Association for Public Policy Analysis & Management

Increasingly, states are using Kindergarten Entry Assessments (KEAs) to measure children’s school readiness when they enter kindergarten. KEAs can be used by policymakers to inform policy decisions about early learning systems and by teachers to identify children’s skills (Regenstein, Connors, Romero-Jurado, & Weiner, 2017). This paper presents another potential use for KEAs – as interim measures of later reading proficiency. However, research is needed to determine how KEAs are related to reading proficiency.

The School District of Philadelphia (SDP) wanted to set a threshold on the KEA to use as an intermediate measure of citywide progress toward substantially increasing the percentage of students who can read on grade level by the end of grade 3. To develop the school readiness threshold, their research partner, Mathematica addressed the following questions:

1. What is the relationship between children’s competencies at the start of kindergarten and their reading skills in grade 3?
2. What threshold scores on the KEA most accurately predict reading proficiency at the end of grade 3?

Analyses to address these questions used district data on 3,521 kindergarteners with scores on two validated dimensions of the state’s KEA (Emerging Academic Competencies [EAC] and Learning Engagement Competencies [LEC]) and the 3rd grade state reading assessment. We first randomly selected “training” and “testing” data to check that the results are not unique to the sample used. We then weighted the analytic sample to reflect the characteristics of students with scores on the KEA (some of whom were missing scores on the 3rd grade reading assessment). We then assessed sensitivity (the rate of accurately identifying students who are proficient) and specificity (the rate of accurately identifying students who are not proficient) for models that set a threshold on the EAC, LEC, or a combination of both scales based on coefficients from a logistic regression. Mathematica worked with SDP to select a threshold that accurately predicted the proportion of proficient students in the kindergarten cohort. We then assessed the accuracy of the selected threshold in the “testing” data. We also explored alternative thresholds that might be more appropriate for other purposes, such as identifying individual students for additional supports. SDP will use KEA thresholds in future cohorts to track progress towards school readiness and likelihood of reading proficiency and could also identify schools and students who may require additional supports.

This paper provides an example of how states or districts can use KEAs to track progress towards school readiness and predicted reading proficiency across cohorts of kindergartners and target supports to schools and students. Moreover, this paper provides important insight into how early teacher-rated skills relate to later, direct assessments of reading.

Findings are embargoed by U.S. Department of Education’s Institute of Education Sciences until the study is published, but will be available to share at APPAM.