Randomized Controlled Trials in Community Colleges Successes Challenges
- Slides: 22
Randomized Controlled Trials in Community Colleges Successes, Challenges, and Next Steps for Research on Developmental Education U. C. Davis March 16, 2015 Alexander K. Mayer Research Associate, MDRC
Developmental Education � Community College Students � Majority enters college under-prepared � Financial need � About 2/3 do not earn a degree or certificate within 6 years �Lots of reform � Limited systematic research � Status quo is changing – implications for research 2
Overview of Four Projects � Learning � Early Communities evaluation and multi-site demonstration � Performance-based � Early scholarships evaluation and multi-site demonstration � Comprehensive program � ASAP � National � 3 3 center to study developmental education large studies and supplemental studies
Learning Communities
Kingsborough: Opening Doors Learning Communities � Design – one semester: � Co-enrollment � Instructor collaboration � Curricular integration � Additional supports (counseling, tutoring, textbooks) � Theory of change � Greater Engagement � Stronger relationships � Benefits of extra support 5
Early Results Kingsborough Learning Communities � Strong implementation � Support from college administration � Students felt more integrated � Impacts during the program term � Students attempted and passed more courses � Earned more credits (about 1. 2 credits) � Advanced more quickly through developmental education � Impacts � No after 4 semesters impact on persistence � Earned more credits (about 2. 4 credits) 6
Long-term Results Long-term follow-up at Kingsborough 60 6 5 Cumulative credits earned 44. 8 42. 0 40 2 52. 3 49. 0 48. 1 3 39. 4 33. 9 30 20 52. 8 4 50 56. 3 31. 1 1 20. 3 18. 2 10 0 Year 1 7 Year 2 Year 3 Program group Year 4 Control group Year 5 Year 6
Learning Communities Demonstration Queensborough Community College: Linked developmental math with a collegelevel course Merced College: Linked developmental English and a variety of courses Kingsborough Community College: Developmental only Community College of Baltimore County: Linked developmental English, a college-level course, and seminar Houston Community College: Linked developmental math with a student success course 8 Hillsborough Community College: Linked developmental reading with a student success course
Results Learning Communities Demonstration 18 0. 5 Cumulative Credits Earned 16 15. 6 14 0. 6* 12 11. 9 10 8 6 4 15. 1 11. 4 Program Group 0. 5*** 6. 9 Control Group 6. 4 2 0 Semester 1 Semester 2 8 9 Semester 3
Taking Stock: Learning Communities � Variation in the programs across the demonstration � Curricular integration was inconsistent � Some scaling challenges, especially in first two semesters � Modest � Can impacts in the short-term on average improve longer-term outcomes � Unanswered � No questions clear explanation for better results at Kingsborough � Impact of ideal program vs. as implemented? 10
Performance-Based Scholarships
Louisiana Results Performance-Based Scholarships � Performance-Based Scholarships: � Need-based grants � Contingent on academic performance and/or student services � Paid directly to students � Paid in addition to other financial aid, such as Pell � Goals � Address unmet financial need and make college more affordable � Improve student success 12
Louisiana: Opening Doors Performance-Based Scholarships � Student participants � Most were women who were single parents � $1, 000/semester if enrolled half-time and earned a C average � Control group � Less than 40% registered in 2 nd and 3 rd semesters � After 3 semesters, average of 7. 7 earned credits � Impacts � Positively affected persistence in 2 nd and 3 rd semesters � Earned more credits (3. 3 credits) 13
Performance-Based Scholarship Programs Parents: Lorain County, Owens, and Sinclair Community Colleges (OH) High school seniors: California Cash for College (CA) Hispanic males: Pima Community College (AZ) 14 Adult learners: Borough of Manhattan and Hostos Communit Colleges (NY) Traditional college students: University of New Mexico (NM) Developmental math sequence: Hillsborough Community College (FL) Parents: Delgado Community College and Louisiana Technical College (Opening Doors Demonstration)
Early Results Performance-Based Scholarships 30 Program Group Control Group 0. 9 25. 7 24. 8 25 20 1. 7 ** 15 16. 1 3. 3 *** 10 5 14. 3 1. 0 14. 9 0. 9 * 16. 3 13. 9 15. 5 1. 7 *** 15. 6 13. 9 11. 0 7. 7 0 Opening Doors Louisiana 15 Arizona Florida New Mexico New York 15 Ohio
Taking Stock: Performance-Based Scholarships � Programs � Flexibility � Modest adhered to core principles to tailor programs to specific student needs but consistent impacts in the longer-term � Including on degrees, and in some cases loans � No effect on persistence � Little detectable variation in impacts � Implications � Impacts and unanswered questions on policy � Scholarship size � Services vs. no services 16
Comprehensive Reform Accelerated Study in Associate Programs
Accelerated Study in Associate Programs (ASAP) Student Responsibility • Enroll Full-time • Take Dev. Ed. Early • Graduate in 3 Years Financial Supports • Tuition Waiver • Metro. Cards • Textbooks Student Services • Advising • Tutoring • Career Services Course Enrollment • ASAP Seminar • Block-Schedule • Early Registration 18
Accelerated Study in Associates Program ASAP � 3 -year Results Program Group Control Group Difference 40% 22% 18 Graduation rates nearly doubled 19
ASAP: Next Steps � ASAP Replication � Replication of ASAP in 3 Ohio colleges � MDRC & CUNY exploring opportunities to bring ASAP to other communities � Long-term follow-up 20
Center for the Analysis of Postsecondary Readiness � MDRC, CCRC, and additional partners � Descriptive � Large study of developmental education national survey and qualitative interviews � Assessment � RCT of new algorithms in colleges across New York � Instruction � RCT study of the New Mathways Project in colleges across Texas 21
Contact Information Alex Mayer Research Associate Alexander. Mayer@mdrc. org Reports available at www. mdrc. org 22
- Advantage of randomized controlled trial
- Joseph stalin main accomplishments
- John cabot successes
- Successes of reconstruction
- Enrollment management models community colleges
- Community colleges
- Community education and training colleges
- Association of canadian community colleges
- Ccccio conference
- Washington state board for community and technical colleges
- California community college chief instructional officers
- Guided pathways california community colleges
- Block design vs matched pairs
- Completely randomized design
- Duncan's multiple range test
- Rcbd meaning in research
- Randomized algorithms and probabilistic analysis
- Rcbd adalah
- Anova table
- Expected running time
- Randomized skip list
- Advantages of randomized complete block design
- Completely randomized design (crd)