A Cautious Survey Statistician’s Approach to Estimation in the Age of Big Data
Abstract:
For survey statistics practitioners, it sometimes feels like the sky is falling. Response rates are declining. Data collection costs are increasing. Federal budgets are shrinking. For survey statistics researchers, the sky isn’t falling but is raining exciting, new data sources and modeling techniques. In this talk, I will present one modern, but cautious, approach to survey estimation where predictive models link survey data with additional data sources. Drawing on my collaborations with the U.S. Bureau of Labor Statistics and the U.S. Forest Inventory and Analysis Program, we will explore the performance of this approach and consider some lessons learned. And, as a firm believer in the transformative impact of undergraduate research, I will close out by describing a summer program I developed in forestry data science and strategies for involving undergraduate students in one’s scholarly work.