Dynamic Assessment with Bilingual Children: Implications for Improving Accuracy of Identification for Language and Reading Disorders

Background: Dynamic assessment (DA) is a means to examine a person’s learning during a short teaching session and is helpful for students disadvantaged in traditional testing settings. DA has been effective in accurately identify bilingual learners struggling with reading and language skills, without mistakenly identifying normal second language acquisition as disability. Correctly differentiating between the two is important for children to receive appropriate supports. Aim: The purpose of this study was: 1) To examine whether a DA measure added predictive value to 1st grade reading scores beyond a commonly used benchmark reading assessment 2) To examine whether or not students benefited from the teaching portion of DA 3) To examine the relationship between a one-time DA screener and teacher perceptions of students’ learning during regular instruction. Method and Procedure: Fifty-nine kindergarten students participated in one DA session using a screener, the PEARL, to assess decoding and storytelling proficiency before and after a short instructional episode. Students were scored on accuracy and responsiveness to teaching. In addition, classroom teachers completed an independent rating of student’s responsiveness to teaching during a typical lesson. Stepwise linear regression was used to determine predictive importance of the DA on 1st grade outcomes. A two-tailed t-test was used to compare student scores before and after teaching. Correlations were used to compare the assessment and teacher rating results. Results: The stepwise regression revealed that the best fitting regression model for predicting students’ 1st grade reading scores was a model with 2 variables, WIDA Literacy Score, a standardized measure of English proficiency, as the first variable and PEARL Story Grammar Score + Teaching Responsiveness, a DA measure, as the second variable. Further results indicated significant changes in pre and posttest data using the DA screener and moderate correlations between teacher rating data and the DA