|
Abstract:
|
The Current Population Survey (CPS), a government survey that is the nation's source for the unemployment rate, recently conducted a study to determine the effects of sending a pre-notice postcard to sampled cases. The goals of this postcard study were to decrease the cost of the survey by reducing the number of contact attempts needed to complete an interview and increase the overall response rate. Decreasing the contact attempts for the cases that are more likely to respond could allow for more resources to be focused on the harder-to-reach cases. Field interviewers conduct the majority of the interviews, either in person or by phone. The postcards included the name and contact number of the interviewer assigned to that particular case, which gives the respondent an opportunity to call their interviewer and respond or set up an appointment. This study analyzed data from 500 interviewers and 25,000 cases over three interview months. Count regression was used to determine what effect the postcard had on the number of contact attempts per case. Mixed effects logistic models were fit to determine what characteristics, including the postcard, predict a cases' propensity to respond.
|
