Improving Physician Survey Response Rates: Compelling Evidence for Non-Contingent Incentives
Submission ID: 5629

Date: Friday, 9:30 AM to 10:15 AM 
Session: Poster Session 3: F9:30 - 10:15 AM

Primary Presenter

Emily Hachey, SSRS

Additional Authors or Round Table Presenters

Eran Ben-Porath, SSRS ,

Gillian SteelFisher, Harvard TH Chan School of Public Health ,

, ,

, ,

, ,

, ,

, ,

, ,

Abstract

The field of survey research is continuously looking for new ways to increase response rates in order to improve data quality along with productivity. Building on previous research, this poster focuses on one feature of mailed survey invitations that has the potential for this type of improvement: the use of non-contingent incentives. Non-contingent incentives refer to incentives offered in the invitation to participate in a survey, irrespective of whether the invited person completes the survey or not. By contrast, contingent incentives are furnished to respondents only after they complete the survey, and are typically larger than non-contingent incentives.  

The  three-month fielding for survey of primary care physicians (PCPs) in the U.S. (n=397) provided an opportunity to test the effect of these two incentive types on response rate. For this study, we mailed survey invitation packages to named physicians. The findings we report are based on an experiment (n=1,125) embedded in that process, which tested the effectiveness of two incentive types on reaching higher response rates : a $25 cash, non-contingent incentive; and a $75 contingent incentive (provided by check or electronic gift-card upon completion). The findings were statistically significant and substantively clear: smaller, non-contingent incentives yielded much higher response rates, and did so at a lower cost-per-complete. Respondents who received non-contingent incentives were also more likely to complete the survey by hardcopy. This poster will present data on response rates based on contingent and non-contingent incentives, demonstrate demographic differences between those who completed the survey in each group, and consider the circumstances most appropriate for including non-contingent incentives in invitation packages.