How to Save Honesty in Human Subject Research

In human subject research, we often face an ethical question: is it ever justifiable to deceive participants? After all, deception can be effective in getting unbiased data in studies where the awareness of the experiment’s purpose is likely to change how participants behave. Yet, there is a deeper ethical dimension here. My recent paper, Honesty in Human Subject Research, tackles an essential question: how can we make human subject research more honest without undermining its scientific validity?

The ethical problem of deception is hard to address mainly due to its two features: irreversibility and opaqueness. First, once the deception takes place, it is hard, if not impossible, to undo it. Second, the subject’s willingness to participate is, in principle, opaque to the consent-seeker unless the latter explicitly asks for consent. Given such difficulties, addressing the ethical issues concerning deception is becoming increasingly important for researchers, especially those in human subject research such as psychology and medicine.

Milgram’s obedience experiment is a famous example. In this experiment, Milgram collected participants’ unfeigned responses to an authoritative figure’s order by hiding that the button they pushed did not actually give the confederates any electrical shock. Also, in some clinical trials, participants are misled about the true nature of the drug being tested, its potential side effects, or the likelihood of receiving a placebo. Researchers do this to ensure unbiased reporting of effects and to study the placebo effect.

The Core Idea: Respect for the Right Not to Be Deceived (RND)

While it may be hard to eradicate deception in human subject research, we should at least try our best to secure the virtue of honesty. We should first understand what honesty is. It is not simply a matter of telling the truth. At the heart of honesty lies the respect for what I call the right not to be deceived (RND). Respecting participants’ RND involves not deceiving them unless there are overriding reasons. This doesn’t imply that no deception should be allowed in human subject research, but it calls for a closer look at why and when we use it. Respect for RND pushes us to think more critically about each instance of deception in our study designs.

Avoiding Dishonesty: ‘After Participation’

In my paper, I outline the limitations of standard practices aimed at managing deception. I divide them into ‘after participation’ and ‘before participation’ phases. The former involves compensating, expressing fitting attitudes, and debriefing. While these methods are helpful in some respects, they have limitations in addressing the ethical issues of dishonesty.

Take compensating participants, for example. Disrespecting someone’s RND would be wronging that person, as opposed to merely harming her, without overriding moral considerations. A wrong committed to a person is not simply neutralized by some benefits given to that person to offset the harm done. The attempt to simply compensate materially after disrespecting the participant can be perceived even as an insult.

The researcher may also try to apologize or thank the participant after the research. However, showing such attitudes after conducting the deceptive research in question does not manifest respect for the subject’s RND, as it is to be deemed insincere. Doing what one would apologize for later shows that one values doing what one plans to do over being honest to the subjects anyway. Also, it would be absurd to express gratitude to the subject for participating in deceptive research, especially one that she would not have participated if only she had known about its deceptive elements.

Now consider debriefing. Debriefing at the end of a study is commonly used to explain to participants what really happened and why deception was necessary. However, one cannot retrospectively consent to the deception that has already happened. Debriefing is neither a way of respecting the subject’s RND nor a way of making up for the disrespect to her RND, insofar as she has been intentionally deceived.

Avoiding Dishonesty: ‘Before Participation’

Reasonable Expectation

Then, we need to consider what we can do before the research to avoid dishonesty. One thing we can do is to make the research plan based on a reasonable expectation of the possible harm (e.g., mental distress) of the subjects or the likelihood that the deception would affect their willingness to participate.

However, even if the deception had no negative effect on the subject’s feelings or willingness to participate, it may still manifest disrespect for her RND. In general, a right may be disrespected even when the right-holder is not aware of the violation or does not feel any discomfort or emotional distress when one finds out the fact that one’s right is violated or disrespected.

The permissibility of a deceptive study depends on whether it involves objectionable kind of disrespect for the participant, rather than on how she subjectively takes it. In this sense, disrespecting the participant’s RND can be ethically problematic regardless of how she subjectively experiences it.

Consent to be Deceived

Now, what’s left? We may ask for consent to being deceived. But is it ever possible for one to be deceived while knowing that one will be deceived? Dave Wendler and Franklin G. Miller say it is. They call the deception based on the deceived person’s consent authorized deception. The basic idea is to ask for the participants’ consent after informing them that some deception may be involved in the study without giving details. This seems promising, since one can get informed consent without losing the participants’ candid responses. However, even if the subject is prospectively informed and warned about the deception involved in the study, the subject still may not know sufficient details about the deception.

So, I suggest two possible strategies to address this issue.

1. Disjunctive Informing

Disjunctive informing allows participants to give informed consent without knowing every detail that could impact the study’s results. In placebo-controlled studies, for instance, rather than telling participants “This is the real medication,” we might say, “This could be either an active medication or a placebo.” This strategy can allow the subject to examine if she is okay with each disjunct without revealing which one will be applied to her. Also, it can ensure unbiased reporting of effects without involving any deception.

 2. Domain Specification

Another strategy is domain specification, which specifies the domain of deception rather than its contents. For example, a researcher might say, “Some elements of the study might involve misleading cues about perceptual stimuli to assess reactions,” without disclosing every detail. To give an analogy, in playing poker, one can state that one may deceive other players about the combination of the cards one is holding but not about whether there is any extra card in one’s sleeve. This strategy allows the participants to give consent within a specific domain of potential deception without revealing details that would compromise the study.

Conclusion: A Call to Elevate Honesty in Human Subject Research

In short, this is a call to action for researchers to rethink how we use deception in studies involving human participants. If we can rethink how we incorporate honesty into our research designs, we might find ways to reduce deception without sacrificing scientific rigor. Researchers, IRB members, and research institutions should work together to implement approaches like disjunctive informing and domain specification as part of a broader effort to respect the right not to be deceived.

We need honesty in any realm of human interaction, and human subject research is no exception.