Project Activities
This study proposed to analyze extant data from Washington State public schools and other sources of public data on youth after high school to examine three malleable factors for students receiving special education services: (1) enrollment in vocational education or workforce skills courses; (2) percent of the school day spent in general education classrooms; and (3) credentials and estimated performance of the students' high school teachers. These factors were considered as predictors of the academic, behavioral, transition, and postsecondary success of students with disabilities.
Structured Abstract
Setting
Secondary data were obtained from public high schools in Washington State, and later postsecondary outcomes of those students were obtained from U.S. 2-year and 4-year colleges, Washington State workplaces, and Washington State penitentiaries.
Sample
Data in the primary analytic cohort were obtained from a state census of 76,260 students who were enrolled in 10th grade in Washington State public high schools during the 2009-2010 school year. Students in this cohort-including 8,209 students who received special education services-were followed for 5 years. Three secondary school observations in 10th-12th grade and two postsecondary observations in the first 2 years after the expected graduation date are included in these data for each student. Additional analyses will be conducted with the 36,004 students currently receiving special education services in Washington State high schools.
Due to the nature of this research, there is no intervention.
Research design and methods
Secondary datasets will be explored to identify malleable factors (e.g., course-taking patterns and value-added measures of teacher performance) and intermediate student outcomes (e.g., absences, test scores, persistence, graduation). Planned analytic methods include a comparison of means; ordinary least squares (OLS), logit, and multinomial logit regressions; and two-stage least squares instrumental variables (IV) models. The nested nature of the data (e.g., within schools and over time) will be considered in various analytic models. Analyses will address the following questions about students in Washington State: (1) What malleable factors at the high school level (i.e., vocational education, general education, and teacher credentials and performance) are predictive of intermediate outcomes (i.e., unexcused absences, test performance, and persistence/graduation) for students with disabilities?; (2) What malleable factors at the high school level are predictive of postsecondary success (i.e., college enrollment, employment, and non-incarceration) of students with disabilities?; and (3) Which intermediate outcomes may mediate the relationship between the malleable factors and postsecondary outcomes?
Control condition
There is no control condition with the exception of analyses examining vocational education or workforce skills courses. The control condition in this case is not being enrolled in vocational education or workforce skills courses. The other malleable factors have a range of levels to consider.
Key measures
The study will consider four student outcome areas based on the variables included in the secondary datasets: (1) academic (scores on 10th-grade math and reading High School Proficiency Exams); (2) behavioral (number of unexcused absences in 11th grade and 12th grade); (3) transition (graduation with a high school diploma); and, (4) postsecondary (enrollment and retention in a 2-year or 4-year college, employment and employment type in the state workforce, and non-entry and/or recidivism in the state penitentiary system).
Data analytic strategy
All analytic models—including OLS, logit, multinomial logit, and two-stage least squares (IV) models—will assess key outcomes listed above as a function of malleable factors the previous year, other observable student characteristics (including base year test scores), and school/district effects (i.e., variations in vocational education, time in general education, and teacher credentials and performance). The intermediate outcomes may also be mediators for subsequent long-term outcomes. Given this, to identify plausible mediation pathways, the long-term outcomes will be estimated both with and without the intermediate outcomes as control variables. All models will be estimated with interaction terms to test for differential effects for students with different diagnosed disabilities and differences in the special education services provided (including variation in those services). An analysis of missing data and selection bias will also be conducted.
People and institutions involved
IES program contact(s)
Project contributors
Products and publications
ERIC Citations: Find available citations in ERIC for this award here.
Questions about this project?
To answer additional questions about this project or provide feedback, please contact the program officer.
