JUSTUS RANDOLPH, MERCER UNIVERSITY ROSE PREJEAN-HARRIS, INDIAN CREEK ELEMENTARY SCHOOL The Negative Consequences of Using FRL As a Measures of Poverty: The Case of the State of Georgia
Poverty in Georgia What proportion of children in Georgia live in poverty*? A.1 in 2 B.1 in 3 C.1 in 4 * $19,000 annual income for a family of three
Poverty in Georgia Approximately what proportion of children in Georgia live in poverty*? A.1 in 2 (Hispanic children [40%]) B.1 in 3 (African American children) C.1 in 4 (All children) * $19,000 annual income for a family of three
Poverty and Academic Achievement
Free and Reduced Lunch as a Measure of Poverty Reduced lunch: About $45,000 for a family of four (1.85 * Federal poverty level). Free lunch: About $30,000 for a family of four (1.30 * Federal poverty level). Cruse and Powers (2006) found the median prediction error rate to be very high: When comparing FRL and census statistics, there was a 30% error rate. Why is the error rate so high when using FRL as a measure of poverty?
Methods: Data Collection & Analysis Pre-existing data from Dr. Prejean-Harris’s dissertation GADOE Mathematics and Science CRCT data ( ) National Center for Educational Statistics FRL data ( ) Correlational design, visual, and regression analysis School-level unit of analysis
Participants & Setting All Georgia elementary schools ( ): About 130,000 Georgian third-grade students 1,244 elementary schools 191 districts 63% of students received free and reduced lunch 46%, 36%, 14%, 4%, 3% were White, African American, Hispanic, Asian and Mixed-Race, respectively
Results: School-Level CRCTs & FRL
MATH CRCT Distribution
FRL Distribution
Imagine Three Schools School A: 95% of students belong to a family with an annual income of $45,0000. School B: 95% of students belong to a family with an annual income of $30,0000. School C: 95% of students belong to a family with an annual income of $5,0000.
Imagine Three Schools On paper, they all have the same percentage of students receiving free and reduced lunch (95% FRL)!
Implications Inequity of allocation of finances Inequity of allocation of resources Inequity of accountability (e..g, value-added models)
Accountability Revisited Expected CRCT values of School A, B, and C with range restriction (817) School B projection w/o range restriction (786); FRL*.67 = CRCT
Conclusions FRL has a restriction in range. As a result, it can lead to biased projections. Using those biased projections can lead to educational inequities. Therefore, don’t use FRL if you have a better measure. If you don’t have a better measure, note the limitations and consequences of using FRL as a measure of poverty.
Thank you A draft of our paper can be downloaded from: