Using quality signaling to enhance survey response rates

ABSTRACT

A growing body of research has found that linking products to charitable causes enhances sellers’ reputations. This paper tests with a field experiment whether the use of the option of giving to charity to signal genuine intentions in mail surveys can enhance the effectiveness of small financial incentives. We find that they do significantly increase individuals’ responsiveness to the differences in the amount of the incentives. Cash incentives lead to an approximated constant response rate. Adding the option to donate the incentive to a specific charity significantly increases the effect of the financial incentive on the likelihood of responding.

KEYWORDS:

• Response rate
• quality signaling
• charitable donations

JEL CLASSIFICATION:

• C81
• C93
• D10

Disclosure statement

No potential conflict of interest was reported by the authors.

Notes

¹ A growing body of research has found that linking products to charitable causes enhances sellers’ reputations. For example, Elfenbein and McManus (2010) found that consumers are willing to pay about 6% more when a portion (or all) of their expenditures goes to a charitable cause selected by the seller. . In follow-up research, Elfenbein, Fisman, and Brian (2012) considered a wider set of products analyzing both sale probabilities and prices and found that charity serves as a signal of seller quality.

² See Heffetz and Reeves 2016) analysis of three official government surveys – the Current Population Survey (CPS), the Behavioral Risk Factor Surveillance Survey (BRFSS), and the Consumer Expenditure Survey (CEX)- documents important reasons for research on strategies for increasing response rates. They observe, in summarizing the importance of their findings that: ‘ … our analysis reveals a consistent picture: in our data, difficulty-of reaching is strongly correlated with important outcomes of interest, even after controlling for the main observables that typical weighting schemes are based on. … the assumption that nonrespondents look like the in sample average is simply hard to defend when said average is a moving target, changing systematically as increasingly difficult respondents are added to the sample.’ (pp 3–5).

³ The supplementary material provides the text of the descriptions used. Two of our three charities are well known in the Phoenix Metropolitan Area. The First Food Bank Alliance is also known as St. Mary’s Food Bank and was one of the first food banks in the US.

⁴ These responses are included in our analysis of the decisions to keep or donate the incentive. Our results are not changed when these observations are omitted.

⁵ These responses were treated as no donation to any charity. Analysis of the results excluding them from the sample did not alter our conclusions.

⁶ A table summarizing these results is available in the online materials.

Additional information

Funding

This material is based upon work supported by the National Science Foundation under grant number [DEB-1637590], Central Arizona-Phoenix Long-Term Ecological Research Program (CAP LTER). Thanks are due Chad J. Kniss and Mark Watkins for their contributions to the survey development and implementation.