Paper Summary
Share...

Direct link:

Using Multivariate Generalizability Theory to Correct Correlations for Common Sources of Measurement Error

Mon, April 11, 2:45 to 4:15pm, Convention Center, Floor: Level Two, Exhibit Hall D Section C

Abstract

Internal consistency and test-retest reliability estimates routinely overestimate reliability of assessment measures because they do not take into account all relevant sources of measurement error (i.e., random response, specific factor, and transient error). This can lead to misrepresentations of relationships among educational and psychological constructs. We apply Generalizability Theory based procedures to more fully estimate reliability of scores from individual difference measures and to estimate relationships between constructs. The procedures quantify the relevant sources of error separately and integrate them into formulas that allow for more comprehensive estimates of the reliability, convergent validity, and discriminant validity of scores. The procedures only require that a measure be administered on two occasions and can by implemented using software in the public domain.

Authors