2014/2015 QMeHSS Workshop (10/3/2014)

Dear Colleagues,

We hope that you all had an enjoyable summer and we would like to welcome you back to a new academic year and another year of the Quantitative Methods in Education, Health, and the Social Sciences Workshop series. Our first workshop will held on Friday October 3rd from 11:00 -12:30 and will be led by Dr. Brian Junker from Carnegie Mellon University. Dr. Junker will be presenting Predictive Inference Using Latent Variables with Covariates. The workshop will be  held in the NORC conference room 344. NORC is located at 1155 E. 60th Street. We hope that you will be able to join us for another exciting year of presentations and engaging discussions.

Brian W. Junker (Carnegie Mellon University)

Joint with Dan A Black (University of Chicago), Lynne Steuerle Schoefield (Swarthmore), and Lowell J Taylor (Carnegie Mellon)

Predictive Inference Using Latent Variables with Covariates

Plausible Values (PVs) have been a standard multiple imputation tool for latent proficiency variables in large scale education survey data since their implementation in the National Assessment of Educational Progress (NAEP) in the 1980's.  Today PVs are used widely in many national and international education surveys.  When latent proficiency is the dependent variable in an analysis, well-constructed PVs provide guarantees of unbiasedness for inferences about latent proficiency.  We review the well-known results that provide these guarantees, and try to extend them to the case in which latent proficiency is one of the independent variables in an analysis.  We show that the same guarantees are impossible in the latter case, and provide an alternative approach, based on Schofield's (2008) mixed effects structural equations (MESE) model.  An example using data from the 1992 National Adult Literacy Survey (NALS) illustrates our results.

 

Relevant papers for the discussion:

·         Predictive Inference Using Latent Variables with Covariates

·         The Use of Cognitive Ability Measures as Explanatory Variables in Regression Analysis