Steven Pedlow is a Senior Statistician in the Statistics and Methodology Department. He is NORC's lead statistician on the Survey of Consumer Finance and other surveys. Pedlow is NORC's expert in imputation, and also has expertise in sample design, sample weighting and other estimation issues, especially for area probability surveys.
Pedlow has designed or consulted on every implementation of hot-deck and multiple imputation at NORC in the past ten years (which includes the Survey of Doctorate Recipients, the Residential Energy Consumption Survey, and multiple surveys under the National Immunization Survey umbrella). His weighting expertise includes combining a cell phone sample with a landline phone sample (for the Minnesota Health Access Survey), combining independent samples with unified probabilities rather than using composite weights (for the National Longitudinal Survey of Youth), constructing complicated panel weights (for the General Social Survey and Making Connection studies), and reducing the impact of outlier weights (for the Racial and Ethnic Approaches to Community Health 2010 study). Pedlow has also been a leading sample designer for the 1997 National Longitudinal Survey of Youth, OSHA's current procedure for the evaluation of Accuracy and Completeness of an Establishment's OSHA-Mandated Employer Records, the 2010 Census Integrated Communication Program Evaluation, and the 2010 NORC National Frame for Area Probability Surveys.
Consulting work he started in graduate school has resulted in co-authorship for articles in Pediatrics, Journal of Pediatric Endocrinology and Metabolism, Journal of Abnormal Child Psychology, Anesthesia and Analgesia, and Journal of Vocational Education Research. Pedlow has also presented papers to the American Statistical Association and American Association of Public Opinion Research on the weighting topics listed above as well as such topics as post-Katrina control totals in New Orleans, linear programming to balance automatic case assignments for field interviewers, variance estimation for ordered categories, the impact of clustering on effective sample sizes, interviewer intervention for data quality, Benford's Law to detect falsified cases, sampling undercoverage, and multiple imputation.
Pedlow is currently the newsletter co-editor for the Survey Research Methods Section of the American Statistical Association.