teacher education. The authors found no effect of CCB adoption on degree production in teacher education, with increases being concentrated solely in states with widespread use of CCB degree programs in teacher education. Surprisingly, the diversity of graduates in teacher education programs declined following CCB adoption. Daun-Barnett and Escalante (2014) also focused specifically on the influence of CCB adoption on degree production in nursing and teaching, finding that CCB degree programs did not result in increased production in nurses or teachers.


Prior work has shown that CCB-adopting institutions aim to establish more CCB degree programs in science, technology, engineering, and mathematics (STEM) fields as a way to “enable community colleges to play a larger role in the production of STEM graduates in the future” (McKinney, Scicchitano, & Johns, 2013, p. 61). A growing body of literature has explored the role of the community college sector along the pathway to STEM baccalaureate success (e.g., Hu & Ortagus, in press; Wang, 2015), but little is known regarding whether increasing a community college’s commitment to STEM bachelor’s degree programs unintentionally limits the opportunities for students seeking to earn their associate degree in a STEM field. Given the limited resources of community colleges relative to four-year institutions (Hendrick, Hightower, & Gregory, 2006), offering baccalaureate degrees in STEM may lead community colleges to decrease their emphasis on associate degree production in STEM fields. This study seeks to complement previous scholarship related to the causes and consequences of CCB adoption by disaggregating between types of degree programs.


Essink’s (2013) mixed-methods study offers considerable insight into the motivations and challenges of community college leaders who adopted CCB degree programs. The author found that several community college presidents viewed CCB adoption as a way to join the ranks of nearby higher status colleges and universities. One community college administrator noted that two-year students constituted four-fifths of the student body, yet the administrator acknowledged that the two-year college mission was getting “pushed away” (p. 74).


Several participants in Essink’s (2013) study also expressed their commitment to the traditional community college mission related to developmental education and associate degree production. However, others indicated that CCB degree programs “had implications for the institution that strained those fundamental priorities” because the new bachelor’s degree programs “required time, energy, and resources from a system widely regarded as overburdened and underfunded” (p. 74). CCB-adopting degree programs tended to devote their already-limited resources to hiring new faculty members (often with PhDs) to teach in the baccalaureate degree programs, which brought a cultural shift aligned more with these faculty members’ professional training rather than the traditional mission of community colleges.


In sum, our review of the existing literature on CCB degree programs indicates that the purpose and research questions addressed in this manuscript have not been explored, at least empirically, in previous work. This gap in the literature precludes scholars, practitioners, and policymakers from fully understanding the impact of CCB degree programs, particularly related to the unintended consequences associated with the adoption of CCB degree programs.


THEORETICAL FRAMEWORK


Historically, community colleges were barred from offering four-year degrees, as the act would constitute a form of mission drift away from their traditional sub-baccalaureate mission (Labaree, 2010, 2017). Institutional theory suggests that colleges and universities often engage in mission drift to improve their status and standing relative to other institutions (Meyer & Rowan, 1977; Morphew & Huisman, 2002; Scott, 1987). More specifically, according to Morphew and Huisman (2002), institutional theory contends that colleges engage in mission drift by adopting new programs rather than trying to maximize efficiency in their market niche. Similar to other types of organizations, community colleges may seek to adopt the rituals, programs, processes, and structures that are seen as legitimate by actors outside of the organization, such as state policymakers and leaders of more prestigious colleges and universities (Meyer & Rowan, 1977; Scott, 1987). We apply institutional theory in this study to better understand the causes and consequences of CCB adoption. Although we recognize CCB degree programs as manifestations of mission drift and attempts to pursue status, we focus primarily on the seeming irrationality of pursuing new four-year degree programs at the expense of fulfilling a more traditional mission of awarding associate degrees.


Morphew and Huisman (2002) argue that institutional theory helps improve upon the idea of academic mission drift. In their application of institutional theory, Morphew and Huisman noted that “[i]n the process of trying to improve themselves with regard to status . . . institutions that engage in the practice of academic drift pay less attention to their pre-existing undergraduate programmes and these programmes suffer as a result” (p. 494). For CCB-adopting community colleges, offering baccalaureate degree programs and behaving like four-year universities would appear to take precedent over “practices that enhance the efficiency of their technical processes or the quality of their organizational outputs” across academic departments (Morphew & Huisman, 2002, p. 496).


In addition to institutional theory, we also draw upon Breneman’s (1976) theory of departmental behavior, which posits that faculty members and departments pursue prestige and resources, and that differences in departmental incentives explain differences in student outcomes, such as graduation rates. Although Breneman developed and applied the theory of departmental behavior to study graduate education, some of the larger points about the way that departments operate are applicable to departments more generally. The author suggested that faculty members in different departments are aware that their deans allocate resources and additional faculty lines based on departmental prestige (this is consistent with Essink’s work, cited above, about community college departments that adopted CCB degree programs). Thus, Breneman concluded that after departments “discover the basis for resource acquisition,” they “behave in accordance with the incentive system in order to maximize command over resources” (p. 14). Breneman argued that prestige can be measured not only in the number of graduates produced by a department but also by the job placement of a department’s graduates.


Because CCB degree programs appear to be in high-demand fields, Breneman’s (1976) work suggests that departments may seek to increase both associate and baccalaureate degree production of CCB degree programs at the relative expense of non-CCB degree programs. Community colleges would likely shift their already-limited resources away from departments or academic fields where graduates are not in high demand and thereby decrease associate degree production in non-CCB degree programs (given the relative lack of prestige, resources, and labor market demand). The logic described above suggests that stand-alone associate degree programs may have been neglected at CCB-adopting community colleges, as resources and attention were devoted to fields with the new four-year degree programs.


Based on our theoretical framework, we developed a series of quasi-experimental models to estimate whether CCB adoption led to decreases in associate degree production. Prior literature (e.g., Essink, 2013) and the application of institutional theory to explain mission drift (Morphew & Huisman, 2002) suggest that community colleges would decrease the number of associate degrees awarded in non-CCB fields after adopting CCB degree programs. Given these dynamics, we developed three hypotheses related to the interplay between CCB adoption and associate degree production.


Hypothesis 1: The adoption of a CCB degree program will lead to significant decreases in associate degree production.


Because community colleges remain committed to providing open-access education to all members of their surrounding community, public community colleges enroll a disproportionate number of historically underrepresented minority students (Bailey, Jaggars, & Jenkins, 2015). Accordingly, the diversity of the student body of community colleges motivates our second hypothesis:


Hypothesis 2: The decrease in associate degree production will have a larger impact on underrepresented minority students (e.g., Black and Hispanic) within higher education.


In addition to overall decreases in associate degree production, concentrated primarily in non-CCB adopting degree programs, we hypothesize that CCB adoption will lead to disproportionate changes in associate degree production for underrepresented minority students. Economists have shown that all types of students who earn associate degrees achieve significantly better labor market outcomes than their peers who start but do not complete college (e.g., Belfield & Bailey, 2017). As noted previously, many community college students who achieve improved labor market outcomes are likely to be underrepresented minorities given that over half of all Hispanic students and more than 40% of Black students enrolled in higher education attend community colleges (AACC, 2018).


