Skip Header

Chuck Nelson, Leah Ericson
Component ID: #ti1566190928

Executive Summary

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.
  • For health insurance coverage (which is not currently asked in the ACS), the objective was to evaluate which of two questions works best in an ACS questionnaire environment. The two formats were: 1) Version I that asked respondents about each of the various types of coverage, and 2) Version II that asked respondents a screener question to find out if they are covered and a second question that asked about the type(s) of coverage.
  • In order to evaluate which set of questions was superior, the Content Follow-up (CFU) used a set of questions that was different than either of the two Content Test (CT) versions. The CFU utilized a more detailed and comprehensive set of questions than the CT that resulted in a higher coverage rate. Thus, the two sets of CT questions could then be evaluated in terms of which came closer to this more comprehensive set of questions.
Research Questions and Results
  • Question 1: Which of the two approaches results in more accurate estimates of insurance coverage?
    Results: Version I (that asks about each type of insurance) was superior to Version II in terms of the net difference rate between the CT and the CFU (this was the most critical of the three evaluations that comprise this research question). In terms of which method results in a smaller percentage of people for whom insurance type cannot be determined, after recoding those who reported “other” types of health insurance, both versions end up with negligible percentages of persons with unclassifiable sources of health coverage. There was no statistically significant difference in the overall health insurance coverage rate between the two versions.
  • Question 2: Which of the two methods for asking the question results in greater reporting (i.e., less under-reporting) of public health care?
    Results: Version I results in less under-reporting of public health coverage (and private coverage as well). The net difference rate for public coverage in Version I was superior to that of the Version II.
  • Question 3: Which of the two methods results in the least amount of item non-response?
    Results: For the question of whether or not respondents were covered by any type of insurance, Version II had a lower nonresponse rate. In terms of the percentage of respondents that did not report a type of coverage, the nonresponse rate of Version I was not statistically different from that of Version II.

Back to Header