The STAR Reading computer-adaptive, norm-referenced reading test and database is a periodic progress-monitoring system that incorporates state-ofthe- art testing technology, including item response theory, to provide teachers with accurate reading scores for students in grades 1 through 12. Research conducted during the development of STAR Reading confirms that the test is reliable, valid, and correlates highly with high-stakes standardized reading tests. The test was normed in spring 1999 using a nationally representative sample of 30,000 students from 269 schools in 47 states across the U.S. The reliability of STAR Reading was established with three reliability studies: test-retest (n=2,095), alternate forms (n=4,551), and generic reliability (n=29,169). The grade-level reliability estimates from all three studies are extremely high, ranging from 0.79 to 0.92 with most estimates greater than 0.85. An additional study (n>12,000) demonstrated the validity of STAR Reading by comparing students’ scores on STAR Reading to their scores on other popular standardized tests such as the California Achievement Test and the Iowa Test of Basic Skills. The high correlation (most are above 0.70) between STAR Reading scores and scores on other tests establishes both the validity of STAR Reading for measuring reading achievement and its ability to predict performance on other tests. This report contains results from each of these studies, details on the test’s standard error of measurement, and an explanation of the criterion- and norm-referenced scores provided by STAR Reading.

Publication Date: 12/01/2001

Updated: 03/01/2009

Literature Number: L0315