Mason Burley ERDC SLDS Conference Olympia, WA May 21, 2014 Big Data, Big Obstacles Assessing the...

29
Mason Burley ERDC SLDS Conference Olympia, WA May 21, 2014 Big Data, Big Obstacles Assessing the Impact of Statewide College Financial Assistance Programs WASHINGTON STATE INSTITUTE FOR PUBLIC POLICY ?

Transcript of Mason Burley ERDC SLDS Conference Olympia, WA May 21, 2014 Big Data, Big Obstacles Assessing the...

Mason Burley

ERDC SLDS Conference

Olympia, WA

May 21, 2014

Big Data, Big ObstaclesAssessing the Impact of Statewide

College Financial Assistance Programs

WASHINGTON STATE INSTITUTE FOR PUBLIC POLICY

?

DebriefStrategyObstaclesMission

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

May 21, 2014 Slide 2 of 28

• Started in 1969

• Available to WA undergrads

• Available to public/private, two/four students

• Award amounts linked to tuition

• Need-based (median family income)

• Means tested (prorated award)

Subject: State Need Grant (SNG) program

DebriefStrategyObstaclesMission

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 3 of 28

Problem: Resident undergraduate tuition growth

SectorTuition and

fees (2000–01)

Tuition and fees

(2011–12)

Public research $3,984 $9,993

Public regional $3,092 $6,969

Private four-year $15,597 $31,618

Community and technical college $1,756 $3,542

May 21, 2014

DebriefStrategyObstaclesMission

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 4 of 28

State Need Grant—Total expenditures (2003-2012)

2003-04

2004-05

2005-06

2006-07

2007-08

2008-09

2009-10

2010-11

2011-12

2012-13

$0

$50

$100

$150

$200

$250

$300

$350

*inflation adjusted (2012) dollars

Dollars

(m

illion

s)

May 21, 2014

DebriefStrategyObstaclesMission

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 5 of 28

State Need Grant eligible students

2003-04

2004-05

2005-06

2006-07

2007-08

2008-09

2009-10

2010-11

2011-12

2012-13

0

20,000

40,000

60,000

80,000

100,000

120,0005

3,6

75

55

,27

8

65

,32

8

66

,36

4

70

,08

5

72

,51

1

70

,37

6

72

,33

8

74

,70

3

73

,98

5

2,6

65

5,4

78

3,5

83

1,8

80

1,6

01

5,4

98

21

,94

8

25

,67

7

30

,96

6

32

,44

3

SNG Eligible, did not receive grantReceived SNG (local funds)Received SNG

May 21, 2014

DebriefStrategyObstaclesMission

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 6 of 28

Mission: Are SNG dollars effective?

2012 HB 2127

“Determine to what extent this [SNG] program has increased access and degree attainment for low-income students and to determine whether the funding for the state need grant has been utilized in the most efficient way possible to maximize the enrollment and degree attainment of low-income students”

May 21, 2014

DebriefStrategyObstaclesMission

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 7 of 28

Mission: Are SNG dollars effective?

• The demographics of recipients, including, but not limited to, gender, race, and income

• The effect on enrollment rates of low-income students at the different institutions of higher education and whether these students attend full-time or part-time

• The effect on persistence, performance, degree completion, and time-to-degree completion

• An inventory of the types of degrees, by field of study

• The interplay of the State Need Grant program with other forms of financial aid

May 21, 2014

DebriefStrategyObstaclesMission

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 8 of 28

Operatives: Washington State Institute for Public PolicyNon-parti san research at legislati ve/board directi on

Board of DirectorsSenator Jeanne Kohl-Welles, Co-Chair

Representative Maureen Walsh,

Co-Chair

Senator Joe Fain Representative Ruth Kagi

Senator Karen Fraser Representative Larry Springer

Senator Mark Schoesler Representative Hans Zeiger

Richard Rodger, Senate Staff Dir.

Jill Reinmuth, House Staff Dir.

David Schumacher, OFM Governor Appointee

Sandra Archibald, UW Rodolfo Arévalo, EWU

Thomas “Les” Purce, TESC Robert Rosenman, WSUMay 21, 2014

ResultsStrategyMission Obstacles

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 9 of 28

Research obstacles: Different sources of aid

Aid category Research Regional Privatefour-year

Community/technical colleges

Total

Federal grants—Pell 23% 25% 14% 33% 28%State Need Grant 28% 25% 18% 15% 20%Institutional and other grants 16% 8% 29% 3% 9%Work study 2% 2% 2% 2% 2%Federal loans 18% 26% 21% 15% 18%Need unmet 14% 14% 16% 31% 23%

