2016-2017 Spring

John Krosnick, Department of Communication, Stanford University

April 3, 2017

Optimizing Design of One of the Most Fundamental Measures in Surveys: “Who Lives Here?”

ABSTRACT: Almost all scientific surveys of random samples of populations must begin by identifying how many adults live at a place of residence, and most surveys then randomly select one of the residents to interview. It might seem simple to determine who lives at a residence, with the goal of assigning every American adult to only one residential address in the country. But in fact, this turns out to be surprisingly difficult. And amazingly, the U.S. Census Bureau, perhaps the organization in the U.S. that spends the most money conducting surveys that have the most impact on policy-making, does not have a single ideal approach to accomplishing this goal in its surveys. Instead, different Census Bureau surveys (and the Census itself) identify household residents differently. A close look at their procedures are both surprising and sometimes even humorous. In this presentation, Jon will review the various ways in which Census Bureau surveys ask this seemingly simple question, highlighting many surprising challenges in the measurement process. And then, PPRG members will brainstorm as a group to see if we can come up with an ideal way to do this that can be tested in the future.