Measurement Invariance and Response Bias: A Stochastic Frontier Approach
The goals of the present paper were to assess measurement invariance using a common econometric method and to illustrate the approach with self-reported measures of parenting behaviors before and after a family intervention. Most recent literature on measurement invariance (MI) in psychological research 1) explores the use of structural equation modeling (SEM) and confirmatory factor analysis to identify measurement invariance, and 2) tests for measurement invariance across groups rather than across time. We use method, Stochastic Frontier Estimation, or SFE, to identify response bias and covariates of response bias both across individuals at a single point in time and across two measurement occasions (before and after participation in a family intervention). We examined the effects of participant demographics (N = 1437) on response bias; gender and race/ethnicity were related to magnitude of bias and to changes in bias across time, and bias was lower at posttest than at pretest. We discuss analytic advantages and disadvantages of SFE relative to SEM approaches and note that the technique may be particularly useful in addressing the problem of “response shift bias” or “recalibration” in program evaluation -- that is, a shift in metric from before to after an intervention which is caused by the intervention itself and may lead to underestimates of program effects.
|Date of creation:||Sep 2010|
|Date of revision:|
|Contact details of provider:|| Postal: |
Web page: http://faculty.ses.wsu.edu/
More information through EDIRC
When requesting a correction, please mention this item's handle: RePEc:wsu:wpaper:rosenman-10. See general information about how to correct material in RePEc.
For technical questions regarding this item, or to correct its authors, title, abstract, bibliographic or download information, contact: (Danielle Engelhardt)
If references are entirely missing, you can add them using this form.