Skip to main content

HRPP Manual Section 12-12

Guidance on Suspicious Survey Submissions

The Michigan State University Human Research Protection Program (HRPP) and the Institutional Review Board (IRB) are currently receiving inquiries requesting assistance with strategies for dealing with compensation for suspicious survey submissions.  Suspicious submissions can include bot driven responses, responses from individuals who do not meet the defined qualifications of the study, and other forms of dubious submissions. The HRPP/IRB cannot provide a comprehensive definition of what constitutes a suspicious submission, but offers some suggestions for strategies to deter them.

There are some tools that study teams can use to identify and reject submissions from suspicious respondents. The following are but a few that researchers have found useful:

  1. AI detection tools such as CAPTCHA tests.

  2. Attention check questions in the survey.

  3. Survey platform tools, such as those provided by Qualtrics that may detect bot and other suspicious submissions. Study teams may want to consult with the survey platform team to see if they have additional options in place to make the survey more secure.

Included below are examples of suggested (but not required) consent language that describe such tools and checks:

  • AI detection tools (e.g., CAPTCHA tests) will be used to confirm the authenticity of human participants. If you are unable to confirm your human authenticity, then you will not be able to participate in this study.

  • This study includes multiple attention-check questions. Failing to answer any attention-check question correctly will result in receiving no payment for the study. This study may only be completed once per individual. Submitting multiple responses will result in receiving no payment for the study for any of the responses. Multiple-submission checks will be conducted based on an analysis of submitted data and meta-data (e.g., IP addresses).

If a study involves a survey with compensation, the study team may want to consider making the following addition to the consent form(s):

“The research team will review all submissions for quality and to confirm that responses are generated by a human participant and if that participant meets the inclusion criteria for the study.  If your responses demonstrate characteristics of automated activity such as [PI may wish to add some study specific language here about the tools used to detect suspicious respondents], or if you do not meet the inclusion criteria, than the submission may be classified as invalid. Invalid submissions are not eligible for compensation.”

This guidance document supersedes those previously drafted.

Version Date: 8-27-2026

Related HRPP Manual Sections

None