Because previous studies outside of higher education scholarship, such as financial lending (Steiner-Khamsi, 2006), anti-smoking policies (Shipan & Volden, 2008), and organic farming (Läpple & Van Rensburg, 2011), have identified significant differences in outcomes between early adopters and late adopters of the identified policies, we hypothesize that the timing of CCB adoption may also play a role in the magnitude of the effect related to associate degree production. This hypothesis is formally depicted as follows:


Hypothesis 3: Early adopters of CCB degree programs will produce fewer associate degrees than their late-adopting peer institutions.


Given that the Florida legislative body broadened the scope of CCB legislation in 2007, CCB-adopting institutions can be categorized as early adopters (2001–2007) and late adopters (2008–2014). Figure 1 provides the time trend of CCB adoption in the state of Florida. As evident by the number of adoptions per year, there appears to be a first phase of adopters between 2001 and 2007 and a second stage of adopters between 2008 and 2014.



Figure 1. Time Trend of Total Number of Florida CCB Degree Programs


[39_22955.htm_g/00001.jpg]




DATA AND EMPIRICAL STRATEGY


To estimate the impact of CCB degree programs on associate degree production, we used institution- and program-level data for the 28 public community colleges within Florida. For this study, we operationalize program level by the two-digit Classification of Instructional Programs (CIP) code. CIP is a standardized reporting taxonomy generated by the U.S. Department of Education’s National Center for Education Statistics (NCES) to track and report fields of study and program completion activity. Program-level enrollment and degree completion data were provided by the Accountability Research and Measurement (ARM) and the PK-20 Education Reporting and Accessibility (PERA) divisions of the Florida Department of Education—the state authorizing entity of the Florida State College System. To this end, we created a dyadic panel dataset where the unit of analysis shifts from the institution (College X) to the dyad of institution and two-digit CIP code (CollegeX_CIP). This approach allows for micro-level analyses while also accounting for trends at individual program and institutional levels.


The data from the Florida Department of Education were then merged with data from NCES’s Integrated Postsecondary Education Data System (IPEDS) to gather institution-level factors. Factors such as cost of attendance, institution-level student demographics, and institutional expenditure data were included as control variables that are assumed to have influenced changes in enrollment and degree completion. In total, our dataset includes 5,349 observations for 435 distinct institution CIP dyads between 1998 and 2015.


INDEPENDENT VARIABLE


The independent variable of interest was the binary measure of CCB adoption at both the institutional and program levels (see Tables 1 and 2). Data on CCB adoption were provided directly by the Florida Department of Education and Florida State Colleges. Given that there is an application process associated with CCB degree program adoption, one could argue that the Florida Department of Education may bias program implementation by limiting CCB adoption to select institutions and programs, but the Florida Department of Education (and the Florida State Board of Education) has not rejected an application for CCB, to date, and therefore satisfies our concerns for the potential state-level adoption influence.


Table 1. Baccalaureate Degree Programs at Florida Community Colleges

Florida Community College

# of Baccalaureate Programs

Broward College

11

Chipola College

10

College of Central Florida

3

Daytona State College

11

Eastern Florida State College

3

Florida Gateway College

4

Florida SouthWestern State College

10

Florida State College at Jacksonville

14

Gulf Coast State College

4

Indian River State College

17

Lake-Sumter State College

1

Miami Dade College

16

Northwest Florida State College

7

Palm Beach State College

3

Pasco-Hernando State College

2

Pensacola State College

2

Polk State College

4

Santa Fe College

7

Seminole State College of Florida

5

South Florida State College

3

St. Johns River State College

3

St. Petersburg College

25

State College of Florida, Manatee-Sarasota

7

Valencia College

3



Table 2. Number of Baccalaureate Degree Programs by Two-Digit CIP

Two-Digit CIP Code

Number of  Baccalaureate Programs

Business, Management, Marketing, and Related Support Services

21

Health Professions and Related Programs

20

Education

14

Computer and Information Sciences and Support Services

9

Homeland Security, Law Enforcement, Firefighting & Related Protective Services

7

Engineering Technologies and Engineering-Related Fields

5

Public Administration and Social Service Professions

3

Communications Technologies/Technicians and Support Services

2

Biological and Biomedical Sciences

2

Legal Professions and Studies

1

Mechanic and Repair Technologies/Technicians

1

Visual and Performing Arts

1




DEPENDENT VARIABLE


Given that traditional higher education institutions cannot readily respond to increases or decreases in demand, this study examines potential changes in associate degree completion along with the proportion of degrees obtained by various student subgroups to gain a holistic view of the extent to which CCB adoption impacts associate degree production. To ease interpretation of results and generate additional models’ specification benefits, we used logged transformed dependent variables. The use of logged or percentage-based dependent variables increases the model’s efficiency and fit by reducing outliers. An additional benefit of a logged or percentage-based dependent variable is that it allows for the interpretation of results in elasticities.


Wooldridge (2009) stated that when predictors and/or dependent variables are expressed in percentage form, their original form efficiently estimates an elastic relationship. However, Meyer (1995) argues that difference-in-differences (DiD) estimations can be sensitive to the selected functional form. Specifically, DiD estimates can actually change their sign if a nonlinear transformation, such as a log, is applied to the dependent variable. To account for this potential limitation, we run model specifications that include our dependent variable in the form of total associate degrees produced as well as associate degrees per full-time equivalent students enrolled. Our results indicate that estimates are not dependent on our functional form choice. Thus, we follow the recommendation of Wooldridge (2009) and log transform our dependent variables for efficiency and ease of interpretation.


COVARIATES


To account for mitigating factors impacting associate degree production, we include institution- and county-level covariates. Specifically, we use institutional factors, such as total student population, proportion of full- to part-time students, in-state tuition and fees, instructional expenditures per FTE, and the percentage of students receiving Pell grants. Additionally, for the counties served by a local Florida State College, we include county-level unemployment rates, the proportion of county population between the ages of 18–25, and per capita income.


HETEROGENEOUS EFFECTS


In addition to testing the impact of CCB adoption on overall associate degree production, we also examine degree production by race/ethnicity. Specifically, we utilize the race/ethnicity definitions provided by the Florida Department of Education and focus on the degree production of non-Latino White, Black, and Hispanic students. We limited our analysis to these three subgroups, as 85% of all associate degrees are awarded by students self-identifying as a member of one of these three subgroups, and Black and Hispanic students represent the overwhelming majority of underrepresented minorities within the Florida College System. As previously discussed, prior scholars have identified the disproportionate impact of institutional polices on degree completion for underrepresented minority students in higher education. Second, we also focus on estimating the interaction effects between CCB adoption and (1) program type (STEM or non-STEM), (2) early versus late adopters, and (3) high versus low adoption—those adopting 10 or more CCBs during our analytical sample.


EMPIRICAL STRATEGY


This study uses two versions of a quasi-experimental approach to estimate the impact of the CCBs on associate degree production at both the institutional and CIP code or program levels. Prior work of scholars provides helpful examples of the effects of policy adoption on student outcomes. Select studies also provide examples of the use of the counterfactual approach when examining the impact of policy adoption (e.g., Khandker, Koolwal, & Samad, 2010), which can be understood in this context as estimating the effect of a scenario in which CCB was never implemented. To produce robust results, we employ a counterfactual approach at both the institutional and program levels.


