Interpolating Population Distributions using Public-Use Data: An Application to Income Segregation using American Community Survey Data

Income segregation measures the extent to which households choose live near other with similar incomes. Sociologists theorize that income can exacerbate impacts of inequality, and have developed indices measure it at metro area level, including information theory index introduced in \citet{reardon2011income}, divergence presented \citet{roberto2015divergence}. To study their differences, we construct both using recent American Community Survey (ACS) estimates features distribution. Since elimination decennial census long form, methods computing these must be updated use ACS account for survey error. We propose a model-based method interpolate distribution accounts this This improves on previous approaches by allowing more types estimates, providing uncertainty quantification. apply estimate U.S. tract-level distributions tabulations, turn indices. find major differences between two relative ranking areas, as well how correlate Gini index.