The February 5, 2015 seminar features Daniel McCaffrey, principal research scientist in Statistical and Psychometric Theory and Practice at Educational Testing Service (ETS). McCaffrey takes up the topic of value-added modeling (VAM), and will help us understand the ways in which VAM might – and might not – be useful for assessing the quality of teaching and teacher preparation.

Educational Testing Service (ETS) presentation



Thursday, February 5, 2015
4:10pm - 6:00pm

  • Daniel McCaffrey
    Principal Research Scientist in Statistical and Psychometric Theory and Practice, Educational Testing Service

About the Speakers

Daniel McCaffrey

  • Principal Research Scientist in Statistical and Psychometric Theory and Practice, Educational Testing Service

Daniel F. McCaffrey is a Principal Research Scientist-STAT at ETS.  He is a national leader on teacher evaluation, in particular on value-added modeling for evaluating teacher performance and has served on eight state and school district teacher evaluation advisory boards. He has conducted extensive methodological and empirical work on value-added, with over 20 scholarly articles, chapters and monographs on the topic. Recently has also conducted work on teacher observation studying issues of rater drift, mode of observation, bias in factor analyses used to test construct validity, and difference in rating across grade level. Dr. McCaffrey is currently collaborating with TeachingWork on the development of the new ETS teacher licensure tests. Dr. McCaffrey was the lead statistician for the MET project and led the efforts to estimate value-added for the project and the analysis of the reliability of teacher observations for that study. He led the effort to randomize the teachers and build the project databases that combine data from districts, observations, surveys and other sources. Dr. McCaffrey was also lead statistician of the Gates Funded Understanding Teaching Quality project and William T. Grant- Spencer funded Toward an Understanding of Classroom context for which conducted extensive analysis of teacher evaluation data and managed the database construction and analysis teams. He was the PI of RAND’s evaluation of New York City’s School-wide Performance Bonus Program for teachers, which was awarded the American Educational Research Association, Division L, Outstanding Report of 2012. He also led RAND’s work for the National Center on Performance Incentives, which conducted the first large-scale randomized trial of teacher performance pay in the United States. Currently Dr. McCaffrey is conducting research on the impact of measurement error in tests on various secondary analyses of test scores for evaluation including value-added and student growth percentiles on IES funded research. He is also working the co-PI on an NIDA grant on causal modeling methods for the evaluation of adolescent substance abuse treatment.  He is a Fellow of the American Statistical Association.