INSTITUTION-LEVEL ANALYSIS


To estimate the overall impact of implementing CCBs, we first examine the institutional impact of CCB adoption on associate degree production. Specifically, we follow methodological approaches from Dynarksi (2000), Hillman, Tandberg, and Gross (2014), and Zhang and Ness, (2010), who each used a combination of the traditional OLS fixed-effects regression parameters and the DiD approach to study institutional and student responses to state-level policy adoption. In an environment where there is a single policy shock, we specified equation (1) as follows:


Yit = α + βInstitutioni + γPostt+ δ(Institutioni*Postt) + εit,

        (1)


where Institution i is an indicator valued at 1 if an institution adopted a CCB at any time during the analytical sample and zero otherwise, Post t has a value of 1 when the time period is after the adopting year—within our study, it would be the 2001 legislation providing authority to Florida community colleges to adopt CCBs—and zero otherwise. Institutioni*Post t is the DiD coefficient that represents the estimate of causal effects of CCB adoption on outcome Y for institution i during time t.


Although the legislative authority to adopt and implement CCB in Florida was provided in 2001 to all community colleges, not all institutions officially adopted a CCB that year. To account for the varying adoption period at the institutional level, we implemented a generalized DiD model. Building on the logic provided by Belasco, Rosinger, and Hearn (2014), we specify Equation (2) as our generalized difference-in-difference (GDD):


Yit = α + β1CCBit+ β2CCB#itYeart+ δCollegei+ εit,

        (2)


where CCBit is a dichotomous indicator equal to 1 during the year of first CCB adoption and afterward and 0 prior to adoption or if an institution has not adopted a CCB. CCB#it is a continuous indicator for the number of CCB degree programs at institution i during year t., Yeart is the year fixed-effect, and δCollegei is the institutional fixed effect. εit is the robust standard error. The quasi-experimental estimates from Equation 1 are approximated within Equation 2 through a two-way fixed effects (Bertrand, Duflo, & Mullainathan, 2004). To this end, the first difference between adopting and non-adopting institutions is accounted for within our institution fixed-effects (δCollegei), and the second difference between pre- and post-time periods is accounted for within our year fixed-effects (γYeart). The resulting coefficient on β1CCBit is our quasi-experimental DiD estimates.


Finally, we extend Equation 2 to account for mitigating factors that may impact associate degree production. Equation 3—our final model specification—now includes Xit, which is a vector of institutional and county level controls captured by β3Xit:


Yit = α + β1CCBit + β2CCB#it + β3Xit + γYeart + δCollegei + εit

        (3)


PROGRAM-LEVEL ANALYSIS


An additional contribution of this study is the use of program-level data to estimate more local impacts of CCB adoption on associate degree production. Similar to our institutional analysis, we first present the standard logic for our difference-in-difference-in-differences (DDD) approach with a single policy shock.


i=1,p=1,t>2001) - i=1,p=1,t≤2001)

        (4)


where i=1,p=1,t>2001) is the logged transformed changes in associate degrees awarded by student demographic groups for program p that adopted a baccalaureate program (treatment = 1), in institution i that has any baccalaureate program (treatment = 1) in years after 2001—the time at which the Florida enacted legislation allowing broad-based baccalaureate degrees within its community colleges.


i=1,p=1,t≤2001) is the same set of outcomes for the same program at the same institutions in the years prior to the implementation of CCB legislation in Florida. The simplicity of the model presented in Equation 1 provides an easy interpretation of the impact of the legislation; however, it fails to account for the global changes in associate degree completion, broader economic trends, or targeted initiatives by institutions to increase enrollment and success.


To identify the causal impact of CCB degree programs, additional sources of variations must be identified to explain away the exogenous impacts of enrollment and completion changes. Within this study, additional variations in the outcomes are partially examined by incorporating an additional DiD parameter to Equation 4. Since institutions can select the programs in which they adopt baccalaureate degree programs, conditions exist where treatment programs (programs in which a community college has decided to grant degrees) can be compared to untreated programs (programs in which a community college only grants associate degrees). With the assumption that the adoption of a CCB within a single program will only impact associate degree outcomes within that program, we can utilize the untreated CIP codes in the same way we compared institutions pre- and post-adoption in Equation 4. We can algebraically express this addition as follows:


=1,p=1,t>2001) - i=1,p=1,t≤2001)) - i=1,p=0,t>2001) - i=1,p=0,t≤2001))

(5)


where the first terms are identical to those in Equation 4 and exploit the within-institutional differences pre- and post-policy. However, this term is now subtracted from the differences between treatment and untreated CIP codes, resulting in a DiD estimation. This additional difference accounts for overall changes in associate degree production as well institution-specific changes in completion.


The final concern not addressed through Equation 5 is the potential for bias in the type of programs selected for CCB adoption. The decision to implement CCBs within certain programs is likely related to heightened student interest or local demand (Moore & Shulock, 2014). Within the state of Florida, degrees in business administration and teacher preparation were early adopters given statewide interest and shortages at the four-year institutions. To account for the potential bias in program adoption, we exploit data available for treated programs within untreated institutions. Equation 5 is thus modified as the following:


=1,p=1,t>2001) - i=1,p=1,t≤2001)) - i=0,p=0,t>2001) - i=0,p=0,t≤2001))

(6)


where now the initial difference pre- and post-adoption within an institution is subtracted from the differences between treated programs in untreated schools. In generating our DDD estimate, we combined equations (4), (5), and (6) to generate the following equation (7):


i=1,p=1,t>2001) - i=1,p=1,t≤2001)) - i=1,p=0,t>2001) - i=0,p=0,t≤2001)) -

i=0,p=1,t>2001) - i=0,p=1,t≤2001)) - i=0,p=0,t>2001) - i=0,p=0,t≤2001))

(7)


In a regression framework, this estimation strategy can be expressed as


Yipt = α + βInstitutioni + γPostt + ωProgramp + δ(Institutioni*Postt*Programp) + εit,

(8)


where Yipt, is the outcome of interest for institution i, program p, in year t. Institution i is a dummy variable that indicates a 1 for institutions adopting a CCB and zero for those that have not. Postt indicates the years post-CCB adoption period, and Programp indicates a 1 if a given program adopted a CCB at any point in the sample and 0 if it did not.


Using the same logic articulated within our institutional analysis, we utilize a generalized difference-in-difference-in-differences (GDDD) approach to account for the variations in adoption periods and CIP program adoption across both institutions and degree programs. To this end, we specify Equation 9 as follows:


ln (Y)ipt  = β0 + β1λipt + β2…nXit + αip + δit + γpy + υi + ρp + θt + εpti,

(9)


where ln (Y)ipt is the logged transformed of associate degrees completed by student demographic characteristics in college i for degree program p for year t. λipt is the coefficient of interest (GDDD indicator), which equals one in the year programs adopt a CCB in the adopting institution and thereafter; otherwise, the indicator equals zero. Xit is a vector of institutional characteristics that impact degree completion. The remaining terms represent a set of fixed effects. αip is vector of program by institutional fixed effects, δit is an institution by year fixed effects, γpy  is a vector of program by year fixed effects, and  υi , ρp , θt  are institutional, program, and year fixed effects. Finally, εpti represents the random idiosyncratic error term that fluctuates across time.


