Census Bureau plans online push for the American Community Survey
Beginning in January 2013, the Census bureau will encourage Americans selected for the Census Bureau's American Community Survey to respond online. If Internet response participation for ACS--the yearly sample survey of about 2.5 percent of American households--is extensive, the bureau will save on survey costs, according to Census Bureau Director Robert Groves.
ACS currently has three methods of survey. First, sampled participants receive a mail out/mail back questionnaire, and if they do not respond, they are contacted for computer-assisted telephone interviews and finally computer-assisted in-person interviews.
In prepared testimony (.pdf) March 6, Groves told the House Oversight and Government Reform subcommittee on health care, District of Columbia, Census and the National Archives that preliminary studies concluded that online data collection will also reduce respondent burden.
As a follow up, a National Research Council panel is assessing ACS survey methods and data products in a report due in late fall 2012, which will be integrated into a final bureau report December 2012, said Groves.
According to the first ACS internet test, conducted in April 2011, when mailings directed households to use the internet first, and did not include a paper questionnaire until a nonresponse follow-up mailing, the self-administered response rate was 2.6 percent higher than the control.
Questionnaire response time is also significantly faster when respondents used the internet than when they mail back the questionnaire, according to a report (.pdf) of the ACS Research and Evaluation Steering Committee.
"Internet respondents were more likely to be younger, Asian, non-Black, 'other' race, with higher education and living in larger households than mail respondents. They were also more likely to speak a language other than English at home," finds the report.
Detailed questions at the end of the survey led some respondents to stop the survey short, which presented a new dilemma for Census to consider, says the report.
"The differences we observed in item nonresponse rates due to Internet break-offs have prompted further consideration on…whether we should treat them as respondents or nonrespondents," notes the report.