Dual Credit,College Type,and Enhanced Degree Attainment

Abstract

In this study, we analyzed data for the Illinois high school class of 2003 to determine the impact of dual credit participation on postsecondary attainment. We matched 8,095 dual credit participants to an equal number of nonparticipants within the same high school at the point of postsecondary enrollment using propensity scores calculated through a two-level hierarchical generalized linear model with college type as a posttreatment adjustment. Results indicate that community college students taking dual credit were significantly more likely to obtain a baccalaureate degree (28% to 19%, respectively) than their matched peers who did not participate in dual credit. We found similar improvements in postsecondary degree attainment for students attending colleges at multiple selectivity levels, although the largest and most robust effects were found for students starting at community colleges and at non/less competitive entry institutions.

Keywords

achievement community colleges curriculum degree attainment dual credit dual enrollment educational policy regression analysis

This study is an analysis of the impact of dual credit participation on postsecondary degree attainment. It is a follow-up to prior studies we conducted as a joint project between the Illinois Board of Higher Education and the Illinois Education Research Council using a statewide data set of the class of 2003 Illinois public high school graduates. Several researchers have examined the potential benefits of participation in dual credit courses, college courses that also count for high school credit. Such benefits include improving postsecondary enrollment, GPA, retention, and completion; facilitating the transition between high school and college; elevating student college aspirations; aiding student adjustment to college; enhancing college readiness (An 2013, 2015; Lichtenberger, Witt, Blankenberger, & Franklin, 2014; Taylor, 2015); and increasing academic performance for Career and Technical Education (CTE) students (Wang, Chan, Phelps, & Washbon, 2015).

A few studies have focused on dual credit and degree completion. Of the studies employing large data sets, only those of An (2013), Wang et al. (2015), and Taylor (2015) have focused on postsecondary degree attainment and also employed propensity score matching and sensitivity analysis in their studies. An’s study using National Center for Education Statistics NELS:88 data found a positive impact of dual credit participation on postsecondary degree attainment. Wang et al. focused on the positive impact of dual credit participation on community college CTE students only. Taylor employed a subset of the Illinois data we used for our enrollment study (2014) including data from 12 of Illinois’s 48 community colleges. He analyzed both likelihood of enrollment and completion together, not completion conditional on enrollment. Not surprisingly then, his results indicated a much greater impact of dual credit participation on completion than An’s.

Our present study builds on these studies and our prior studies using the 2003 Illinois graduating class data. In our initial study, we conducted an exploratory analysis of this data set using Cox regression to identify factors associated with increased likelihood of postsecondary completion (Blankenberger, Lichtenberger, Witt, & Franklin, 2017). We found that participation in dual credit was one of several factors that were significantly related to greater likelihood of completion. Those students starting their postsecondary careers at a community college in the fall after high school graduation who had participated in a dual credit course were significantly more likely to complete a baccalaureate degree than their peers who did not participate in dual credit (1.026 odds ratio). This relationship was even stronger for those students who were low income (1.076 odds ratio). Given the possible importance of this for education policy in Illinois, we conducted a follow-up study focusing on the impact of dual credit on postsecondary enrollment (Lichtenberger, Witt, Blankenberger, & Franklin, 2014). Using propensity score weighting to address selection bias and controlling for a large number of covariates, we found that dual credit participation was significantly related to substantially higher odds of enrolling at both community colleges (8.027 odds ratio) and baccalaureate institutions (6.686 odds ratio).