The benefits of this model specification is its full control of college-specific time effects common across academic programs combined with time-varying program effects. This specification models out any influence on the outcomes related to (1) specific programs at individual institutions, (2) specific programs across institutions for a particular year, and (3) specific institutions in a particular year. The estimates on  β1λipt  indicate the outcome and enrollment impacts related to the adoption of baccalaureate programs.


Limitations and Validation of Design Assumptions


A major challenge for any quasi-experimental design relates to the identification of the counterfactual in the absence of a policy adoption. The use of a GDDD design allows this study to approximate the impact of non-adoption in adopting programs by using non-adopting programs and non-adopting institutions as controls. This approach produces estimates of what could have occurred within the outcomes if the CCB had not been adopted. The assumption of this counterfactual approach is that treatment and control units following similar (or parallel) pre-policy patterns and the resulting variations in the outcome can be attributed to policy adoption. Figure 2 provides visual evidence supporting this assumption by showing the average degree production among non-adopting and adopting community colleges in Florida.


Figure 2. The average degree production among non-adopting and adopting community college in Florida.

[39_22955.htm_g/00002.jpg]



Although this assumption is difficult to test, this study utilizes two techniques to account for the parallel assumption. First, this study added a program- and institution-specific trend to the set of covariates (Angrist & Pischke, 2009; Belasco, Rosinger, & Hearn, 2014). This inclusion controls for the potential that adopting programs within adopting institutions—and adopting institutions as a whole—may have experienced differences in the outcomes of interest prior to adopting a CCB. Accordingly, program and institutional trend variables were created by regressing dummy time variables for the years prior to 2001 on each of the dependent variables and by multiplying the resulting coefficient by year to create a unique program and institutional trend variable. One concern with our design is the potential presence of various composition effects. To test for the possible presence of composition effects, we re-estimated our main-effects tables to include institution-by-year for our institution level analysis and institution-by-year and program-by-year for our program-level analysis and interaction terms. After employing these checks for composition effects, we obtain virtually the same results in both magnitude and directionality. Additionally, the Florida Department of Education has yet to deny any institution’s application to create a baccalaureate degree program. This point provides additional credibility that CCB legislation was not targeted in the subset of community colleges or programs.


Second, we utilize falsification or placebo tests (Cook & Campbell, 1986) to overcome a major concern with quasi-experimental approaches by untangling the policy effect from a potential corresponding time effect. To this end, we artificially created the adoption of CCB degree programs years prior to the actual adoption. This approach allows the results to be viewed in context. Significant results prior to the actual adoption signal that the estimated impact on associate degree production was not a product of CCB adoption, but rather a time effect that happens to coincide with the adoption of CCB degree programs.


Finally, given that we are estimating the effect of CCB legislation on the entire population of two-year institutions within the state of Florida, our study presents an interesting methodological challenge. Specifically, Abadie et al. (2014) showed that when the “sample” is the entire population, conventional methods for estimating standard errors are inefficient. The authors concluded that the standard error estimates are typically overestimated in such cases due to the lack of sampling error. Our study aligns with this context given that the population is the same as our sample. This indicates that the conservative standard errors within this study may underestimate the significance of estimated coefficients. Nonetheless, inferential statistics serve the purpose of estimating the expected value in the population, but we have access to population-level data over time and consider it important to assess the estimated effects without discarding probability levels that do not meet the 5% probability-level threshold utilized in the social sciences.


RESULTS


Tables 3 and 4 contain the means and standard deviations of total associate degrees for Florida’s 28 community colleges at the institution and program level, respectively. We also compared the proportional differences between CCB-adopting and non-adopting institutions in the final year prior to the adoption of the CCB legislation (2001) and the final year of our analytical sample (2015). For the total number of associate degrees, we found that non-adopting institutions, on average, produced 792 associate degrees in 2001 compared to 1,266 associate degrees for adopting institutions during that same year. The proportional difference grew very slightly from 60% in 2001 to 61% in 2015. We saw similar trends in associate degree production for subgroups of White and Black students. Taken together, our descriptive statistics appear to suggest that CCB legislation may not have affected associate degree production.



Table 3. Institution-Level Associate Degree Production Means and Standard Deviations

 

Pre-Legislation (2001)

 

Post-Legislation (2015)

 

Non-Adopting Colleges

Adopting Colleges

Proportional Difference

 

Non-Adopting Colleges

Adopting Colleges

Proportional Difference

Total Associate Degrees

791.75

1,265.58

1.60

 

1,652.50

2,653.25

1.61

 

(779.2656)

(961.9769)

  

(1773.304)

(2437.57)

 

Associate Degrees: White

576.25

875.29

1.52

 

812.00

1,245.54

1.53

 

(554.3545)

(546.5507)

  

(799.7879)

(745.6265)

 

Associate Degrees: Black

105.75

134.38

1.27

 

305.25

388.42

1.27

 

(116.4256)

(181.8671)

  

(339.9455)

(485.3807)

 

Associate Degrees: Hispanic

64.00

167.38

2.62

 

306.00

687.17

2.25

 

(89.52839)

(460.022)

  

(415.8036)

(1456.459)

 
        





Table 4. Program-Level Associate Degree Production Means and Standard Deviations

 

Pre-Legislation (2001)

 

Post-Legislation (2015)

 

Non-Adopting Programs

Adopting Programs

Proportional Difference

 

Non-Adopting Programs

Adopting Programs

Proportional Difference

Associate Degrees: Total

       

Programs

119.82

94.47

0.79

 

203.60

96.04

0.47

 

(417.66)

(115.12)

  

(825.63)

(150.65)

 

Institutions

89.26

118.09

1.32

 

129.61

184.04

1.42

 

(303.75)

(384.59)

  

(506.36)

(748.40)

 

Associate Degrees: White

       

Programs

99.11

64.35

0.65

 

134.09

63.35

0.47

 

(272.04)

(65.80)

  

(366.75)

(88.47)

 

Institutions

83.20

91.59

1.10

 

90.22

116.77

1.29

 

(243.96)

(239.02)

  

(277.41)

(319.01)

 

Associate Degrees: Black

       

Programs

15.39

13.72

0.89

 

44.94

14.60

0.32

 

(58.81)

(31.05)

  

(172.29)

(25.12)

 

Institutions

17.07

14.72

0.86

 

33.92

36.41

1.07

 

(53.04)

(53.49)

  

(115.79)

(150.21)

 

Associate Degrees: Hispanic

       

Programs

23.54

11.97

0.51

 

77.32

20.12

0.26

 

(185.15)

(40.87)

  

(464.90)

(58.59)

 

Institutions

10.50

22.02

2.10

 

34.00

64.42

1.89

 

(34.68)

(171.80)

  

(128.55)

(417.28)

 
        
        



INSTITUTION LEVEL: DIFFERENCE-IN-DIFFERENCES


Although this study also examines results at the academic program level, this section describes our institution-level generalized DiD approach to identify whether CCB-adopting institutions decreased their associate degree production following CCB adoption (Table 4). Table 5 illustrates two specifications for measuring the impact of CCB degree programs on associate degree production. The first is our primary specification, which was used to examine the effect of implementing at least one CCB degree program on institution-level degree production. The second specification estimates the interaction between our CCB adoption and a continuous indicator for the total number of active CCB degree programs within the institution, which was used to identify the impact of each additional CCB degree program on associate degree production.


