The Short-Term Effects of the Kalamazoo Promise Scholarship

on Student Outcomes

Timothy J. Bartik and Marta LachowskaW.E. Upjohn Institute for Employment Research

October 11, 2013

Presentation at Lumina Foundation, Indianapolis, IN

Presentation briefly summarizes full paper, available at

http://research.upjohn.org/up_workingpapers/186/

Paper is forthcoming in Research in Labor Economics

Main Findings

• The Kalamazoo Promise college scholarship program significantly improves high school students’ behavior, based on comparing changes for eligible vs. ineligible students, before and after the Promise’s announcement.

• Among African-American high school students, the Promise also significantly improves GPA.

• Policy implication: Promise-style programs can improve K-12 academics.

• These high school comparisons of eligible vs. ineligible students do not reflect possible Promise benefits for school climate or post-secondary success.

Our research approach

• Exploited aspect of KP that is natural experiment.

• In Nov. 2005, some KPS high school students discovered they were eligible for KP, others ineligible, based on prior enrollment decisions.

• Use individual HS student data on eligible/ineligible students, from 2 yrs. before (2003-04) to 2 yrs. after (2007-08) KP announcement.

• Do behavior/academics show post-announcement trends for eligible vs. ineligible students that differ from pre-announcement trends?

• Dependent variables: days suspended, GPA, etc. Controls: year, grade, student fixed effects, student socioeconomics.

• Eligible/ineligible differentials do not reflect overall school climate.

KP effects on annual days in suspension, all students

-3-2

-10

1D

iff-i

n-D

iff P

oint

Est

ima

te

2003-04 2004-05 2005-06 2006-07 2007-08School Year

Note: dots indicate statistical significance at 10 percent level, p<0.10

Estimated effect of KP on days in suspension

KP effects on GPA, all students

-.2

0.2

.4.6

Diff

-in-

Diff

Poi

nt E

stim

ate

2003-04 2004-05 2005-06 2006-07 2007-08School Year

Note: dots indicate statistical significance at 10 percent level, p<0.10

Estimated effect of KP on GPA

KP effects on annual days in suspension, African-American students

-6-4

-20

2D

iff-in

-Diff

Poi

nt E

stim

ate

2003-04 2004-05 2005-06 2006-07 2007-08School Year

Note: dots indicate statistical significance at 10 percent level, p<0.10

Estimated effect of KP on days in suspension

KP effects on GPA, African-American students

0.5

11.

5D

iff-in

-Diff

Poi

nt E

stim

ate

2003-04 2004-05 2005-06 2006-07 2007-08School Year

Note: dots indicate statistical significance at 10 percent level, p<0.10

Estimated effect of KP on GPA

Summary

• Strong evidence that generous & broad eligibility

Promise programs can improve the behavior of high

school students, and GPA for at least some groups.

• Extension is to look at how this natural experiment

affects post-secondary outcomes.

• Other extensions: look at WHY these effects occur.

• Other techniques needed to incorporate

spillover/school climate effects on student

outcomes, which could increase benefits.