For our current study, we analyzed the impact of dual credit participation on postsecondary completion within seven years of graduation from high school using the same statewide data set as our previous studies. We controlled for selection bias using a nearest neighbor propensity score matching approach along with a posttreatment adjustment for college type to determine whether the effect would vary across institutional settings. Out of the total class of about 115,000 public Illinois high school graduates from 2003, 63,495 directly enrolled in college the fall semester of 2003. Of this group, 10,987 participated in dual credit courses. Among these, we were able to successfully match 8,095 dual credit participants to an equal number of nonparticipants within the same high school and with a similar student-level profile across a large number of observed baseline covariates in the hierarchic generalized linear model (HGLM). We defined college type using Barron’s selectivity rankings and added community college as another category. Additional data on college enrollment and completion were acquired from the National Student Clearinghouse (NSC) and incorporated into the data set, allowing us to track student completions at out-of-state institutions. We attempted to isolate the impact of dual credit on completion from that of enrollment by conducting propensity score matching for the students conditional on their postsecondary enrollment. That is, unlike Taylor’s (2015) study, dual credit and non–dual credit students were matched only from those who were enrolled in postsecondary education. Also, unlike Taylor’s, we used the full statewide data set of the 2003 high school graduates. We employed a large number of covariates to strengthen our design (see Table 1). Propensity scores were calculated using a HGLM, with students matched within high schools to try to control for school- and student-level covariates. We determined that the matching process had provided adequate balance on the baseline covariates used in the HGLM (see Table 1). We used inferential and predictive analyses (multinomial logistic regression) to determine the overall impact of dual credit participation on postsecondrary degree attainment. We employed sensitivity analysis to determine whether the results were robust to unobserved confounding factors.