Need unmet average (dollars) $ 3,349 $ 2,914 $ 6,384 $ 4,921 $ 4,676 Average need per student $ 22,852 $ 19,982 $ 38,910 $ 15,655 $ 19,877

• Calculated need equals cost of attendance minus expected family contribution• Note that not all students receive award/loan in each category• Private loans not included

May 21, 2014

ResultsStrategyMission Obstacles

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 10 of 28

Research obstacle: Differences among studentsStudent background and profi le• 64% – Community and technical college students• 44% – Independent, over age 25 (37% dependent under age 25)• 84% – Income under half of state median family income• 16% – Independent and married

Enrollment level• 58% – Enrolled full-time, full year• 18% – Enrolled full-time, part year• 33% – Summer enrollment

May 21, 2014

ResultsStrategyMission Obstacles

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 11 of 28

Research obstacle: Comparison group

…may not be comparable• Higher family incomes• Late filers• More likely to enroll part-time

Unserved SNG students…

…don’t stay “unserved”• 40% receive SNG in following year

…have long term outcomes that are unknown• 2009-10 first year unserved population grew

May 21, 2014

ResultsStrategyMission Obstacles

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 12 of 28

Data challenge: Timing

Persistence: on-going enrollment• Academic year completion (fall to spring)• Fall-to-fall continuation

Academic progress• Accumulation of credits for on-track graduation

Graduati on • Degree or certificate completion within 6 years (3 years for associates)

• Electronic databases available from 2005 (fall) forward

• Final report due January 2014

Potential outcomes

May 21, 2014

ResultsStrategyMission Obstacles

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 13 of 28

Data challenge: Linking data

Unit Record Report (WSAC)

CTC enrollment(SBCTC)

Public four-year(PCHEES-COP)

Private four-year(12/15 colleges)

Private career(6/11 schools)

Education Research and Data

Center (ERDC)

WSIPP SNG

research data

Linked, de-identified

May 21, 2014

ResultsStrategyMission Obstacles

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 14 of 28

Data challenge: Standardized data

Comparability of public/private college data• Cooperation• Definitions• Institutional aid

IPEDS measures• Institutional crossover• First-time enrollment

Academic progress • Credit accumulation, satisfactory academic progress

May 21, 2014

ResultsMission Obstacles Strategy

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 15 of 28

• Difference in SNG amount, but similar profile/aid• Permits reliable inference on SNG effect

Approach: Take advantage of “discontinuities” in SNG design

Median family income range

Income limits (family of four)

Percent of State Need Grant award

0-50% $41,000 100%51-65% $53,000 75%66-70% $57,500 50%71%+ ineligible

May 21, 2014

ResultsMission Obstacles Strategy

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 16 of 28

0 10 20 30 40 50 60 70 $-

$1,000

$2,000

$3,000

$4,000

$5,000

$6,000

$7,000

$8,000

Avg SNG Award

Avg Pell Awd

Median Family Income (MFI) Level

Aver

age

Stat

e N

eed

Gra

nt A

mou

nt

Avg Pell Awd

Average State Need Grant and Pell by MFI

If SNG has an effect, we would expect to see it here

May 21, 2014

ResultsMission Obstacles Strategy

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 17 of 28

Financial aid as a percentage of COA by MFI range and sector: 2011-12

0-50

MFI

51-7

0 M

FI

0-50

MFI

51-7

0 M

FI

0-50

MFI

51-7

0 M

FI

0-50

MFI

51-7

0 M

FI

35%23% 29%

19% 19% 12% 18% 13%

20%

9%

23%

10% 12%5%

31%

16%

17%

22%

23%

28% 21%

22%

15%

14%

6%

13%

5%

7% 23%31%

2%

4%

State Need Grant Pell Grant Loans Institutional Aid Other AidRemaining Need EFC

Research Regional Private bacc. CTC

May 21, 2014

ResultsMission Obstacles Strategy

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 18 of 28

Know thy data….

Acronym soup• COA, EFC, SAP, Need, CIP….

Sector-specifi c defi niti ons• Student classification• Outcome milestones

Verifi cati on and transparency

May 21, 2014

Mission StrategyObstacles Results

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 19 of 28

Effects* of a 25% reduction SNG on outcomes

CTC students Public Baccalaureate students-9%-8%-7%-6%-5%-4%-3%-2%-1%0%

-3.8

% -2.2

%

-4.6

%

-8.3

%

Persistence (enrolled next fall)

Degree completion

*Based on median coefficient estimates derived from 12 different models testing four dis-continuity bandwidths (2-5%) and three different specifications

May 21, 2014

Mission StrategyObstacles Results

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 20 of 28

• 4% change in expected fall-to-fall enrollment for CTC students

1,791 students (of total 47,134 enrolled) may change enrollment decision based on $600 change in SNG amount

Fall-to-fall enrollment

A 25% change in the SNG award is associated with….

