|Title:||Novel Models and Methods to Address Measurement Error Issues in Educational Assessment and Evaluation Studies|
|Principal Investigator:||Cai, Li||Awardee:||University of California, Los Angeles|
|Program:||Statistical and Research Methodology in Education [Program Details]|
|Award Period:||3 years (7/1/14 – 6/30/17)||Award Amount:||$895,108|
|Goal:||Methodological Innovation||Award Number:||R305D140046|
Measurement error issues adversely affect results obtained from typical modeling approaches used to analyze data from assessment and evaluation studies. In particular, measurement error can weaken the validity of inferences from student assessment data, such as the inferences made from the results of using value-added models. Measurement error also can reduce the statistical power of impact studies and can diminish the ability of researchers to identify the causal mechanisms that lead to an intervention improving the desired outcome.
The purpose of this project is to develop models and statistical software to account properly for the impacts of measurement error. Researchers plan to use multilevel latent variable modeling to develop multi-stage and single-stage estimation methods that will address some problems created by measurement error. By the end of the study, researchers will develop statistical models, computer software, and accompanying manuals that will be useful for addressing these difficulties related to measurement error. In addition, the researchers plan to offer workshops and web-based materials that will help applied researchers make use of the models and software they have developed.