Non-Response Bias Associated With Statewide Angler Surveys Conducted By Internet

Monday, September 9, 2013: 1:20 PM
Fulton (Statehouse Convention Center)
Larry M. Gigliotti , South Dakota Cooperative Fish and Wildlife Research Unit, United States Geological Survey, Brookings, SD
Kjetil Henderson , Natural Resource Management, South Dakota State University, Brookings, SD
Statewide angler surveys have been used to provide estimates of fishing activity, harvest, and opinions. The South Dakota Game, Fish and Parks Department had traditionally conducted statewide angler surveys by mail. In 2010 the agency experimented with using e-mail to invite anglers to participate in an on-line (Internet) survey. A large number of responses could be achieved at great savings (no printing, mailing or encoding costs); however, Internet response rates were very low compared to mail surveys (e.g. 26% vs. 61%: resident licenses). A two-year study was initiated to investigate the extent and nature of Internet non-response bias associated with conducting statewide angler surveys.

 The potential for Internet non-response bias is large due to the relatively low Internet response rate. The overall Internet response rate was 34% in 2011 and 32% in 2012 (15% to 47% and 20% to 46% for each of the fishing license types; respectively). Of 337 variables tested among nine license types in 2011, 36 variables (11%) had significant differences between the Internet data and data collected by mail from a sample of Internet non-respondents. This presentation will summarize the nature and extent of Internet non-response bias and possible methods to correct for the biases.