ERIC Number: ED563776
Record Type: Non-Journal
Publication Date: 2015
Abstractor: As Provided
Reference Count: 60
Statistical Properties of School Value-Added Scores Based on Assessments of College Readiness. ACT Research Report Series. 2015 (5)
The 2001 reauthorization of the Elementary and Secondary Education Act (ESEA) known as No Child Left Behind and more recent federal initiatives such as Race to the Top and the ESEA flexibility waiver have brought student growth to the forefront of education reform for assessing school and teacher effectiveness. This study examined growth projections and school value-added scores that are based on the ACT college and career readiness system for different growth periods (grades 8-10, 9-10, 10-11, 10-12, and 8-12). For this report, I investigated methodological questions related to value-added modeling. The analyses are based on large samples of students who took two or more of the tests in the ACT college and career readiness system in grades 8, 9, 10, and 11/12, respectively. For each growth period, I examined high schools' effects on student achievement in English, mathematics, reading, and science by controlling for prior test scores and varying time between tests. Across the various growth periods, I found differences in the variance of school effects (value-added scores), as well as differences among types of schools in the proportion of their students classified as significantly above or below average. Most school effects were not significantly different from the average school effect and could not usually be distinguished from average with confidence. This was more often the case for small schools than for large schools. I found evidence of consistency in value-added scores by examining same-year and same-cohort correlations for different growth periods and by examining correlations of value-added scores for adjacent cohorts (one, two, and three years apart) at the same schools. Value-added scores based on grades 8-12 data show greater reliability over time. The grades 8-12 data also generated value-added scores that perform better than scores based on shorter growth periods in terms of differentiating schools and achieving statistical significance with smaller sample size. I also found positive associations between value-added scores and mean prior academic achievement and an inverse relationship with school poverty level and the proportion of racial/ethnic minority students. Generally, compared to other school characteristics, grade level enrollment and the proportion of students tested had weaker associations with value-added measures. Overall, the relationships of school characteristics and value-added scores suggest that different types of schools would be expected to have different mean value-added scores. The following are appended: (1) Tables and figures; and (2) A description of the statistical procedure used.
Descriptors: Value Added Models, College Readiness, Statistics, Grade 8, Grade 9, Grade 10, Grade 11, Grade 12, School Effectiveness, High Schools, Academic Achievement, Institutional Characteristics, Correlation, Reliability, Scores, Minority Group Students, Poverty, Career Readiness, High School Students
ACT, Inc. 500 ACT Drive, P.O. Box 168, Iowa City, IA 52243-0168. Tel: 319-337-1270; Web site: http://www.act.org
Publication Type: Reports - Research
Education Level: Grade 8; Junior High Schools; Middle Schools; Elementary Education; Secondary Education; Grade 9; High Schools; Grade 10; Grade 11; Grade 12
Authoring Institution: ACT, Inc.