Table 1

Propensity Score Balancing Table

	Prior to Matching			After Propensity Score Matching (Nearest Neighbor)
Covariate	Dual Credit (N = 10,987)	No Dual Credit (N = 52,508)	Standardized Differences	Dual Credit (N = 8,095)	Non–Dual Credit (N = 8,095)	Standardized Differences
School percent White	0.81	0.71	0.38^a	0.80	0.80	0.00
School percent limited English proficient	0.02	0.03	−0.17^b	0.02	0.02	0.00
School percent low income	0.20	0.21	−0.05	0.17	0.17	0.00
School ACT average	20.02	20.55	−0.25^a	20.11	20.11	0.00
ACT Math	21.51	21.55	−0.01	21.10	20.51	0.11^b
ACT English	21.2	21.11	0.02	20.76	19.96	0.14^b
ACT Reading	21.85	21.61	0.04	21.40	20.62	0.13^b
ACT Science	21.46	21.34	0.02	21.11	20.52	0.13^b
High school GPA: 3.5+	0.28	0.25	0.07	0.25	0.20	0.12^b
High school GPA: 3.0–3.4	0.22	0.22	0.02	0.23	0.22	0.02
High school GPA: 2.5–2.9	0.13	0.14	−0.03	0.14	0.18	−0.11^b
High school GPA: <2.5	0.13	0.15	−0.06	0.15	0.17	−0.05
High school GPA: Missing	0.24	0.25	−0.03	0.24	0.24	0.00
Completed core curriculum	0.46	0.49	−0.06	0.44	0.41	0.06
Did not complete core	0.45	0.42	0.06	0.47	0.49	−0.04
Missing answer about core	0.09	0.1	−0.03	0.10	0.11	−0.03
High school program: College prep	0.45	0.43	0.04	0.43	0.37	0.13^b
High school program: Career and Technical Education	0.11	0.1	0.02	0.11	0.13	−0.06
High school program: General	0.2	0.22	−0.05	0.21	0.24	−0.07
High school program: Missing	0.24	0.25	−0.02	0.25	0.26	−0.02
Highest planned degree: < Two-year degree	0.02	0.01	0.04	0.02	0.03	−0.06
Highest planned degree: Associate	0.05	0.04	0.03	0.05	0.06	−0.04
Highest planned degree: Bachelor’s	0.28	0.28	−0.01	0.28	0.30	−0.04
Highest planned degree: Master’s	0.17	0.19	−0.04	0.17	0.17	0.00
Highest planned degree: Professional	0.25	0.24	0.02	0.23	0.21	0.05
Highest planned degree: Other	0.03	0.03	−0.02	0.03	0.03	0.00
Highest planned degree: Missing	0.21	0.21	0.01	0.21	0.21	0.00
Gender: Male	0.40	0.46	−0.12^b	.40	.46	−0.12^b
Race: White	0.75	0.64	0.25^a	0.74	0.72	0.05
Race: Latino	0.04	0.07	−0.11^b	0.04	0.04	0.00
Race: Asian	0.03	0.05	−0.11^b	0.03	0.03	0.00
Race: African American	0.06	0.1	−0.13^b	0.06	0.07	−0.04
Race: Other	0.03	0.04	−0.05	0.03	0.04	−0.05
Race: Missing	0.10	0.10	−0.07	0.10	0.10	0.00
Income: High $80,000+	0.16	0.20	−0.10^b	0.16	0.15	0.03
Income: Mid high $50,000–<$80,000	0.20	0.18	0.05	0.19	0.19	0.00
Income: Mid low $30,000–<$50,000	0.20	0.18	0.05	0.20	0.20	0.00
Income: Low <$30,000	0.14	0.14	0.00	0.14	0.17	−0.08
Income: Missing	0.30	0.30	0.00	0.30	0.30	0.00
Aid expectation: Yes	0.65	0.63	0.04	0.64	0.64	0.00
Aid expectation: No	0.13	0.15	−0.06	0.14	0.15	−0.03
Aid expectation: Missing	0.22	0.22	0.01	0.22	0.22	0.00
College type preference: Public four-year	0.50	0.54	−0.08	0.50	0.50	0.00
College type preference: Private four-year	0.12	0.13	−0.03	0.10	0.09	0.03
College type preference: Community college	0.11	0.07	0.14^b	0.11	0.12	−0.03
College type preference: Private two-year	0.01	0.01	0.00	0.01	0.01	0.00
College type preference: Vo-tech school	0.02	0.01	0.08	0.02	0.02	0.00
College type preference: School of nursing	0.01	0.01	0.00	0.01	0.01	0.00
College type preference: Missing	0.23	0.24	−0.02	0.23	0.24	−0.02
AP English: Yes	0.34	0.3	0.09	0.32	0.28	0.09
AP Math	0.32	0.29	0.07	0.29	0.26	0.07
AP Social studies	0.23	0.25	−0.05	0.23	0.19	0.10^b
AP Science	0.29	0.27	0.05	0.27	0.23	0.09
AP foreign language	0.17	0.19	−0.05	0.17	0.15	0.05
Expected work hours in college: None	0.10	0.12	−0.08	0.09	0.08	0.04
Expected work hours in college: 1–10	0.17	0.19	−0.05	0.17	0.15	0.05
Expected work hours in college: 11–20	0.32	0.31	0.02	0.33	0.34	−0.02
Expected work hours in college: 21–30	0.15	0.13	0.06	0.15	0.16	−0.03
Expected work hours in college: 31+	0.04	0.03	0.06	0.04	0.05	−0.05
Expected work hours in college: Missing	0.22	0.23	−0.02	0.22	0.23	−0.02
State preference: In state (IL)	0.59	0.55	0.08	0.59	0.57	0.04
State preference: Out of state	0.18	0.20	−0.06	0.18	0.19	−0.03
State preference: Missing	0.24	0.24	0.00	0.24	0.25	−0.02
Expected distance from home: <10 miles	0.07	0.05	0.09	0.07	0.07	0.00
Expected distance from home: 10–25	0.12	0.09	0.10	0.13	0.11	0.06
Expected distance from home: 26–100	0.18	0.16	0.05	0.18	0.18	0.00
Expected distance from home: 100+	0.21	0.24	−0.07	0.20	0.20	0.00
Expected distance from home: No college in mind yet	0.2	0.23	−0.07	0.20	0.21	−0.02
Expected distance from home: Missing	0.22	0.23	−0.02	0.22	0.24	−0.05
Live: Residence hall	0.39	0.48	−0.18^a	0.37	0.35	0.04
Live: Off campus	0.18	0.14	0.11^a	0.19	0.21	−0.05
Live: Parent’s house	0.18	0.13	0.14^a	0.19	0.18	0.03
Live: Married housing	0.01	0.01	0.00	0.01	0.01	0.00
Live: Fraternity or sorority	0.03	0.04	−0.05	0.03	0.04	−0.05
Live: Missing	0.21	0.20	0.03	0.21	0.21	0.00
BA rate	0.51	0.52	−0.02	0.48	0.40	0.16^a
Propensity	0.40	0.13	1.37	0.35	0.34	0.03
Barron’s: Most/highly competitive	0.10	0.12	−0.06	0.08	0.08	0.00
Barron’s: Very competitive	0.09	0.14	−0.15^b	0.08	0.08	0.00
Barron’s: Competitive	0.21	0.26	−0.12^b	0.20	0.20	0.00
Barron’s: Less/noncompetitive	0.07	0.06	0.04	0.06	0.06	0.00
Community college	0.51	0.40	0.23	0.58	0.58	0.00
Bachelor’s rate chi-square	73.169		0.00	97.994		0.00
.25 SD of propensity	0.0375
Degree or credential	0.65	0.61	0.09	0.64	0.57	0.15^a
No degree/credential but still enrolled	0.07	0.08	−0.04	0.08	0.10	−0.07
No degree and not enrolled	0.28	0.31	−0.07	0.29	0.33	−0.09
B^a	0.51	0.52	−0.02	0.48	0.40	0.16^a
Associate	0.12	0.07	0.17^a	0.13	0.15	−0.06
Certificate	0.03	0.01	0.14^a	0.03	0.02	0.06
No degree	0.35	0.39	−0.08	0.36	0.43	−0.14^a