• 2% change in expected fall-to-fall enrollment for public baccalaureate students

378 students (of total 17,171 enrolled) may change enrollment decision based on $2,000 change in SNG amount

May 21, 2014

Mission StrategyObstacles Results

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 21 of 28

• 5% change in expected degree completions for CTC students

Total degrees earned may change for 436 entering students (of 10,895 first-time students) with a $600 change in SNG amount

Degree completion

A 25% change in the SNG award is associated with….

• 8% change in expected degree completions for public baccalaureate students

Total degrees earned may change for 439 entering students (of 5,289 first-time students) with a $2,000 change in SNG amount

May 21, 2014

Mission StrategyObstacles Results

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 22 of 28

Finding 1: Effect of need-based grants on subsequent fall enrollment (elasticity)

Goldrick-Rab et al, 2012

(WI)

Bania et al., 2013 (WA)

Castleman & Long, 2013

(FL)

Bettinger, 2010 (OH)

0.00

0.05

0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.098 0.109

0.162 0.172

Ela

sti

cit

y

Weighted average = 0.140

May 21, 2014

Mission StrategyObstacles Results

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 23 of 28

Finding 2: Longitudinal (cross sector) outcomes

CTC Earned award No award

Type of student

Transfer to four-

yearNo

transferTransfer to four-

year

Preparedtransfer ready

Still enrolled

Not enrolled Total

Academic 329 (15%)

313 (14%)

166 (7%)

64 (3%)

374 (17%)

998 (44%) 2,250

Workforce 71(4%)

524 (29%)

46(3%)

408 (22%)

127(7%)

651 (36%) 1,826

May 21, 2014

Mission StrategyObstacles Results

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 24 of 28

Finding 2: Longitudinal (cross sector) outcomes

Baccalaureate

SectorCompletion at starting institution

Complete at different

institutionStill

enrolledNot

enrolled TotalFour-year

college

Two-year

college

Public research 765 (64%) 23 (1%) 33 (2%) 69 (5%) 289 (24%) 1,179

Public regional 445 (43%) 48 (4%) 53 (5%) 92 (8%) 394 (38%) 1,032

Private four-year 239 (56%) 20 (4%) 11 (2%) 23 (5%) 127 (30%) 420

May 21, 2014

Mission StrategyObstacles Results

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 25 of 28

Finding 3: Variation in “need”

Pell Grant receipt among SNG students by income level (2011-12)

MFI range Received Pell Pell ineligible Total students with SNG

0-50% 61,255 (99%) 692 (1%) 61,947

51-60% 4,440 (77%) 1,339 (23%) 5,779

61-70% 2,459 (65%) 1,345 (35%) 3,804

Total 68,154 (95%) 3,376 (5%) 71,530

May 21, 2014

Mission StrategyObstacles Results

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 26 of 28

Alternative state awarding formulas

State Recipients(2009-10)

Expenditures (millions) Model

WA 70,376 $223 Shared responsibility (MFI)

MN 103,544 $168 Shared responsibility (Pell)

NC (UNC) 61,952 $133 Shared responsibility (Pell, tax)

OR 43,111 $77 Shared responsibility (Pell, tax)

PA 171,702 $413 Percentage of need

• Shared responsibility modelaward amount=COA-(student+family+govt Pell)

• Percentage of need declines as EFC increases

May 21, 2014

Mission StrategyObstacles Results

Washington State Institute for Public Policy

http://www.wsipp.wa.gov

Slide 27 of 28

Finding 3: Variation in “need”

Hypothetical state examples—students served and award amounts

State formula

Estimated students served

Estimated students unserved

Estimated average grant Maximum grant

WA 72,060* 28,450* $3,460 $10,868

MN 81,319 15,549 $3,143 $10,450

NC 73,474 - $2,514 $3,800

OR 96,653 - $1,382 $1,950#

PA 98,257 - $1,883 $3,700

May 21, 2014

LESSONS LEARNED

http://www.wsipp.wa.gov

Slide 28 of 28

• Statewide programs do not always include statewide information

• Within/between differences

• An integrated database can be used to test eligibility and awarding formulas

• Right team and structure necessary for program evaluation – THANKS to ERDC, WSAC, and institutional partners

May 21, 2014

THANK YOU

http://www.wsipp.wa.gov

Questions?

Mason Burley

[email protected]

360-528-1645