Predictors of Mandatory Third Grade Retention From High-Stakes Test Performance for Low-Income, Ethnically Diverse Children




Tavassolie, Tanya

Journal Title

Journal ISSN

Volume Title



As a result of the No Child Left Behind (NCLB) policy, Florida, like many other states, has mandatory third grade retention for children who fail the high-stakes reading test, the Florida Comprehensive Assessment Test [FCAT]. Since implementation, the percentage of third graders retained has increased dramatically. Various exceptions and loopholes exist to help with retention and promotion decisions. Little research exists, however, on actual practices in enforcing mandatory retention policies from high-stakes test results. Using data from the Miami School Readiness Project (MSRP), I examined a large (N = 27,980) and ethnically diverse (59.2% Latino, 33.3% Black, 7.4% White/other) sample of third graders over five cohorts who completed third grade between 2006-2010. I ask the following research questions: (1) What proportion of children take and fail the FCAT reading test, and of those children, how many actually repeat 3rd grade? (2) To what extent is failing the FCAT, and subsequent retention, related to demographic variables and children’s academic performance in 3rd grade? (3) When considering all demographic factors, what are the unique and combined predictors of failing the FCAT, and actually being retained? Multiple logistic regression analyses revealed that even after accounting for demographic variables (e.g. free/reduced lunch, gender, ethnicity, ELL and special education status), those receiving free/reduced lunch, who were not English proficient, and in special education in third grade were more likely to fail the FCAT reading test, and be retained after failing. Latinos and Blacks were less likely to be retained even after failing the FCAT, compared to Whites. Implications for high-stakes testing policy in Florida are discussed.



High-stakes testing, Achievement tests, Grade retention, Retention policy