Presentation to MRIA Ottawa Chapter Doug Anderson, SVP Public Affairs Harris/Decima 160 Elgin Street Suite 160 Ottawa, ON K2P 2P7 Richard Leigh-Bennett, CMRP SVP Harris/Decima 160 Elgin Street Suite 160 Ottawa, ON K2P 2P7
Non-response Rate The collective body of empirical work suggests no consistent relationship between response rate and non-response bias. RDD telephone surveys with low response rates generally still have excellent demographic representativeness. (Keeter et al. 2006).
Questions we had…. With declining response rates, are we just getting the same respondents who do multiple surveys or are we actually getting different people but they are harder to reach? –Number of calls received asking to do a phone survey and number of times agreed to do the survey in past year. –Likelihood to complete a phone survey in the next year. Online coverage is coming close to telephone coverage, but who is doing online surveys? –Number of times participated in any sort of on-line survey. –Membership in an on-line panel. –Likelihood to complete an online survey in the next year.
Expected Future Trend In the next 12 months, do you think you will be more likely, less likely, or just as likely to complete a phone survey, if asked??
Who has done any sort of online survey in past year? In the past 12 months, that is since May 2009, how many times have you agreed to participate in any sort of online survey conducted over the Internet?
Who has joined online panels? Have you ever joined an online group, community or panel for the purpose of completing surveys over the Internet?
More questions… The toughest strata to reach by phone are among the most participatory online. –Is dove-tailing samples a way to help, or is it more risk than benefit? We build/weight both phone and online samples to look like Census profiles. –If we know that either of the two are not truly identical to the Census profile and vary in identifiable ways, would we design samples to reflect these limitations? Phone response rates have continued to decline, but perhaps that decline has slowed. Our findings show as online panels grow, response rates also decline. –Should we read anything into the responses on respondents expectations for the future?