Last revised: September 10, 2009

• 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.
• The objective of the test question on year structure built was to facilitate the collection of more current data throughout the decade for units built in 2000 or later.

• The Content Test compared two versions of the question on year structure built. The control version replicated the wording and response categories used in the current ACS question.

The test version included the following changes to the question on year built:

• Collapsing the “2000 to 2004” and “2005 or later” into “2000 or later” and adding the instruction “Specify year” as the most recent category

• Research Question 1: Do the changes introduced by the test question affect item nonresponse?

The results support the hypothesis that the item non-response rate for the test version will be equal to or less than that for the control version because there were no significant differences between the control and test versions at the national level as well as in high and low response areas.

• Research Question 2: Do respondents who indicate built in the year 2000 or later also provide a specific year?

Among the respondents who indicated built in the year 2000 or later on the test version, approximately 12 percent did not provide a specific year of construction at the national level and in the high response areas. In the low response areas, the nonresponse rate was about 13 percent.

• Research Question 3: Do the changes affect the distribution of the year built question?

Generally, the results support the hypothesis that the test version will not cause a statistically significant change in the year built distribution. The overall distribution for year built was not significantly different between the control and test panels.

The item nonresponse rate for the test question is similar to that of the control and the majority of respondents who check the “2000 or later” checkbox are entering the year built. The distribution rates remain unchanged for the nation as well as high and low response areas. Based on these empirical results, the test version did not adversely affect data quality, but did not suggest any significant improvement over the control version. However, the test version does allow for more detailed data on year built for units built in 2000 or later. This approach would facilitate the collection of single year estimates of newly constructed units throughout the decade.