Note. Standardized differences, D, are based on the following formula:

$D = ({\bar{x}}_{treatment} - {\bar{x}}_{control}) / \sqrt{(s_{treatment}^{2} + s_{control}^{2}) / 2} .$

Standardized differences over .20, which are considered significantly different.

Cells between .10 and .20 are considered indicative of a slight imbalance.

Given the results of our initial Cox regression analysis, we hypothesized that students who took dual credit courses would be significantly more likely to complete a postsecondary degree than their matched peers who did not take dual credit. We examined outcomes separately for postsecondary certificate, associate and baccalaureate degrees combined, and baccalaureate degrees alone. Also, given the positive results on completion for community college students we found in the initial Cox study, we hypothesized that this impact would vary by level of college selectivity. Our new analysis supported these hypotheses. Degree completion results overall and for most levels of selectivity were significantly better for students participating in dual credit compared to their matched peers who did not participate. This was true both for attaining any postsecondary degree and for attaining a baccalaureate degree (see Table 2).

Table 2

Results of Dual Credit Participation Comparison After Propensity Score Matching: College Type by Highest Degree Achieved

Institution Type for Student’s First Year of Postsecondary Attendance	Dual Credit Student	Non–Dual Credit Comparison	Absolute Difference	F Statistic	p Value	Gamma (Γ)
Community college (n = 9,390)
Percentage with bachelor’s degree or higher	28	19	9	112.993	<.001	2.0
Percentage obtaining any postsecondary credential	53	45	8	57.736	<.001	<1.5
Less/noncompetitive (n = 1,046)
Percentage with bachelor’s degree or higher	56	48	8	7.118	.008	<1.5
Percentage obtaining any postsecondary credential	64	57	7	4.641	.031	<1.5
Competitive (n = 3,194)
Percentage with bachelor’s degree or higher	73	67	6	11.825	.001	<1.5
Percentage obtaining any postsecondary credential	76	70	6	14.47	<.001	<1.5
Very competitive (n = 1,244)
Percentage with bachelor’s degree or higher	80	77	3	2.115	.146	<1.5
Percentage obtaining any postsecondary credential	83	80	3	0.898	.344	<1.5
Highly/most competitive (n = 1,316)
Percentage with bachelor’s degree or higher	90	84	6	10.563	.001	<1.5
Percentage obtaining any postsecondary credential	92	85	7	14.21	<.001	<1.5
Overall (N = 16,190)
Percentage with bachelor’s degree or higher	48	40	8	98.578	<.001	<1.5
Percentage obtaining any postsecondary credential	64	57	7	79.24	<.001	<1.5

The positive impact on degree completion was greatest for students starting at community college and less/noncompetitive colleges, although some positive impact was found for all results. The smallest impact was found for students starting at very competitive institutions, where the results were not significant. However, only the baccalaureate completion results for community college students who had participated in dual credit when compared to their matched peers who did not were significant and robust to possible confounding variables. These dual credit community college students finished a bachelor’s degree at a 28% rate compared to their non–dual credit peers who finished at a 19% rate, a 9 percentage point difference. The results were relatively robust to sensitivity analysis, with a Γ of 2. This is especially significant given that in Illinois, community colleges do not grant baccalaureate degrees. So, every student who started at a community college had to transfer to complete a baccalaureate degree. Results for other college selectivity levels indicated substantial differences between dual credit participants and nonparticipants in completion of both any postsecondary credential and baccalaureate degrees; however, the sensitivity analysis findings were not robust to possible hidden bias, yielding a Γ < 1.5. This indicates that there may be unobserved confounding factors that could be affecting the results of the analysis. For example, there is no differentiation between CTE and degree-oriented courses in this data set, and this could significantly impact the analysis. Thus, it is likely that this study may underestimate the positive impact of dual credit participation on baccalaureate completion for community college students. If we were able to isolate the dual credit courses to exclude CTE courses and only include those that are meant to apply to degrees, this would likely strengthen the connection between dual credit participation and degree completion.

