The paper was presented at the Annual Meeting of the American Statistical Association (ASA), August 1995.

This paper reports the results of research and analysis undertaken by Census Bureau staff. It has undergone a more limited review than official Census Bureau publications. This report is released to inform interested parties of research and to encourage discussion.

Every 10 years the Census Bureau embarks on the enormous task of counting every person in the United States. In addition to collecting population and housing counts and basic demographic information, the decennial census obtains detailed data from a sample of households. This sample information, often referred to as long form data, is crucial to Federal, state, and local governments for planning and distribution of governmental funds. Private companies also make major business decisions on programs and day-to-day operations based on decennial census data. Considering the many users of these data and the constantly changing face of the Nation, the Census Bureau is responding to the ever increasing need for more frequent collection of the sample data. The Continuous Measurement Survey (CMS) is the vehicle the Census Bureau is proposing to use to collect and distribute this information on a more timely basis over the decade.

The CMS will be a large monthly household survey implemented in 1999. It will use a tri-modal data collection methodology which will 1) Mail or deliver to a large sample of addresses, 2) Attempt to interview the nonrespondents by phone, and 3) Select a subsample of the remaining nonresponding sample units, upon completion of the telephone follow-up operation, for personal interviews.

The mail component of the test relies on self-response using a paper questionnaire, while the telephone and personal interview follow-up will be computer assisted. We will begin testing all three data collection modes and their integration in 1996.

Beginning in November 1994, we tested the computer assisted telephone interviewing (CATI) phase through the 1995 Continuous Measurement CATI Test. This test had two objectives. The first objective was to test our ability to gather long form data over the telephone. To help with this evaluation, we tested different wording and reference periods. The second objective was to gather operational and cost data to help with planning of the 1996 test and subsequent phases of the CMS.

The 1995 test was the first Census Bureau survey to use the CASES authoring software and the new CENCATI Control System in production. November data were used only to ensure that the system was working properly. All analyses in this paper use data for December 1994 though April 1995.

The purpose of this paper is to discuss selected results of the 1995 test. Information on the sample, vendor-supplied telephone lists, response and refusal rates, length of interview, best time to call, and the number of calls to complete a case are presented in this paper.