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.
• The proposal to collect property value information as an open-ended format, similar to the approach followed for monthly rent paid, is an attempt to introduce more precision to the value distribution. Economists and housing analysts at HUD have encountered considerable difficulty using the bracketed data and have recommended collecting property value information as point estimates in the ACS rather than continuing with the current categorical approach established in prior decennial census data collection efforts. HUD states that the categories used in Census 2000 may not serve them well in the coming years if the housing market continues at the pace established in the first half of this decade. In addition, this is the only dollar value on the ACS questionnaire that is currently collected as categorical data. This makes it difficult to inflation-adjust the data from year to year.

Methodology

• The Content Test compared two versions of the property value question.
• The control version was not exactly the same as the questionnaire used in production. It included the following changes to the question on property value: o Updating the response categories by incorporating more high-end categories than are shown on the ACS production questionnaire. o Eliminating the open-ended category for those who indicated a value of “$250,000 or more.”
• The test version included the following changes to the question on property value: o Revising the question wording by adding the word “About” to the beginning of the question. (About how much do you think this house and lot, apartment, or mobile home (and lot, if owned) would sell for if it were for sale?) o Using an open-ended/write-in field rather than categories.

Research Questions/Results

• Research Question 1: Which of the two approaches for changing the current ACS property value question (control or test) results in a lower item non-response overall and for those in mobile homes?
  The results do not support the hypothesis. The item non-response rate for the test version is higher than that of the control version for owner-occupied units. There is no difference in non-response rates for owner-occupied mobile homes.
• Research Question 2: Do the changes made to the property value question using the two different approaches produce similar median values?
  The results support the hypothesis that the median value for the test version does not significantly differ from the control version for owner-occupied units, with the exception of Low Response Areas. For owner-occupied mobile homes, the results show no significant differences between control and test.
• Research Question 3: For the test version of the property value question, does the median value directly calculated from the open-ended responses differ from the interpolated median calculated from the recoded responses?
  The median value directly calculated from the open-ended responses was slightly lower than the interpolated median calculated by categorizing the open-ended response into the 19-value categories at the national and high response areas. While the differences in the low response areas were also lower, they were more pronounced. This is for informational purposes.
• Research Question 4: Do the changes made to the property value question using the two different methods result in similar distributions of property values overall, and for those in mobile homes?
  Overall, the results do not support the hypothesis. The test distributions for owneroccupied units at the national level and within high response areas are significantly different from the control distributions. The property value distributions resulting from the test version are not significantly different from the control version for owner-occupied mobile homes and for all owneroccupied units within low response areas.
• Research Question 5: Do the changes made to the property value question using the two approaches produce distributions that are similar to property value distributions produced by other surveys?
  The median property value is similar to the median produced by the American Housing Survey (AHS). This is for informational purposes only.
• Research Question 6: Do the changes made through the test and control version of the property value question produce comparable levels of reliability overall and for those in mobile homes?
  The results support the hypothesis that the index of inconsistency for the test version is comparable to that of the control version for all owner-occupied units. The index of inconsistency for owner-occupied units, mail only, shows that both the test and control are in the moderate range. Therefore, the test version maintains the reliability of the property value question.

Summary of Empirical Results

The median values are similar for the ACS test and control versions, and the reliability between the control and test is the same or better. The national median is also similar to the median property value from the 2005 American Housing Survey.

Incorporating the changes may come at the cost of higher non-response rates to the property value question. However, it will facilitate the collection of more precise data and thus serve the needs of HUD, the major Federal user of housing statistics.

The empirical results show that the test version, with the exception of the higher non-response rates, performed as well as the control version.