Browse By Day
Browse By Time
Browse By Panel
Browse By Session Type
Browse By Topic Area
The Texas Kindergarten Entry Assessment (TX-KEA) is a standardized assessment designed to screen students when they enter kindergarten and to help identify students in need of further diagnostic assessment. TX-KEA is a screener that covers six domains of school readiness, of which two were recently expanded: language and literacy. The screening measure was launched in 2017-18 school year. The expanded measure includes two more progress monitoring time points; this version will launch in the 2018-19 school year. Three measurement time points of assessment permit ongoing monitoring of students’ learning, which is essential to inform instruction. This study sought to validate the psychometric properties of the expanded progress monitoring system.
An iterative item development process was used that included feedback from content experts and educators to develop a test blueprint. The second step involved calibrating a large initial pool of items within a scaling study that administered sets of test items to more than 2,575 children in several regions across the state of Texas. On average, children were 5.67 years, 53.59% male, and represented diverse ethnicities. Our third step was to select the most robust items from the calibration phase and submit these for additional psychometric study. This involved administering TX-KEA test items along with other standardized measures (e.g., Woodcock Johnson IV Test of Achievement, 2014) at two time points within the kindergarten year (i.e., beginning- and end-of-year) to 208 new students. The final phase of work established methods to interpret and use test scores to inform instruction. We used a criterion-referenced approach to scoring with benchmark categories including: “needs support,” “monitor,” and “on track.” We also developed an online platform with teacher reports, student grouping tools, and leveled lesson plans linked to each recommended student group.
The final language screening subtests (Vocabulary and Listening Comprehension) included approximately 35 total items at each of the three assessment time points. The final literacy subtests (Letter Names, Letter Sounds, Blending Sounds, Spelling) had 45 total items at each assessment time point. Inter-scorer reliability was high (.995) and internal consistency was acceptable - Raykov Reliability range .80 to .96. Tables 1 and 2 show generally moderate correlations between the screening and progress monitoring subtests and standardized measures (r’s = .38 to .79). Additionally, educators’ feedback surveys indicate good satisfaction with the score interpretation framework and the linked grouping and lesson resources. The current measure is publically available at no cost to ISD-based programs and has been adopted by over 80% of school districts in Texas to date; we will report on usage of the expanded version as well as ongoing psychometric analyses.
Tricia Zucker, University of Texas Health Science Center at Houston
Michael Assel, University of Texas Health Science Center at Houston
Susan Landry, University of Texas Health Science Center at Houston
Janelle J. Montroy, University of Texas Health Science Center, Children's Learning Institute
Hsien-Yuan Hsu, University of Texas Health Science Center at Houston
Colby Hall, University of Texas Health Science Center at Houston
April Crawford, University of Texas Health Science Center at Houston