A comparison of two early education literacy benchmark assessments

The need for public education accountability and recent federal legislation such as No Child Left Behind (NCLB) and Individuals with Disabilities Education Act (IDEA) requires effective assessment of all children to ensure adequate instruction. Assessment measures that attempt to provide this accountability vary greatly in terms of type of measure, required assessment time, and predictive relationship with end-of-year state assessment measures. The most prevalent assessment tool in Virginia is the Phonological Awareness Literacy Screener (PALS) which was developed by the University of Virginia to facilitate the Lady Intervention Reading Initiative, an initiative established to reduce the number of children with reading problems by early diagnosis and immediate intervention. PALS has an adoption rate by public school divisions in Virginia close to 100%. Another literacy screener is AIMSweb, developed by Edformation and currently owned by Pearson. It is used in progress monitoring in other states, and commonly used to facilitate Response to Intervention (RtI), a system-wide early identification and intervention process that attempts to provide research-based and alternate forms of instruction within the general education environment.This study compared these two operationally divergent universal screeners for early literacy. The first comparison determined whether AIMSweb identified the same population as at-risk that PALS identified as at-risk, This was a direct comparison of 2007-2008 fall benchmark results. Findings indicated that AIMSweb measures identified the same students that the PALS measures identified, and many component measures had highly significant correlations. This second comparison determined the sensitivity and specificity of PALS 1-3 in predicting the success or failure on the third grade Standards of Learning assessment (SOLA). Third grade fall and spring PAIS scores from 2006-2007 were compared to performance on the third grade 2000-2007 SOL assessment. These findings demonstrate that PALS 1-3 fall scores have a lower correlation with SOLA results than is stated in the PALS technical manual, but higher than the most recent study that examined the correlation between PALS and the SOLA. It also determined that the measure had less sensitivity and specificity in the spring benchmark assessment period than the fall. Recommendations were made to consider the efficiency of using AIMSweb as an initial screener, and then utilizing PALS for those identified students for whom more diagnostic assessment is needed. AIMSweb would also provide an additional and needed tool for predicting SOLA performance.