Census Bureau

Improvement of Response Rates on the Manufacturers' Shipments, Inventories, and Orders Survey with CATI

Douglas Bond and Richard Sigman

RR93/07,10/18/93

ABSTRACT

We used CATI follow-up for nonresponse to improve response rates to mailed questionnaires in a nine-month study of the Manufacturers' Shipments, Inventories, and Orders (M3) survey. Return rates increased 30 to 40 percentage points for a group of chronic nonrespondents, and reporting was more timely. In a group of good, but late reporters, return rates improved an average of 20 and 40 percentage points in time for the advance and preliminary M3 reports respectively, and imputation rates declined. Interviewers needed a little over three calls and 10 to 11 minutes logged on to the computer (including about six minutes of conversation)to complete CATI case by the end of the study. Recommendations for CATI use and future research are presented.