Getting Students to Graduate:Developmental Education

Thomas BaileyNational Center for Postsecondary Research

Community College Research CenterTeachers College, Columbia University

Presented at

Higher Education in Michigan: Looking Back and Looking Ahead

Ann Arbor, MichiganDecember 10, 2009

CCRC and NCPR Research on Developmental Education

• Community College Research Center– Assessment, completion, pathways and sequences,

outcomes, evaluation of models and initiatives– http://ccrc.tc.columbia.edu/

• National Center for Postsecondary Research– US Dept. of Education Funded Center– Partnership of CCRC, MDRC, UVA– Analysis of effectiveness using state data– Experimental evaluations of learning communities and

intensive summer bridge programs– http://www.tc.columbia.edu/centers/ncpr/

Developmental Education

• Huge barrier to increased completion

• Assessment and placement systems need fundamental reform

• Evidence suggests that current developmental practices do not work well

• Most students do not complete their developmental sequences

Incidence of Remediation

• 58 percent—at least one course (NELS)

• 44 percent—1 to 3 courses (NELS)

• 14 percent—more than 3 courses (NELS)

• 59 percent—at least one course (ATD)

Referrals to Levels of Dev. Ed.

Math - Full Sample

3 levels below19%

2 levels below16%

1 level below24%

Not referred41%

Reading - Full Sample

Not referred67%

3 levels below3%

2 levels below7%

1 level below23%

0.2

.4.6

.81

-50 -25 0 25 50CPT Score Relative to Math Cutoff

OutcomeE[Y|Z=z]

Outcomes for remedial students

Outcomes for non -remedial students

Local treatment effect

00.

10.

20.

30.

40.

5

-50 -25 0 25 50Assessment Score Relative to Cutoff

StudentOutcome Outcomes for

college ready students

Local treatment effect

Outcomes for remedial students

Completion of First College-Level Course and Retention by CPT Score and Subject

0.2

.4.6

.81

-50 -40 -30 -20 -10 0 10 20 30CPT Score Relative to Math Cutoff

Estimated Discontinuity (Math ITT Table 4) = -0.061(0.013)Outcome: Completion of First College-Level Course

0.2

.4.6

.81

-50 -40 -30 -20 -10 0 10 20 30CPT Score Relative to Math Cutoff

Estimated Discontinuity (Math ITT Table 4) = -0.021(0.011)Outcome: Fall-to-Fall Retention

0.2

.4.6

.81

-50 -40 -30 -20 -10 0 10 20 30CPT Score Relative to Reading Cutoff

Estimated Discontinuity (Reading ITT Table 4) = -0.068(0.008)

0.2

.4.6

.81

-50 -40 -30 -20 -10 0 10 20 30CPT Score Relative to Reading Cutoff

Estimated Discontinuity (Reading ITT Table 4) = -0.009(0.008)

0.2

.4.6

-50 -40 -30 -20 -10 0 10 20 30CPT Score Relative to Math Cutoff

Estimated Discontinuity (Math ITT Table 5) = -0.019(0.008)

Outcome: Transfer to SUS

0.2

.4.6

-50 -40 -30 -20 -10 0 10 20 30CPT Score Relative to Math Cutoff

Estimated Discontinuity (Math ITT Table 5) = -0.030(0.008)

Outcome: Degree Completion

0.2

.4.6

-50 -40 -30 -20 -10 0 10 20 30CPT Score Relative to Reading Cutoff

Estimated Discontinuity (Reading ITT Table 5) = -0.019(0.004)

0.2

.4.6

-50 -40 -30 -20 -10 0 10 20 30CPT Score Relative to Reading Cutoff

Estimated Discontinuity (Reading ITT Table 5) = -0.029(0.004)

Transfer to State University System and Degree Completion by CPT Score and Subject

What Does This Say About Assessment?

