How Accurate is the STAR Math Assessment?

The STAR Math Enterprise assessment is a computer-adaptive test that gauges general mathematics achievement in students in grades 1 to 12. According to its publisher, Renaissance Learning Inc., “STAR Math delivers a wealth of valid, reliable, and actionable Response to Intervention (RTI) data.” This claim of validity, reliability and accuracy has been examined by researchers who are part of the national Regional Educational Laboratory Program to determine how accurately the STAR Math tests assess a student’s achievement.

The word "accuracy" refers to the match between the target population and the sample, those students who are actually tested. "Validity" refers to the degree to which an exam actually assesses the specific concept that the test claims to measure. "Reliability" refers to the fact that the assessment produces the same result each time it is given. The STAR Math Enterprise Program has strong results in all three categories.

The test data for STAR Math indicates that it is an accurate, valid, reliable assessment. In a 2002 to 2004 study completed at the Regional Educational Laboratory at Pennsylvania State University, researchers found the math assessment had internal consistency correlation coefficient scores of 0.77 to 0.80. Statistically, a high value approaching +1.00 is considered a strong direct relationship, so these scores are very high. Similar results were found in studies completed in the national development trials from 2008 to 2012.

Versions of the STAR Math assessment released between 2002 and 2011 used the scale score norms from the Pennsylvania State University study. But starting in 2008, the norming sample to test accuracy, validity and reliability was changed to reflect the changing student population in the United States. With the focus on reliability and measurement precision, over 29,000 students in grades 1 to 12 were tested across the country.

From the total sampling base of 29,288, the scores of 7,389 students were examined for alternate forms of reliability. The reliability coefficients of the STAR Math assessment remained high in the 0.7 to 0.8 range. The test is an accurate indicator of a student’s ability and growth in math.

The National Center on Response to Intervention is a branch of the American Institute for Research. In 2010, the NCRI scrutinized the STAR Math and its claim to be useful as an assessment tool to be used for students with disabilities who need response to intervention services. The 2004 accuracy, validity and reliability data again held true with generic performance level coefficients ranging from 0.79 to 0.834.