Nonetheless, the results of this study offer further evidence of the benefits of dual credit participation on postsecondary completion. The primary contribution of our study is the differentiation of impact for those dual credit participants starting at community colleges. The results of this study indicate that the greatest improvement to degree attainment associated with dual credit participation accrued to those students who began their postsecondary careers at community colleges rather than more selective institutions. This was true even when community college students had to transfer to complete a baccalaureate degree, as in Illinois. This is especially important given the growing numbers of students attending community colleges and the need to create pathways that improve the likelihood of higher degree attainment for these students. In Illinois, community college students make up about 40% of postsecondary enrollments each fall, making improving their success even more crucial to meeting state degree completion goals.

Unlike prior studies, for our analysis, we matched at the point of enrollment in postsecondary. Matching at postsecondary enrollment appears to have reduced the magnitude of impact on postsecondary completion compared to the results of Taylor (2015). This indicates that some of the prior results showing degree completion gains by dual credit participants could have been due to increased likelihood of enrollment. Furthermore, with the exception of our findings related to baccalaureate completion, our results were not as robust to confounders as earlier studies (An, 2013; Taylor, 2015), suggesting that the analysis of the positive benefits of dual credit may benefit from additional studies that match at the point of enrollment.

Lastly, our study results support the findings of An (2013), who found an 8 percentage point increase in likelihood of completing some postsecondary degree for dual credit participants and a 7 percentage point increase in baccalaureate completion for dual credit participants. Our results were similar overall for dual credit participants, with a 7 percentage point increase for achieving any postsecondary credential and an 8 percentage point increase for bachelor’s degree completion when compared to matched nonparticipants. However, An’s results were based on students taking at least two dual credit courses, and he found little improvement in likelihood of postsecondary degree attainment for those who took only one dual credit course. In our study, the degree attainment improvements for dual credit students occurred even for students who took only one dual credit course.

Footnotes

Authors

BOB BLANKENBERGER, PhD, is an associate professor in the public administration department at the University of Illinois Springfield, One University Plaza, PAC 420, Springfield IL 62703; rblan2@uis.edu . His research focuses on educational policy, assessment, and educational attainment.

ERIC LICHTENBERGER, PhD, is the deputy director for information management and research with the Illinois Board of Higher Education, 1 N. Old State Capitol Plaza, Suite 333, Springfield, Illinois 62701; lichtenberger@ibhe.org . His research focuses on education policy and high school to college transitions.

M. ALLISON WITT, PhD, is a teaching assistant professor and the director of international programs in the College of Education at the University of Illinois Urbana-Champaign, 363 Education Building, 1310 S. Sixth St., Champaign, IL 61820; awitt1@illinois.edu . Her research focuses on internationalizing teacher education.

References

B. P.

(2013). The impact of dual enrollment on college degree attainment: Do low-SES students benefit? Educational Evaluation and Policy Analysis, 35(1), 57–75.

B. P.

(2015). The role of academic motivation and engagement on the relationship between dual enrollment and academic performance. The Journal of Higher Education, 86(1), 98–126.

Blankenberger

Lichtenberger

Witt

M. A.

Franklin

(2017). Diverse students, high school factors, and completion agenda goals: An analysis of the Illinois class of 2003. Education and Urban Society 49(5), 518–545.

Lichtenberger

Witt

M. A.

Blankenberger

Franklin

(2014). Dual credit/dual enrollment and data driven policy implementation. Journal of Community College Research and Practice, 38(11), 959–979.

Taylor

J. L.

(2015). Accelerating pathways to college the (in)equitable effects of community college dual credit. Community College Review, 43(4), 355–379.

Wang

Chan

H. Y.

Phelps

L. A.

Washbon

J. I.

(2015). Fuel for success academic momentum as a mediator between dual enrollment and educational outcomes of two-year technical college students. Community College Review, 43(2), 165–190.