The%20Impact%20of%20the%20Promise%20of%20Scholarships%20and%20Altering%20School%20Structure%20on%20College%20Plans,%20Preparation,%20and%20Enrollment

1
The Impact of the Promise of Scholarships and
Altering School Structure on College Plans,
Preparation, and Enrollment 

• By
• Jerald R. Herting, Charles Hirschman and Nikolas
  Pharris-Ciurej
• Department of Sociology and Center for Studies in
  Demography and Ecology,
• University of Washington

Thanks to the Andrew W. Mellon and the Bill and
Melinda Gates Foundations and to the Tacoma
Public Schools, their staff and students
2
Washington State Achievers Program

• Increase college attendance rates through various
  means (e.g. scholarship programs, school
  redesign, and mentoring programs) at selected
  high schools
• Reduce financial barriers for low income students
  that are motivated to attend college
• Schools targeted were historically lower
  achievers, lower SES population, and higher
  population of minority youth

3
Program altered

• Motivation
• Preparation
• Ability
• Mechanisms
• Scholarships
• Mentor
• Restructuring schools

4

• Research Question
• We are interested in assessing the effect of the
  WSA program on college plans, actual preparation,
  and college attendance.

5
Evaluating the Achievers ProgramA Serendipitous
Natural Experiment

• Two data sources available
• UW Beyond High School Project
• Prior to Implementation 2000 Senior Survey
• Post Implementation 2002 to 2005 Senior Surveys
• 3 Program High Schools and 2 Non-program High
  Schools.
• School records from large metropolitan school
  district from mid 1990s to present
• Prior School enrollment data for the class of
  2000
• Post Enrollment data for the classes of 2001 to
  2005
• 3 Program Schools and 2 Non-program schools

6
The Natural Experiment

• Given lack of true experimental design the
  natural experiment allows a pre- vs post-program
  evaluation
• Our expectations are that relative to
  pre-program, post-program educational
  attitudes/behaviors should increase
• And that we should observe this increase in
  program schools but not in non-program schools
• This observed change vs non-change provides
  evidence of a WSA effect
• This observed change should narrow the gap
  between the WSA schools and the non-WSA schools

7
The comparisons we would like to make have
problems

• Comparisons of WSA schools to non-WSA schools
• Schools were selected on specific criteria
• Schools have different input population
• Direct comparisons of WSA scholarship recipients
  to non-recipients poses problems
• Without a strict experimental design comparisons
  are problematic
• Youth are not necessarily comparable (i.e. not
  all differences among them have been nullified by
  randomization)
• Selection bias and endogeneity

8

• Given these limitations our best comparison is
  youth in the same school pre-program to youth in
  the same school post-program
• We expect change in post-program WSA schools that
  narrows the gap in educational oriented behaviors
  between the WSA and non-WSA schools
• We might expect differences in WSA effects across
  settings

9
Pattern of Change Expected
10
Data Sources
1. UW Beyond High School Project

• Focus on race/ethnic disparities in transition to
  college
• 2000, 2002, 2003, 2004 and 2005 Senior Surveys
• 5 high schools in metropolitan school district
• Paper and pencil survey modeled on NELS
• About 70-80 coverage of all potential high
  school seniors
• High school seniors an elusive concept and
  population
• In school survey and multiple follow ups
• Only 2 percent refusal rate
• One year follow up survey 90 response rate
• 4,300 (approx) seniors in merged file 2000, 2002
  to 2005

11
Description of Data Sources (cont.)
2. School records from large metropolitan school
district in Puget Sound area

• School district in which UWBHS survey was
  conducted
• School records for each semester from 1994 to
  2005
• Can track individual students, courses, grades,
  credits earned, and school attended
• Trace the progress of 4 cohorts of 9th graders
  graduating classes of 2000 to 2005
• Roughly 9,000 students in the analysis

12
Outcomes of Interest

• College Plans and Preparation
• Plan to attend 4-year college?
• Have taken the SAT/ACT?
• College Enrollment -- one year later
• Enrolled in any college 2 year or 4 year
• Enrolled in a four year college or university

13
Pattern of Change Expected
14
Control variables

• Race/Ethnicity
• Gender
• Generational Status (Senior Survey only)
• Social Class (Parental Education, Home ownership
  Family Income above/below poverty level)
• Family Structure (Senior Survey only)

15
Findings

• The social experiment to increase the transition
  from high school to college is effecting
  students behaviors consistently in 2 of 3 WSA
  high schools.
• An observed effect on 4 year college plans
• Strong effect by 2004 and 2005
• An observed effect on SAT/ACT
• Strong effect by 2004 and 2005
• An observed effect on attending 4 year college
• Consistent impact for HS 1 for 2003, 2004, and
  2005
• Consistent impact of HS 3 by 2004, 2005

16
(No Transcript)
17
(No Transcript)
18
Implications

• WSA initiatives appear to have success on key
  outcomes of plans, preparation, and college
  attendance
• Since initiatives are implemented in different
  settings with different forms of
  implementation--not all strategies and settings
  were equal in outcomes
• One setting failed to show significant effects
• Expectation of delayed effects

19
Pattern of Change Expected
20
Thanks and Comments
21
(No Transcript)
22
(No Transcript)
23
Effects of WSA on Educational Outcomes
24
Source Stoops 2004 2
25
Underlying Analytical Design

• Y B0 B1 (WSA) B2 (Year 2) B3
  (Year 3) B4 (Year 4) B5 (Year 5) B6(Year 2
  WSA) B7 (Year 3 WSA) B8 (Year 4 WSA)
  B9 (Year 5 WSA) Bn (Xn)
• This model is applied separately for each WSA
  program school

26
Interpreting Results

• We want to see a significant value greater than 0
  for B4 (Year 1 WSA), B5 (Year 2 WSA) and B5
  (Year 3 WSA) which implies that being in a WSA
  school post-program implementation has increased
  the probability of these behaviors compared to
  pre-program