Comparing the Accuracy of Two Popular Early Literacy Screeners

Comparing the Accuracy of Two Popular Early Literacy Screeners

First Author: Gail Lovette -- University of Virginia
Additional authors/chairs: 
John L. Hosp; William J. Therrien
Keywords: Assessment, At Risk Students, Early Literacy, Correlational studies
Abstract / Summary: 

Purpose:
Universal screening is critical to identifying children at-risk for reading difficulties. The Phonemic Awareness Literacy Screening (PALS) is administered in grades K-3 throughout Virginia. Likewise, the Dynamic Indicators of Basic Early Literacy Skills, 8th edition (DIBELS-8) is a validated K-8 literacy screener used widely across the country. Despite widespread implementation, no published studies have explored the correlation and classification accuracy between the two screeners. The purpose of this study was to compare the diagnostic efficiency between PALS and DIBELS-8.

Method:
Participants (n=417) enrolled in grades K-3 in a district in Virginia. DIBELS-8 and the fall PALS screening measures were administered at the beginning of the school year. Bivariate correlations between PALS-K and PALS-1-3, and grades K-3 on the DIBELS-8 for overall scores and subtests and classification accuracy statistics were calculated.

Results:
Descriptive statistics demonstrate differences in the number of students identified as at-risk. Across grade levels, DIBELS-8 identified between 46-66% more students. Additional analyses for kindergarten only are presented below due to space constraints. Results for all grades will be provided on the poster. The correlation between overall scores of DIBELS-8 kindergarten and PALS-K was .814. However, Overall Correct Classification (.495), Kappa (.429), Sensitivity (.275), Specificity (1.000), Negative Predictive Power (.375), and Positive Predictive Power (1.000) indicate that there are functional differences between the two instruments.

Conclusion:
Despite both purporting to be early literacy screeners, DIBELS-8 and PALS identify different groups of students as at-risk. Further research comparing performance to a common, external criterion is warranted.