Abstract
In this paper, I focused upon improving the predictive validity of the Florida Assessments for Instruction in Reading (FAIR) as it pertains to subsequent proficiency on the Florida Comprehensive Assessment Test (FCAT) in reading for middle school students within a participating school district. The Florida Center for Reading Research (FCRR) (2009) has published moderate correlations between the risk status designation of the FAIR and subsequent FCAT reading proficiency. I supported these conclusions in my previous research (Clark, 2011). This study sought to improve the practical significance of the predictive relationship by identifying candidate variables for addition into the predictive model. I analyzed three years of middle school data from a participating school district with an exploratory logistic regression and identified 6 significant predictors from among 33 candidate variables. The predictors included three years of prior FCAT reading levels, the reading comprehension and Maze percentiles from the first FAIR assessment periods, and reading comprehension percentile from the second assessment period. The resultant experimental model predicted subsequent FCA T reading proficiency significantly better than the control model for the 2009-201 0 and 2010-2011 years. The experimental model was also a significant predictor of FCA T reading proficiency in 2011-2012, a school year without released risk-status data. I concluded with a discussion about the appropriateness of the FAIR as a progress monitoring tool and presented an alternative procedure for predicting reading success as Florida transitions into the new assessments accompanying the Common Core State Standards.