Contrary to our first hypothesis, we found that the presence of CCBs had a positive impact on associate degree production, as community colleges that adopted a CCB degree program experienced a 6.4% increase in associate degree production. Additionally, the level of an institution’s commitment to CCB degree programs, as measured by accounting for the total number of active CCB degree programs, did not have a statistically significant impact on associate degree production. We also estimated the heterogeneous effects of CCB adoption on associate degree production by race/ethnicity subgroups (Table 5). CCB-adopting institutions experienced a 15.9% increase in the number of associate degrees awarded to White students and a 25.7% increase in the number of associate degrees awarded to Black students. The relationship between CCB adoption and the number of associate degrees awarded to Hispanic students was directionally positive but not statistically significant.



Table 5. Community College Baccalaureate and Associate Degree Production: Institution Level

 

Total Associate Degrees

 

White

 

Black

 

Hispanic

 

 

(1)

(2)

 

(1)

(2)

 

(1)

(2)

 

(1)

(2)

 

CCB Adopt

0.067*

0.064+

 

0.159+

0.105

 

0.257+

0.167+

 

0.170

0.053

 
 

(0.029)

(0.033)

 

(0.086)

(0.069)

 

(0.135)

(0.092)

 

(0.218)

(0.175)

 

Adopt * Total CCBs

 

0.001

  

0.012

  

0.020

  

0.026

 
  

(0.006)

  

(0.011)

  

(0.021)

  

(0.023)

 
             

Year Fixed Effects

Yes

Yes

 

Yes

Yes

 

Yes

Yes

 

Yes

Yes

 

Institutional Fixed Effects

Yes

Yes

 

Yes

Yes

 

Yes

Yes

 

Yes

Yes

 

Time-Varying Institutional Covariates

Yes

Yes

 

Yes

Yes

 

Yes

Yes

 

Yes

Yes

 

Time-Varying County Covariates

Yes

Yes

 

Yes

Yes

 

Yes

Yes

 

Yes

Yes

 
             

# of Observations

627

627

 

627

627

 

627

627

 

627

627

 

Adj. R2.

0.861

0.861

 

0.985

0.985

 

0.931

0.931

 

0.868

0.869

 

Note: Robust standard errors in parentheses; + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001




PROGRAM LEVEL: DIFFERENCE-IN-DIFFERENCE-IN-DIFFERENCES


Tables 6 and 7 outline the main and heterogeneous effects of program-level adoption of a CCB on associate degree production. Model 1 provides the overall effect of CCB adoption on associate degree production, Model 2 displays the interaction effect between CCB adoption and STEM CIP code, and Model 3 estimates the interaction effects between the number of CCB degree programs at an institutional level and individual two-digit CIP code adoption to compare low- and high-institutional adopters. Finally, Model 4 examines the interaction between two-digit CIP-level CCB adoption and whether the adoption was within the first legislation (2001–2007) or the second modified CCB legislation (2008–2014).


We found that the adoption of a CCB degree program within a given CIP code increased the number of associate degrees awarded by 17.9% two years post-adoption. When examining the differences based on STEM and non-STEM focused two-digit CIP codes—Model (2)—we uncover results that indicate significantly different responses between STEM and non-STEM degree programs. Specifically, adopting a CCB within a non-STEM CIP code led to a significant increase of 19.7% in the number of awarded associate degrees two years after CCB adoption, but adoption within a STEM CIP code led to a 20.6% decrease in associate degree production during the same time period. This indicates that the overall gain in associate degree production appears to be driven primarily by non-STEM degree programs. Consistent with Hypothesis 2, the decrease in associate degree production in STEM programs appears to be concentrated among Black and Hispanic students, as the number of STEM associate degrees awarded decreased two years after CCB adoption for Black students (44.8% decrease) and Hispanic students (29% decrease).


Additional evidence revealed that associate degree production was greater for community colleges with a small number (fewer than 10) of active CCB degree programs—a 38.5% increase—compared to community colleges with a greater emphasis on CCB degree programs (ten or more), which experienced a 10.7% increase in associate degree production after two years of CCB adoption. Consistent with Hypothesis 3, we found that the positive effects of CCB adoption on associate degree production were concentrated primarily within late-adopting institutions (a 53.1% increase) rather than early-adopting institutions (a 7% increase for early adopters).



Table 6. Community College Baccalaureate and Associate Degree Production: DDD Estimates (logged)

 

 

Associate Degrees

 

 

Overall

STEM

Adoption Level

Adoption Timing

CCB Adopt

 

0.179+

0.197+

0.385*

0.531***

  

(0.107)

(0.110)

(0.170)

(0.159)

      

CCB Adopt * STEM

  

-0.400*

  
   

(0.202)

  
      

CCB Adopt * High Adopter

   

-0.278

 
    

(0.202)

 
      

CCB Adopt * Early Adopter

    

-0.461*

     

(0.201)

      

Year Fixed Effects

 

Yes

Yes

Yes

Yes

Institutional Fixed Effects

 

Yes

Yes

Yes

Yes

CIP Fixed Effects

 

Yes

Yes

Yes

Yes

CIP by Year Fixed Effects

 

Yes

Yes

Yes

Yes

Institution by Year Fixed Effects

 

Yes

Yes

Yes

Yes

CIP by Institution Fixed Effects

 

Yes

Yes

Yes

Yes

Time-Varying Institutional Covariates

 

Yes

Yes

Yes

Yes

Time-Varying County Covariates

 

Yes

Yes

Yes

Yes

      

# of Groups

 

435.00

435.00

435.00

435.00

# of Observations

 

5,349

5,349

5,349

5,349

R2

 

0.387

0.387

0.387

0.388

Note: Robust standard errors in parentheses; + p < 0.10, * p < 0.05, ** p < 0.01, *** p < 0.001




Table 7. Community College Baccalaureate and Associate Degree Production by Race/Ethnicity: DDD Estimates (logged)

[39_22955.htm_g/00004.jpg]



FALSIFICATION AND ROBUSTNESS CHECKS


The difficulty in any quasi-experimental design lies in identifying the counterfactual in the absence of policy adoption. The use of both DiD and DDD designs allows this study to approximate the impact of non-adoption in the adopting two-digit CIP codes, using non-adopting institutions and non-adopting CIP codes within adopting institutions as controls. This approach yields estimates of what could have occurred with respect to our associate degree production outcomes had the CCB not been adopted. Such a counterfactual approach assumes that treatment and control units follow similar (or parallel) pre-policy patterns and that the resulting variations in the outcome can be attributed to policy adoption. A chief concern around any DiD (or DDD) approach is how to distinguish the policy impact from the time impact. To illustrate the lack of pre-adoption significance and to highlight potential time trends in our primary outcomes, we used leading and lagging adoption indicators to test the sensitivity of estimates to a false adoption year. We specifically focused on the year prior to adoption to assess whether two-digit CIP codes within adopting institutions made significant adjustments to their associate degree production in anticipation of CCB adoption; this prior dip is commonly called the Ashenfelter’s dip (Abadie, 2005).


