Assessing the Statistical Validity and Reliability of Student Surveys by Grade Level

• In Event: Poster Session 12
  In Poster Session: Division L Section 9 Poster Session

Abstract

We test the convergent validity of student survey responses with teacher value-added and observation ratings from teachers’ principals, as well as the inter-rater and test-retest reliabilities of a student survey instrument. By comparing results across grade levels, from 3rd through 12th grades, we provide evidence to guide empirical considerations of the efficacy of feedback from student surveys for either formative or summative teacher evaluation. Preliminary results suggest considerable increases in the reliability of student survey feedback beginning in 6th grade, though responses from elementary school students demonstrate statistical properties supportive of their inclusion in formative teacher evaluation systems.

Authors

• J. Edward Guthrie, Vanderbilt University

• Ryan Balch