Co-Principal Investigator: Laura Giuliano (University of Miami)

Purpose: The goal of this project is to evaluate the impact of gifted education on student achievement in the Broward County Public Schools (BCPS) in Florida. The BCPS gifted program is district-wide and characterized by high-intensity models with separate gifted classrooms, specialized teacher training, and accelerated and differentiated instruction. The research team will also evaluate a BCPS universal screening program aimed at the under-representation of disadvantaged groups in the gifted program. Applying quasi-experimental methods to secondary data, researchers will answer three main questions: (1) What is the effect of gifted education in early elementary grades on subsequent academic achievement? (2) What is the impact of universal screening to identify potentially gifted children on the size and composition of the gifted population? (3) To the extent that universal screening increases the size of the gifted population, how do the program impacts for newly identified students compare to impacts for students identified though traditional channels?

Project Activities: The research team will carry out several quasi-experiments using student longitudinal data from BCPS as well as school longitudinal data from both BCPS and other Florida school districts. To address the impact of gifted education, researchers will use a regression discontinuity design based on BCPS second graders' IQ scores (used to assign students to gifted education) and universal screening scores during the 1998–99 to 2009–10 school years. Student outcomes in later grades (such as test scores, attendance, and disciplinary referrals) will be compared for students who scored above and below the cut points upon which assignment was based. A fidelity of implementation study will be done at all 150 BCPS elementary schools to identify the type of gifted program implemented each year. To address the impact of universal screening, two comparative interrupted time series will be done. Fidelity of implementation studies will examine the extent to which universal screening occurred and to which students who were identified as gifted by the screening actually entered gifted programs, and whether universal screening affected the timeliness of student placement into gifted programs.

Products: Products include evidence of the effectiveness of the BCPS gifted education program for improving students academic outcomes and the effectiveness of the universal screening for gifted students in increasing participation of students from groups who have traditionally had low participation in gifted education. The evidence will be presented to the research community and other education practitioners and policymakers through presentations and peer-reviewed publications.

Structured Abstract

Setting: This study will take place in Broward County Public Schools (BCPS) in Florida. BCPS is the 6th largest U.S. school district with about 260,000 students. Elementary students are 26 percent Hispanic, 37 percent Black non-Hispanic, 30 percent white non-Hispanic, and 7 percent other groups. About half of BCPS students receive free or reduced price lunches.

Population: The population for this study includes all students in a BCPS public school between 1998–99 and 2009–10. The sample analyzed will include students enrolled in a BCPS elementary school in second grade between 1998–99 to 2008–09 who were evaluated for gifted education before third grade and who have state standardized test scores for at least 3rd through 5th grades.

Intervention: The BCPS gifted program is district-wide and characterized by high-intensity models with separate gifted classrooms, specialized teacher training, and accelerated and differentiated instruction. For grades K–5, schools may choose from one of four gifted models depending on the number of gifted students they have. Schools with 20 or more gifted students per grade or grade combination are to have full-time, self-contained, gifted-only classes. Schools with less than 20 gifted students per grade or grade combination may either: (1) have full-time, self-contained classes of gifted and high-achieving students (but at least half the students are to be gifted), or (2) hold half-time (2.5 hours per day) gifted-only classes in the core content. For grades K–3, if there are less than 11 gifted students per grade or grade combination, these gifted students may be taught a differentiated curriculum within a regular classroom by a gifted-endorsed teacher. Program admission originally required: (1) a referral for evaluation by a BCPS psychologist or private IQ testing submitted for review by a BCPS psychologist, and (2) an IQ score above a fixed cutoff. Students who achieve an IQ test score of 130 or more are eligible for gifted educations. A lower threshold of 115 is set for students with limited English proficiency and students who receive a free or reduced price lunch. In 2005, universal screening was introduced so that all 2nd grade students who exceed a threshold score on the screening test (the Naglieri Nonverbal Ability Test) are automatically referred for an evaluation and a full IQ test.

Research Design and Methods: To assess the impact of gifted participation on academic achievement, researchers will use regression discontinuity designs based on the IQ cutoff for gifted eligibility and the screening test cutoff used to generate referrals for gifted evaluation. The data will be examined to determine if IQ scores were manipulated or whether there is a selection bias due to multiple IQ testing and additional analyses will be done to address the potential for such bias. To assess the impact of universal screening on overall and subgroup participation, the project will use two comparative interrupted time series designs. One will compare trends in BCPS schools to trends in Florida districts that do not use universal screening. The other will use only schools within BCPS but will provide additional evidence on the differential effects across subgroups. Fidelity analyses will be done using district administrative data, school records, and gifted teacher interviews to determine the type of gifted programs used at each school in each year. These analyses will also determine the degree to which universal screening was done and led to students who were screened as gifted to be placed in a gifted program.

Control Condition: Students who are not in a gifted program will receive the general educational program. Instead of universal screening, most Florida districts rely on teachers and parents to make referrals for gifted evaluation. BCPS followed this same approach before it began universal screening in 2005.

Key Measures: Scores on Florida's state standardized tests will be used as the key outcome measures. These include the Stanford Achievement Test (SAT-9 or SAT-10) and the Florida Comprehensive Achievement Test (FCAT-SSS) in Math, Reading, Writing and Science. IQ test scores are the key assignment measure and the Naglieri Nonverbal Ability Test will be used for universal screening.

Data Analytic Strategy: The regression discontinuity design will be examined using both parametric and non-parametric approaches. The study will test for manipulation of IQ score manipulation. If found, several alternative analysis will also be done (e.g., using first IQ score to address selective re-testing). Differential attrition will be tested for and, if necessary, selective trimming of observations will be done. The interrupted time series design will be estimated using ordinary least squares regression models. In all analyses, standard errors will be clustered by school and cohort, by school, or by district.

Products and Publications

Journal article, monograph, or newsletter

Card, D. and Giuliano, L. (2016). Can Tracking Raise the Test Scores of High–Ability Minority Students? American Economic Review, 106, 2783–2816.

Card, D. and Giuliano, L. (2016). Universal Screening Increases the Representation of Low–Income and Minority Students in Gifted Education., Proceedings of the National Academies of Sciences, 113, 13678–13683.

Working paper

Card, D. and Giuliano, L. (2014). Does Gifted Education Work? For Which Students? (No. w20453). National Bureau of Economic Research.