Co-Principal Investigator: Thomas White (University of Virginia) and Geoffrey Borman (University of Wisconsin)

Purpose: This project will evaluate the impact of Core Knowledge (CK) charter elementary schools in Colorado on student achievement primarily in kindergarten through third grade. It will also examine whether there are differential impacts by student and school characteristics, whether the level of implementation of the Core Knowledge program is associated with the impacts on student achievement, as well as the cost-effectiveness of the CK charter schools.

Project Activities: This evaluation will primarily be based on a randomized control trial using a lottery to assign students the opportunity to enroll in a CK charter school for kindergarten. Treatment and control students (the latter were not assigned to the treatment schools) will be followed to the end of third grade. In addition, a multivariate analysis of student achievement scores on state tests in grades 3 through 6 will be done to compare growth in school achievement at different types of schools: CK charter schools, non-CK charter schools, CK public schools, and non-CK public elementary schools.

Products: The products of this project will be evidence of the effectiveness of the CK charter schools in Colorado to increase student achievement as measured by standardized tests in language arts, math, science, and social studies. The evidence will be directly provided to the Colorado Department of Education through the participation of department personnel on the project and to the research community and other education practitioners and policymakers through publications.

Setting: The evaluation will take place across the state of Colorado.

Population: In Colorado, 59 charter schools have implemented the Core Knowledge curriculum. Of these, approximately 18 are expected to be oversubscribed and so take part in the evaluation. About 75% of the students in these schools are white and 15% receive free- or reduced-price meals. Approximately 1,300 students are expected to join the study in their pre-kindergarten year and are to be followed through the end of third grade.

Intervention: The intervention is a combination of charter schools (an organizational reform) with Core Knowledge (a curriculum reform). Approximately 5% of students in Colorado attend one of the state's 141 charter schools and of these 59 have adopted Core Knowledge, as have a small number of public schools. Core Knowledge provides a K–8 curriculum that identifies the topics to be taught in each grade for language arts, science, math, history, geography, visual arts, and music. These topics are to build knowledge, concepts and vocabulary from grade to grade. Working in teams, teachers use teacher handbooks (that include the topics, background knowledge, vocabulary, and instruction ideas) plus some professional development to develop their lesson plans and assessments. This project is operating under the hypothesis that charter schools will provide a better environment for implementing CK (and therefore a better test of its efficacy) because they can focus on the program without the potential competing district mandates and programs that public schools must address.

Research Design and Methods: The evaluation is based on a randomized control trial in which a minimum of 18 oversubscribed CK charters use a lottery to determine which students will enter their kindergarten classes. Treatment students (those who the lottery accepts) and control students (those who the lottery rejects) are followed through the end of third grade. In each grade, they take a varying set of standardized tests in language arts, math, science, and social studies. In addition, fidelity of implementation is measured in treatment schools using five measures and in control schools using one measure.

Control Condition: Students who were not chosen by lottery to attend a CK charter school may be enrolled in other type of school (private, other charter, public) depending on their families' decision.

Key Measures: A varying set of standardized assessments are given to the students in each grade. In pre-K, students take the Peabody Picture Vocabulary Test 4th Edition (PPVT-4) for oral vocabulary and two Woodcock-Johnson III math tests: Quantitative Concepts and Calculation. In kindergarten, students take the PPVT-4 and two Woodcock-Johnson reading tests: Sound Awareness and Letter-Word Identification. In grade 1, students take the PPVT-4, two Woodcock-Johnson reading tests: Word Attack and Passage Comprehension, and four Woodcock-Johnson math tests: Quantitative Concepts, Calculation, Applied Problems, and Math Fluency. In grade 2, students take the PPVT-4, a Core Knowledge Vocabulary Test, and two Woodcock-Johnson reading tests: Word Attack and Passage Comprehension. In grade 3, students take the PPVT-4, a Core Knowledge Vocabulary Test, the Iowa Test of Basic Skills in Science and Social Studies, three Woodcock-Johnson III tests of listening comprehension: Oral Comprehension, Story Recall, and Story Recall-Delayed, and the Colorado Student Assessment Program for reading, writing, and math.

Five measures are used to collect fidelity of implementation data. In the pre-K year, the School Observation Measure will be used in a maximum of 12 randomly-selected K–3 classrooms per school. Every year, teachers with treatment students will take the Core Knowledge Implementation/Program Coherence Survey and be observed during language arts, science and social studies using Foorman and Schatschnieder's procedures. Both treatment and control teachers will be surveyed using items for the Survey of Enacted Curriculum for English-Language Arts. In addition, treatment students in grades 1–3 will take tests covering the CK topics in language arts, math, history and geography, and science.

Data Analytic Strategy: The results from the randomized experiment will be analyzed using several models including school fixed effects Ordinary Least Squares (OLS) models, mediation models for reading comprehension, and two level students in schools random effects models. In addition, growth over time will be examined.

In addition, a statistical analysis using non-experimental data from schools across the state will examine the growth in student state assessment scores as they move from grade 3 to grade 6 for four cohorts of students. This analysis compares results for 59 CK charter schools, 50 non-CK charter schools, 18 CK public schools, and 824 non-CK public elementary schools. The data will be obtained from the Colorado Student Assessment Program (student test scores), Colorado Department of Education (student, teacher and school-level data), and the Common Core Data (district data). The analysis will use an OLS model with robust standard errors.

Products and Publications

Journal article, monograph, or newsletter

Altenhofen, S., Berends, M., and White, T.G. (2016). School Choice Decision Making Among Suburban, High–Income Parents. AERA Open, 2(1).