IRB: Ethics & Human Research

When Is It Ethical to Withhold a Research Incentive?

Researchers want to conduct an online survey to determine respondents’ attitudes toward and concerns about uses of deidentified tissue in research biobanks to see whether their concerns affect their willingness to donate a tissue sample under a blanket consent to future research uses. Respondents will be asked to react to several hypothetical scenarios in which they are asked to donate tissue to a biobank. Follow-up questions inquiring about the reasons for their decision to donate or not will be determined by their answers to previous questions. In addition, a variety of demographic data will be collected, along with answers to questions regarding their attitudes toward medical research. The survey does not ask sensitive or stigmatizing questions and will not collect identifying information. It is estimated to take 15–20 minutes to complete.

Respondents will be given a $10 gift certificate by e-mail on completing the survey, and this incentive will be mentioned in recruitment materials. Online surveys are particularly vulnerable to being completed multiple times by the same person, and the researchers are concerned that the incentive may motivate some people to try this. To identify such attempts, the researchers require persons interested in taking the survey to contact the research assistant to receive a unique URL link to the survey by e-mail. When they complete the survey, respondents will provide an e-mail address to receive an authorization code and instructions for redeeming the gift certificate. The researchers will keep a database of these e-mail addresses. When new e-mail inquiries are received, the e-mail addresses will be compared with those in the database. If an e-mail address has already been used by a respondent, a survey link will not be sent. When the survey is closed, the database of e-mail addresses will be destroyed.

This is not a foolproof method of preventing individuals from taking the survey multiple times. Individuals often have several e-mail accounts or can easily create new accounts for this purpose. The researchers want to further discourage multiple attempts by making the collection of additional gift cards more onerous. To do this, the researchers propose to make the gift card available only at the end of the survey, which respondents can reach only by answering all the survey questions. Respondents who decide not to answer a question will not reach the end, and so will not receive a gift certificate. These conditions are spelled out in the project consent document, which respondents will read and accept before entering the survey.

The project is submitted for expedited review to the local institutional review board (IRB). Some IRB members reject the plan to provide a gift card only to respondents who answer all of the survey questions and put forward several arguments against it. Are these arguments against the proposed plan for the gift card persuasive?

Withholding the Gift Card Is a Penalty and Thus Coercive

The IRB members first point out that according to federal regulations regarding informed consent requirements (45 CFR 46.116 [a] [8]), individuals recruited to participate in research must ordinarily be told that “participation is voluntary, refusal to participate will involve no penalty or loss of benefits to which the subject is otherwise entitled, and the subject may discontinue participation at any time without penalty or loss of benefits to which the subject is otherwise entitled.” This provision reflects the Belmont Report’s insistence that a valid consent “requires conditions free of coercion.”1

The IRB members argue that by withholding the gift card from participants who do not answer all of the survey questions, researchers are penalizing those participants and thus violating their right to be free of coercion. If someone receives a traffic ticket and pays a fine, the fine is a penalty: the driver has less money after paying the fine than had she not broken the traffic laws. By analogy, IRB members argue that the researchers’ plan to withhold the gift card will also exact a penalty: participants who don’t complete the survey will have less money than had they answered all the questions and received the gift card. But this argument misunderstands the nature of a “penalty” and the nature of “coercion”—misunderstandings that are closely related.

Let us start with the concept of a penalty. In the case of the traffic ticket, the fine is a penalty because it is levied against funds to which the person already has a rightful claim. Unless she can get someone else to pay the fine, the driver will have to use her own money. Those who don’t complete the survey will also have less money, but they won’t have less oftheirmoney. They will have less of the project’s money.

But what about the additional clause in the federal regulations, “or loss of benefits to which the subject is otherwise entitled”? How does it apply in this case? The clause certainly provides some further elaboration of the concept of “penalty.” Someone can be penalized by the loss of goods in their possession, as in the example of the traffic ticket. Or she can be penalized by the loss of future goods that rightfully would have been hers. An example would be a military veteran whose access to promised health services from the Veterans Administration was made contingent on her agreement to participate in clinical research. Such a person would not only have been penalized; she would also have been wrongly coerced into her agreement to be a research subject. As Wertheimer and Miller argue, A wrongly coerces B’s agreement when A is threatening to violate B’s rights or not meet obligations owed to B in order to make agreement the only reasonable option left for B.2This is exactly what would have been done in the veteran’s case.

