Journal article

Identifying disengaged survey responses: New evidence using response time metadata

2018

By: James Soland, Steven Wise, Lingyun Gao

Abstract

Disengaged responding is a phenomenon that often biases observed scores from achievement tests and surveys in practically and statistically significant ways. This problem has led to the development of methods to detect and correct for disengaged responses on both achievement test and survey scores. One major disadvantage when trying to detect disengaged responses on surveys is that, unlike on achievement tests, there are no correct answers. As a result, validating decision rules for detection methods is problematic. In this study, we condition results from a variety of detection methods used to identify disengaged survey responses on response times. We then show how this conditional approach may be useful in identifying where to set response time thresholds for survey items, as well as in avoiding misclassification when using other detection methods.

See More
View journal article

This article was published outside of NWEA. The full text can be found at the link above.

Related Topics