• No obvious cutoff point—in practice many different assessments and standards

• Confusion about what it means to be “college ready”

• Assessments do not provide much diagnostic information

Implications for Effectiveness

• Current system neglects the academic needs of weaker “college level” students

• Distinction between college level and developmental students is arbitrary and doesn’t serve either group

• We know little about the effectiveness of services for students far from the cutoff scores

• We do know that few referred to multiple levels of remediation finish

Achieving the Dream Database

• 250,000 students

• All first time (in the college) degree seeking students (full or part time)

• 57 colleges in CT, FL, NC, NM, OH, PA, TX, VA, WA

• Not representative of all CCs—similar to large, urban institutions with lower funding per student

Not enrolled

18%

3 levels below

Not Completed

25%

Not enrolled 16%

2 levels below

Not completed

12%

Completed

16%

Enroll

Pass

Referred to Lev. 3 46824

Not enrolled

7%

1 level below

Enroll

Pass

Not completed

6%

Enroll

In-Order Course Completion and Enrollment for Math Remediation

82%

57%

41%

29%

22%

Not enrolled 31%

3 levels below

Not completed

13%

Not enrolled 17%

2 levels below

Not completed

4%

Completed 22%

Enroll

Pass

Referred to Lev. 3 7987

Not enrolled 10%

1 level below

Enroll

Pass

Not completed

4%

Enroll

In-Order Course Completion and Enrollment for Reading Remediation

39%

56%

69%

35%

25%

Enrollment and Progression Patterns Among Achieving the Dream Students

Math

42%

40%

38%

40%

42%

31%

18%

29%

16%

29%

44%

33%

0% 20% 40% 60% 80% 100%

3 levels below

2 levels below

1 level below

Total

Ref

erre

d to

Not Enrolled Not Passed Completed

Student Progression by Enrollment and Gatekeeper in Math

Math - Full Sample

18%

26%

38%

28%

24%

13%

11%

42%

32%

18%

31%

5%

10%

17%

11%

2%

4%

6%

4%

8%

15%

21%

16%

0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%

3 levels below

2 levels below

1 level below

Total

Ref

erre

d to

Never enrolled Not re-enrolled Not completed dev GK Not enrolled GK Not passed GK Passed

E3, no C3,exit

R=3

3

2

1

GK

exit withoutever enrolling

C3, exit

C3, E2, no C2, exit

C3, C2,exit

C3, C2, E1,no C1, exit

C3, C2, C1, exit

C3, C2, C1E GK, exit

No E3, skip to 2

No E3, skip to 1

No E3, skip to GK

E2, no C2,skip to 1

E1, no C1,skip to GK

C2, skip to GK

E2, no C2, skip to GK

E3, no C3, E2

E3, no C3, E GK

E3, no C3, E1C3, skip to 1

C3, skip to GK

1

1

1

1

1

1

1

1

C GK1

18

6

2

6

2

2

18

6 6

2

2

2

Research Partnerships

• Connecticut

• Washington (ABE)

• Florida

• CUNY

• Virginia

Enrollment in Developmental Courses• However, many students did not take a

recommended developmental course

55% 60%51%

4%5%

10%

41% 35% 39%

0%10%20%30%40%50%60%70%80%90%

100%

Reading Writing Math

Dev

elo

pm

enta

l E

nro

llm

ent

Took Same Took Another Took None

Among students recommended to a developmental course . . .

Performance in Gatekeeper Courses• Once enrolled, gatekeeper course passing rates also

did not vary strongly by developmental recommendation compliance

76% 75% 78%75%74% 72%74%68%

74%

0%10%20%30%40%50%60%70%80%90%

100%

Reading Writing Math

Gat

ekee

per

Pas

s (%

)

College-Level Dev - Took Dev - Skipped

http://www.vccs.edu/Portals/0/ContentAreas/AcademicServices/DevelopmentalEducationTaskForce_200909.pdf

Policy Implications

• Assessment and placement is a mess

• No consensus on college readiness

• Profound pipeline problems—many opportunities to leave

• Data have been effective in influencing legislatures and state agencies

For more information:

Please visit us on the web at http://ccrc.tc.columbia.edu,where you can download presentations, reports,

CCRC Briefs, and sign up for news announcements.

CCRC funders include: Alfred P. Sloan Foundation, Lumina Foundation for Education, The Ford Foundation, National Science Foundation (NSF), Bill & Melinda Gates Foundation

and Institute of Education Sciences of the U.S. Department of Education

Community College Research CenterInstitute on Education and the Economy, Teachers College, Columbia University

525 West 120th Street, Box 174, New York, NY 10027 E-mail: ccrc@columbia.eduTelephone: 212.678.3091