Figure 3 shows the pre- and post-adoption trend for associate degree production. As seen in Figure 3, there were no significant changes prior to CCB adoption. Given the variations in CCB adoption years, robust estimates are indicated by statistically significant impacts on associate degree production after the adoption of a CCB policy. Two years after CCB adoption, institutions experienced their first significant increase in associate degree production. This significant effect held for the third year following CCB adoption. Given that the completion of an associate degree typically takes two years, one would not expect the full adoption effect until at least two years following CCB adoption. This display provides evidence that the prior-adoption parallel trend assumptions have been met, and that the estimates described earlier are robust and appear to be attributable to the impact of the CCB degree program adoption.



Figure 3. Institution-Level Falsification Timing Test


[39_22955.htm_g/00006.jpg]




Figure 4. Program-Level Falsification Timing Test


[39_22955.htm_g/00008.jpg]




DISCUSSION AND IMPLICATIONS


The increasing demand for skilled and educated workers, especially those possessing a bachelor’s degree, has created a strain on the institutions that supply higher education. In response to this dynamic, states have responded by providing community colleges with the authority to confer bachelor’s degrees, with a particular focus on workforce-based degree programs (Russell, 2010; Walker, 2005). However, associate degree production remains a core mission of the community college sector given that students who earn associate degrees have better, higher-paying jobs than students who earn high school diplomas (Belfield & Bailey, 2017; Minaya & Scott-Clayton, 2017).


The present study examines whether community colleges decreased their production of associate degrees after adopting CCB degree programs. The concern that community colleges would move away from associate degree production in response to CCB adoption was due to the perception of the inelasticity of higher education, but our results show that community colleges appear to possess the flexibility required to adapt to changes in their degree programs without sacrificing a core aspect of their institutional identity (e.g., associate degree production).


We offer two possible explanations for our findings at the institution and program levels. At the institution level, increases in overall associate degree production following CCB adoption signal that CCB-adopting institutions can supplement, rather than supplant, their traditional academic mission with CCB offerings. Although previous scholars have noted that different degree programs continue to compete for classroom space, instructors, and institutional resources (Liefner, 2003; Van Vught, 2008), the adoption of CCB degree programs appears to increase, rather than decrease, the production of associate degrees at CCB-adopting institutions. Further research is needed to identify why community colleges increase their associate degree production following CCB adoption.


At the program level, the significant increase in associate degree production within CCB-adopting programs can be explained by two distinct forces. Because the state of Florida must certify that the proposed CCB degree programs address local workforce shortages before they can be implemented, CCB adoption appears to signal an increased institutional commitment to local workforce and community needs in the form of prioritizing these high-demand CCB degree program areas. This institutional response can also lead to additional resources to be used for additional class sections, full-time instructors, and academic support—all three of which have been shown to enhance student success in higher education (Angrist, Lang, & Oreopoulos, 2009; Bettinger & Baker, 2011). In addition, CCB adoption may facilitate a streamlined degree plan to ensure that community college students complete their associate degree requirements as a stepping stone to earning their bachelor’s degree (Cruce, Wolniak, Seifert, & Pascarella, 2006), but further research is needed to empirically examine whether students enrolled in CCB degree programs earn their associate degree before completing their bachelor’s degree.


HETEROGENEOUS EFFECTS


Although our main findings show institution-level increases in associate degree production following CCB adoption, several interesting findings are revealed through our heterogeneous effects. First, program-level results show that CCB adoption has a positive impact on non-STEM associate degree production but a negative impact on STEM associate degree production. In line with Hypothesis 2, this decrease in associate degree production within STEM CIP codes appears to be concentrated primarily within Black and Hispanic student subgroups. STEM completion among community college students has been identified as a major concern, as 69% of community college students who begin as STEM majors either change majors or drop out within six years of enrolling in college (Chen, 2013). STEM attrition is especially disconcerting for underrepresented minority students, as they have been found to be less likely to persist in STEM fields when compared to non-minority students (Chang, Sharkness, Hurtado, & Newman, 2014).


Second, and in alignment with both Hypothesis 3 and much of the policy-based literature (e.g., Walker, 1969), our findings reveal that early adopters of CCB degree programs experience different outcomes than late-adopting institutions. This particular heterogeneous finding indicates that the positive impact of CCB adoption on associate degree production may diminish as institutions become more entrenched in their baccalaureate degree program offerings. Future research can examine whether this timing effect is a product of a long-run substitution between associate and bachelor’s degree programs within CCB-adopting institutions. Regardless of a timing effect related to CCB adoption, community colleges may consider programmatic and policy adjustments to ensure that CCB adoption is not disproportionately harming underrepresented minority students seeking to gain workforce training and postsecondary credentials, particularly those underrepresented student populations within STEM fields.


CONCLUSIONS


This study uses IPEDS data to evaluate the effect of CCB adoption on the production of associate degrees at both the institutional and program level. Our methodological approach, combined with the use of program-level data, improves upon prior research examining the role of CCB adoption on institutional degree production and decision-making (Floyd, Falconetti, & Hrabak, 2009; Hrabak, 2009; McKee, 2001; Townsend, 2005). We offer a novel contribution to the literature by analyzing the impact of CCB adoption at multiple levels. Although institutional theory would suggest that higher education institutions are continually striving to increase their status through mission expansion to bolster their perceived legitimacy, our results demonstrate that community colleges appear to be able to maintain (or improve upon) their historical mission of institution-level associate degree production while expanding their level of degree offerings to meet local workforce demands.


Our results at the program level show that associate degree production increased for CCB-adopting programs, but it remains unclear whether these results are a product of institutional support for CCB-adopting programs or merely a product of student interest in high-demand programs that offer the CCB. However, the positive impact of CCB adoption appears to be concentrated within non-STEM program areas. Because the national demand for STEM graduates has grown considerably in recent years due to rising employment rates in science and engineering fields (National Science Foundation, 2010), multiple studies have described the need for more STEM graduates (Anderson & Kim, 2006; Chen & Weko, 2009). Our findings suggest that policymakers seeking to implement CCB degree programs to respond to local workforce demands should also consider the potentially negative impact of CCB adoption on STEM associate degree production.


The availability of rich data allows this study to show conditional differences in associate degree production by degree type, the timing of CCB adoption, and a range of other heterogeneous effects. These program-level results suggest that earlier studies at the institutional level may have confounded the effects of CCB, as prior studies have failed to take advantage of disciplinary differences when examining the impact of CCB. In addition to our main effects, the previously identified heterogeneous effects speak to both institutional commitment and disciplinary differences. This study also provides insight into the interplay between mission expansion, responsiveness to local labor market conditions, and the traditional focus of community colleges on associate degree production. A growing number of states have already adopted or are considering adopting legislation that provides the authority to two-year institutions to award four-year baccalaureate degrees. This study explores the impact of adopting CCB degree programs on associate degree production and extends previous research by analyzing the effects of CCB adoption at both the program and institutional level. The constraints on four-year institutions to increase capacity and meet the local demand for baccalaureate degree completion create a need for innovative approaches to improve access to four-year degrees.


Opponents of CCB legislation have long argued that giving community colleges the authority to confer baccalaureate degrees will dilute the opportunity for community colleges to fulfill their sub-baccalaureate institutional mission, but our results suggest that the adoption of a CCB is associated with an overall increase in associate degree production. However, this effect does not hold for STEM associate degree production, which creates complications for community colleges seeking to improve degree completion, respond to local workforce demands, and address the need for more STEM graduates. Future research can explore how community colleges can explore how CCB-adopting institutions can improve their associate degree production in STEM disciplines, particularly given the national demand for STEM employees with postsecondary credentials.