With regard to the biobank survey, the critical question is whether participants who don’t complete the survey are entitled to the gift certificate. The answer is no, because the conditions of entitlement are set out in the consent document, and the participant has agreed to them. Had the consent document promised a gift certificate regardless of how much of the survey was completed, then it’s true that participants would be penalized if the researchers later threaten to withhold the incentive until all items are answered. But that’s not what the researchers are proposing. Individuals will have freely decided whether to take the survey under the conditions described. Those who then answer questions in order to receive the gift certificate will also do so freely, without coercion, because failing to do so leads only to a consequence of their voluntary choice, not to a penalty wrongly imposed on them.

Withholding the Gift Card Is an Undue Influence

Even if it’s not coercive, the IRB members claim that withholding the gift card exerts an undue influence on prospective research participants. Both 45 CFR 46 and the Belmont Report forbid the use of undue influence when an individual is deciding whether to enter a research project or remain in it.

The IRB members argue that an incentive exerts an undue influence unless it is a form of reasonable compensation, either for expenses that research participants incur (e.g., gas to travel to the research site) or for time spent in the study. They also claim that the incentive to participate would not constitute an undue influence if participants received a sum proportional to the amount of the survey they completed—say, $10.00 for completing the whole survey, $7.50 for completing 75%, or $5.00 for answering half of the questions.

This reasoning is problematic on several levels. First, it is based on no substantive account of “undue influence,” which IRB members characterize only in terms of what it’s presumably not. Thus, on their reasoning, they can’t explain when compensation itself is a form of undue influence—i.e., when it’s “unreasonable.” The question why a $10 compensation isn’t itself an undue inducement to complete 100% of the survey is not a question the argument can answer.

Second, their reasoning fails to understand ethically important differences between “compensations” and “incentives.” As Grant and Sugarman point out, incentives are not a form of compensation. Compensation redresses a loss. An incentive is instead a “benefit designed as a motive or incitement to action.”3A year-end bonus is an incentive, not a compensation. Grant and Sugarman note that unlike compensation, a bonus or other incentive can’t be reasonable or unreasonable, fair or unfair. Incentives are only effective, or not. Of course, there is nothing in the nature of incentives that precludes their exerting an undue influence. So the next question is whether the gift card incentive does so.

A tempting first response is to ask a rhetorical question: if $10 isn’t an undue influence on the decision to start answering survey questions (i.e., to enter the study), how could it be an undue influence on a decision to continue answering survey questions? However, this riposte makes the glib assumption that the two decisions are similar in all relevant ways. They may not be: the first decision is made in ignorance of the specific nature of the questions in the survey, and the second is not. If the nature of the questions turns out to be ethically relevant to a judgment about undue influence, then an incentive could exert an undue influence on the decision to continue, but not the decision to start.

This is not the place to undertake a comprehensive analysis of “undue influence.” The Grant and Sugarman article referenced earlier, however, attempts to provide one. The most pertinent of their criteria is that an incentive doesn’t exert an “undue” influence unless it leads people to act against their principles or contrary to their best interest. This is compatible with Wertheimer and Miller’s claim that an inducement is “undue” only when it predictably leads to irrational decisions. However one describes it, such behavior suggests that the individual’s capacity to make an informed and competent choice has been compromised by the offer, and that this is what makes the inducement wrong.

In being asked to answer all questions in order to receive the gift certificate, are survey respondents induced to do anything that might lead them to act against their principles or against their best interest? It’s difficult to see how. None of the questions touch on morally sensitive matters that respondents might have qualms about answering, or that might lead to discrimination or other harm to them if the confidentiality of their answers was breached.

The researchers’ plan, then, is ethically permissible. It neither exacts a penalty against participants’ right to withdraw from the study, nor coerces them to complete the survey, nor creates an undue influence that works to their detriment. Thus, IRBs should approve a project and recruitment incentive like the one described here.


I would like to thank the journal’s anonymous reviewers for additional references and helpful suggestions.

Tom Tomlinson, PhD,is Professor of Philosophy; Director of the Center for Ethics and Humanities in the Life Sciences; and Codirector of the Bioethics, Humanities, and Society Program at Michigan State University, East Lansing, MI.


1. United States National Commission for the Protection of Human Subjects of Biomedical and Behavioral Research.The Belmont Report: Ethical Principles and Guidelines for the Protection of Human Subjects of Research. Washington, DC: U.S. Department of Health, Education, and Welfare, 1979.

2. Wertheimer A, Miller FG. Payment for research participation: A coercive offer?Journal of Medical Ethics2008;34:389-392.

3. Grant RW, Sugarman J. Ethics in human subjects research: Do incentives matter?Journal of Medicine and Philosophy2004;29(6):717-728.