Notice: the WebSM website has not been updated since the beginning of 2018.

Web Survey Bibliography

Title Unintended Consequences of Incentive Induced Response Rate Differences
Source The American Association for (AAPOR) 60th Annual Conference, 2005The American Association for Public Opinion Research (AAPOR) 60th Annual Conference, 2005
Year 2005
Access date 28.04.2005

Studies are finding that many previously held hypotheses about the effectiveness of varying incentive approaches remain intact when implemented in Web surveys. Given low per respondent costs in Web surveys, researchers take advantage of the ability to increase sample sizes, and frequently use sweepstakes based incentives, where respondents are entered into a drawing to win a fixed prize or prizes. The effectiveness of this kind of incentive has been mixed. However, a recent experiment of first contact mode and incentives types in a Web survey of 2,500 undergraduate college students conducted by these authors demonstrated that: 1. First respondent contact send via US Mail improves response rates over the use of email for first contact. 2. A sweepstakes drawing for $500 improves response rates over the use of no incentive. 3. A pre-paid $2 bill improves response rates over the use of the sweepstakes drawing for $500. 4. Use of a pre-paid $2 bill AND a drawing for $500 improves response rates over the use of either single incentive alone. In this study, the resulting response rate (AAPOR RR2) spread between treatment groups was 28%. The common assumption would be that with the increased response rates comes higher data quality. In looking at key measures in this study, we have found that there may in fact be significant data quality differences between the modes, including a finding of significantly different measures of smoking prevalence between experimental conditions. Singer (2002, Chapter 11 in Survey Nonresponse, eds. Groves, et al.) described that while incentives have their well documented intended consequences, they may as well have many unintended consequences on data quality. In this presentation we will discuss what we have found with regards to the impact these incentives have had on missing data, key variable response distributions, and sample composition.

Access/Direct link

Conference program

Year of publication2005
Bibliographic typeConferences, workshops, tutorials, presentations