Given that data are not currently available to estimate the long-term effects of CCB adoption, our results demonstrate its short-term effects and lay the foundation for analyzing the long-term impacts of CCB adoption on degree production. As the higher education market continues to respond to the increasing demand for baccalaureate degrees, CCB adoption may become an increasingly viable option for state policymakers considering mechanisms to expand access to four-year degree programs and address local workforce demands. Future studies can focus on other CCB-adopting states and continue to codify the effects of CCB adoption on associate degree production and other outcomes identified throughout this study.



References


Abadie, A. (2005). Semiparametric difference-in-differences estimators. The Review of Economic Studies, 72(1), 1-19.


Abadie, A., Athey, S., Imbens, G. W., & Wooldridge, J. M. (2014). Finite population causal standard errors (No. w20325). National Bureau of Economic Research.


American Association of Community Colleges. (2018). Fast facts 2018. Washington, DC.


Anderson, E. L., & Kim, D. (2006). Increasing the success of minority students in science and technology. American Council on Education.


Angrist, J., Lang, D., & Oreopoulos, P. (2009). Incentives and services for college achievement: Evidence from a randomized trial. American Economic Journal: Applied Economics, 1(1), 136–163.


Angrist, J. D., & Pinschke, J. S. (2009). Mostly harmless econometrics: An empiricst’s companion. Princeton, NJ: Princeton University Press.


Bailey, T. R., Jaggars, S. S., & Jenkins, D. (2015). Redesigning America's community colleges. Cambridge, MA: Harvard University

Press.


Belasco, A. S., Rosinger, K. O., & Hearn, J. C. (2015). The test-optional movement at America’s selective liberal arts colleges: A boon for equity or something else? Educational Evaluation and Policy Analysis, 37(2), 206–223.


Belfield, C., & Bailey, T. (2017). The labor market returns to sub-baccalaureate college: A review (CAPSEE Working Paper). Retrieved from http://ccrc.tc.columbia.edu/media/k2/attachments/labor-market-returns-sub-baccalaureate-college-review.pdf


Bemmel, E. P. (2008). A cost-effectiveness analysis of two community college baccalaureate programs in Florida: An exploratory study (Doctoral dissertation, Florida Atlantic University Boca Raton, Florida).


Bemmel, E. P., Floyd, D. L., & Bryan, V. C. (2008). Perceptions and reflections of administrators: Community colleges transitioning to baccalaureate colleges. Community College Journal of Research and Practice, 33(2), 151–176.


Bertrand, M., Duflo, E., & Mullainathan, S. (2002). How much should we trust differences-in-differences estimates? (No. w8841). National Bureau of Economic Research.


Bettinger, E., & Baker, R. (2011). The effects of student coaching in college: An evaluation of a randomized experiment in student mentoring. NBER Working Paper, (w16881).


Breneman, D. W. (1976). The Ph.D. production process. In J. T. Froomkin, D. T. Jamison, & R. Radner (Eds.), Education as an industry (pp. 3–52). Cambridge, MA: National Bureau of Economic Research.


Chang, M. J., Sharkness, J., Hurtado, S., & Newman, C. B. (2014). What matters in college for retaining aspiring scientists and engineers from underrepresented racial groups. Journal of Research in Science Teaching, 51(5), 555–580.


Chen, X. (2013). STEM attrition: College students' paths into and out of STEM fields. Statistical Analysis Report. NCES 2014-001. National Center for Education Statistics.


Chen, X., & Weko, T. (2009). Students who study science, technology, engineering, and mathematics (STEM) inpPostsecondary education. U.S. Department of Education NCES—161.


Cook, T. D., & Campbell, D. T. (1986). The causal assumptions of quasi-experimental practice. Synthese68(1), 141–180.


Cruce, T. M., Wolniak, G. C., Seifert, T. A., & Pascarella, E. T. (2006). Impacts of good practices on cognitive development, learning orientations, and graduate degree plans during the first year of college. Journal of College Student Development, 47(4), 365–383.


Daun-Barnett, N. (2011). Community college baccalaureate: A fixed effects, multi-year study of the influence of state policy on nursing degree production. Higher Education Policy, 24(3), 377–398.


Daun-Barnett, N., & Escalante, S. (2014). Local influences of community college baccalaureate legislation on nursing and teaching degree production in Florida. Community College Journal of Research and Practice, 38(11), 1030-1043.


DiMaggio, P., & Powell, W. (1983). The iron cage revisited: Institutional isomorphism and collective rationality in organizational fields. American Sociological Review, 48(2), 147–160.


Dynarski, S. (2000). Hope for whom? Financial aid for the middle class and its impact on college attendance (No. w7756). National Bureau of Economic Research.


Essink, S. E. (2013). The community college baccalaureate: A mixed methods study of implementation and best practices (Unpublished doctoral dissertation). University of Nebraska, Lincoln.


Florida College System. (2017). Bachelor’s degree programs. Retrieved from https://www.floridacollegesystem.com/students/programs.aspx


Florida Department of Education. (2005). History of the need for baccalaureates policy paper. Retrieved from http://www.fldoe.org/core/fileparse.php/5592/urlt/0082822-histbacc.pdf


Florida Department of Education. (2008). Baccalaureate programs in community colleges: A program review. Retrieved from http://www.fldoe.org/core/fileparse.php/5592/urlt/PR2008_02_Baccalaureate_Program_Review.pdf


Florida Senate. (2008). 1004.65 Community Colleges: Definition, mission, and responsibilities. Retrieved from http://www.flsenate.gov/laws/statutes/2008/1004.65


Floyd, D. L., Falconetti, A. M. G., & Hrabak, M. R. (2009). Baccalaureate community colleges: The new Florida College System. Community College Journal of Research and Practice,   33(2), 195.


Floyd, D. L., & Walker, K. P. (2009): The community college baccalaureate: Putting the pieces together. Community College Journal of Research and Practice, 33, 90–124.


Fulton, M. (2015). Community colleges’ expanded role into awarding bachelor's degrees. ECS Education Policy Analysis. Education Commission of the States. Retrieved from http://www.ecs.org.


Furlong, T. E., Jr. (2005). St. Petersburg College: Increasing baccalaureate access in critical program areas. In D. L. Floyd, M. L. Skolnik, & K. P. Walker (Eds.), The community college baccalaureate: Emerging trends and policy issues. Sterling, VA: Stylus Publishing.


Floyd, D. L., Falconetti, A. M., & Hrabak, M. R. (2009). Baccalaureate community colleges: The new Florida college system. Community College Journal of Research and Practice, 33(2), 195-202.


Gonzales, D. A. (2005). Identifying peer institutions: Creating a utilized-focused analysis of community colleges conferring select baccalaureate degrees. Retrieved from Proquest (UMI Number 3151912)


Henderson, C. E. (2014). The community college baccalaureate degree in the United States: An event history analysis. Retrieved from Proquest. (UMI Number 3638004).


Hendrick, R. Z., Hightower, W. H., & Gregory, D. E. (2006). State funding limitations and community college open door policy: Conflicting priorities? Community College Journal   of Research and Practice, 30(8), 627–640.


