Test Objective

In January through March of 2006, the American Community Survey (ACS) conducted the first test of new and modified content since the ACS reached full implementation levels of data collection. The results of that testing will determine the content for the 2008 ACS.

At the request of the Department of Health and Human Services (DHHS), a series of questions related to the marital history of the population 15 years and over were placed on the 2006 American Community Survey (ACS) Content Test. The questions were designed to collect annual estimates of the numbers of people who marry and divorce, the number of times people were married, and the duration of their current marriage. The motivation for these questions was to use the ACS as the primary federal vehicle for the collection of marital data to replace the discontinued marriage and divorce registration area that had previously provided this information to DHHS on an annual basis.

Methodology

There was only one set of questions in the content test. Following the basic item on the marital status of the respondent, a matrix designed set of questions was asked if the respondent had married, widowed, separated or divorced in the last 12 months. Two further questions asked the number of times married and the year the person was last married. A followup test was used to determine the consistency of the results from the original interview and included two qualitative items concerning the legal decree status of divorces and separations obtained in that 12-month period.

Research Questions and Results

The results indicate that the series of four questions used to identify the occurrence of a marital event in the last 12 months (either a marriage, divorce, separation, or death of a spouse) failed to pass as a grouped question all of the selection criteria established in the analysis plan. For one or more questions in this group, either item nonresponse rates were too high, the index of inconsistency for the items were either moderate or high, or the analytical results proved illogical or inconsistent with benchmark data from other surveys or administrative records. Empirical analysis, however, suggests that the divorce in the last 12 months test data did appear to produce reasonable estimates when comparisons were made with current estimates from the existing vital statistics system. The current version of asking a respondent’s current marital status will provide acceptable results in its new location within the survey. The two items asking about the number of times people were married and the date of their last marriage proved acceptable. These latter two items in the ACS will fulfill three components of the DHHS request for obtaining estimates of the number of marriages occurring in the last year, times married, and the duration of the current marriage.