689 – Evaluating Robustness and Instrumentation in Economic Problems
The Effect of the U.S. Census Bureau Realignment on the National Crime Victimization Survey and the Consumer Expenditure Quarterly Interview Survey
Lindsay Longsine
U.S. Census Bureau
Danielle Castelo
U.S. Census Bureau
In 2012, the U.S. Census Bureau reduced the number of regional offices from twelve to six. There were concerns that the changes in the field might adversely affect estimates from our demographic surveys, including the National Crime Victimization Survey (NCVS) and the Consumer Expenditure Quarterly Interview Survey (CEQ). To study this, we developed statistical models to measure the effect of the new management structure on certain key response variables and monitored them on a daily basis. The four key variables chosen for the NCVS were total person crime, total violent crime, total property crime, and total overall crime. The CEQ survey had seven key variables quarterly rent, monthly rent equivalent, property value, monthly gasoline expenditures, quarterly health and medical expenditures, weekly grocery expenses, and quarterly contributions. Using logistic regression and linear regression models, we measured the effect of cases coming in under the new management. In this paper, we will discuss the details of the models, an interpretation of the results, and discuss further research ideas.