Hillman, N. W., Tandberg, D. A., & Gross, J. P. (2014). Performance funding in higher education: Do financial incentives impact college completions? The Journal of Higher Education85(6), 826–857.


Hrabak, M. R. (2009). The community college baccalaureate movement: Cutting-edge dissertation research. Community College Journal of Research and Practice, 33(2), 203–212.


Hu, X., & Ortagus, J. (in press). A national study of the influence of the community college pathway on female students’ STEM baccalaureate success. Community College Review.


Kerr, C. (2001, April 4). Letter to Senator Dede Alpert. Retrieved from http://www.ucop.edu/acadinit/mastplan/edd/kerralpert.pdf


Khandker, S. R., Koolwal, G. B., & Samad, H. A. (2010). Handbook on impact evaluation: quantitative methods and practices. Herndon, VA: World Bank Publications.


Labaree, D. F. (2010). Understanding the rise of American higher education: How complexity breeds autonomy. Peking University Education Review, 8(3), 24–39.


Labaree, D. F. (2017). A perfect mess: The unlikely ascendancy of American higher education. Chicago: Chicago University Press.


Läpple, D., & Van Rensburg, T. (2011). Adoption of organic farming: Are there differences between early and late adoption?. Ecological Economics70(7), 1406-1414.


Levin, J. S. (2004). The community college as a baccalaureate-granting institution. Review of Higher Education, 28, 1–22.


Liefner, I. (2003). Funding, resource allocation, and performance in higher education systems. Higher Education, 46(4), 469–489.


Lumina Foundation. (2017). Strategic plan for 2017 to 2020. Indianapolis, IN: Author.


McKee, J. V. (2001). Factors and issues surrounding development of one community college baccalaureate degree program. Retrieved from Proquest (UMI Number 251725700).


McKinney, L., Scicchitano, M., & Johns, T. (2013). A national survey of community college baccalaureate institutions. Community College Journal of Research and Practice, 37(1), 54–63.


Meyer, B. (1995). Natural and quasi-experiments in economics. Journal of Business and Economic Statistics, 13(2), 151–161.


Meyer, J. W., & Rowan, B. (1977). Institutionalized organizations: Formal structure as myth and ceremony. American Journal of Sociology, 340–363.


Minaya, V., & Scott-Clayton, J. (2017). Labor market trajectories for community college graduates: New evidence spanning the Great Recession (CAPSEE Working Paper). Retrieved from http://ccrc.tc.columbia.edu/media/k2/attachments/labor-market-trajectories-community-college-graduates-R1.pdf


Moore, C., & Shulock, N. (2014). From community college to university: Expectations for California's new transfer degrees. San Franscisco, CA: Public Policy Institute of California.


Morphew, C. C., & Huisman, J. (2002). Using institutional theory to reframe research on academic drift. Higher Education in Europe, 27(4), 491–506.


National Science Foundation (2010). Science and engineering indicators 2010. Arlington, VA: National Science Foundation.


Park, T. J., Tandberg, D. A., Shim, H., Hu, S., & Herrington, C. D. (2016). Community college teacher education baccalaureate programs: Early evidence yields mixed results. Educational Policy, 1–23.


Petry, D. K. (2006). The transformation of five Florida community colleges: Converting to baccalaureate degree-producing programs. Doctoral dissertation, The University of West Florida.


Povich, E. (2018, April 26). More community colleges are offering bachelor’s degrees—and four-year universities aren’t happy about it. Huffington Post. Retrieved from https://www.huffingtonpost.com/entry/more-community-colleges-are-offering-bachelors-degrees_us_5ae1de02e4b0d538c22dfbfe


Rephann, T. J. (2007). Community college growth opportunities: untapped potential in America's heartland? Growth and Change, 38(3), 443–459.


Russell, A. (2010). Update on the community college baccalaureate: Evolving trends and issues. American Association of State Colleges and Universities. Retrieved from http://www.aascu.org/policy/publications/policymatters/2010/communitycollegebaccalaureate.pdf


Shipan, C. R., & Volden, C. (2008). The mechanisms of policy diffusion. American Journal of Political Science52(4), 840–857.


SteinerKhamsi, G. (2006). The economics of policy borrowing and lending: A study of late adopters. Oxford Review of Education32(5), 665–678.


Stevens, M., & Kirst, M. W. (Eds.). (2015). Remaking college: The changing ecology of higher education. Stanford, CA: Stanford University Press.


Van Vught, F. (2008). Mission diversity and reputation in higher education. Higher Education Policy, 21(2), 151–174.


Walker, J. (1969). The diffusion of innovations among the American states. American Political Science Review, 63(3), 880–890.


Walker, K. P. (2005). History, rationale, and the CCBA. In D. L. Floyd, M. L. Skolnik, & K. P. Walker (Eds.), The community college baccalaureate: Emerging trends and policy issues (pp. 9–23). Sterling, VA: Stylus Publishing, LLC.


Wang, X. (2015). Pathway to a baccalaureate in STEM fields: Are community colleges a viable route and does early STEM momentum matter? Educational Evaluation and Policy Analysis37(3), 376–393.


Wooldridge, J. (2009). Introductory econometrics: A modern approach (4th ed.). Mason, OH: South-Western Cengage Learning.


Zhang, L., & Ness, E. C. (2010). Does state merit-based aid stem brain drain? Educational Evaluation and Policy Analysis, 32(2), 143–165.





Cite This Article as: Teachers College Record Volume 122 Number 1, 2020, p. 1-36
https://www.tcrecord.org ID Number: 22955, Date Accessed: 12/2/2021 11:58:52 PM

Purchase Reprint Rights for this article or review
 
Article Tools
Related Articles

Related Discussion
 
Post a Comment | Read All

About the Author
  • Justin Ortagus
    University of Florida
    E-mail Author
    JUSTIN C. ORTAGUS is Assistant Professor of Higher Education Administration & Policy and Director of the Institute of Higher Education at the University of Florida. His research typically examines the growing impact of online education and technology, the role and influence of community colleges, and the effects of various state policies on the opportunities and outcomes of historically underrepresented students.
  • Dennis Kramer
    University of Florida
    E-mail Author
    DENNIS A. KRAMER II is Assistant Professor of Higher Education Administration & Policy and Director of the Educational Policy Center at the University of Florida. His prior research has focused on the economics of higher education, the evaluation of federal and state policy adoption, and impact of state decisions on community colleges and four-year institutions.
  • Manuel González Canché
    Graduate School of Education, University of Pennsylvania
    E-mail Author
    MANUEL S. GONZÁLEZ CANCHÉ is Associate Professor of Higher Education at the University of Pennsylvania. His research examines issues of access, persistence, and success, with an emphasis on institution effects on students’ outcomes. He also focuses on higher education finance, with emphases on spatial modeling and competition based on spatial proximity and spillover effects.
  • Frank Fernandez
    College of Education, University of Houston
    E-mail Author
    FRANK FERNANDEZ is Assistant Professor of Higher Education at the University of Houston. His research typically examines community colleges, legal issues, graduate education, international/comparative education, and institutional theory.
 
Member Center
In Print
This Month's Issue

Submit
EMAIL

Twitter

RSS