Ethical Concerns in Social Psychological Research

Most people would want to conduct research geared toward changing people’s attitudes only if they believed the direction of change was for the better. We wouldn’t support research that might have the effect of encouraging people to engage in unhealthy or dangerous behaviors. If there is a suspicion that a study might have harmful effects on some participants, it should be reviewed by people unconnected with the research. Research conducted at universities that has the potential for harm has to be approved by an institutional review board (IRB), a committee that examines research proposals and makes judgments about their ethical appropriateness. An IRB includes at least one scientist, one nonscientist, and one person who is not affiliated with the institution. If some aspect of the study’s procedures is deemed overly harmful, that procedure must be changed before the research can be approved. Before 2018, all social psychological research conducted with federal funds had to be submitted to an IRB. Now investigators are not required to obtain IRB approval if a study involves only a manifestly “benign intervention,” such as ordinary interviews and surveys, ability and personality tests, economic games, or tests of basic cognitive and perceptual processes.

Proposed research may be approved even if it makes people uncomfortable or embarrassed or causes physical pain—as long as the research is deemed sufficiently likely to yield scientific information of significant value and the discomfort or harm to the participants is not too great. For example, the Milgram studies on obedience (1963, 1974) were conducted before IRB committees existed. Today, Milgram’s proposal would be thoroughly examined by an IRB, and it’s not clear whether it would be approved. On the one hand, there is no question that Milgram’s research made some participants extremely uncomfortable; their psychological distress was manifest to observers. On the other hand, many (if not most) people would consider the knowledge gained to be enormous. It’s impossible to think about Nazi Germany the same way after learning the results of the Milgram studies. And we can no longer blithely assume that ordinary, decent people would refuse to obey commands that are patently harmful. Different IRBs would undoubtedly reach different conclusions about the admissibility of the Milgram experiments today. What do you think? Would you permit research like Milgram’s to be conducted? (We discuss a couple of IRB-approved modifications to Milgram’s procedure in the discussion of obedience in Chapter 8.)

Informed consent—a participant’s agreement to participate after learning about all relevant aspects of the procedure—is required for that small fraction of social psychological research that poses the possibility of significant harm. However, for certain types of studies, known as deception research, it’s not possible to obtain informed consent from the participants. Darley and Batson (1973) couldn’t have told their seminary participants that the stated reason for their need to hurry was bogus and that the apparent person in need was actually a confederate who was merely acting as part of the experiment. Informed consent would have defeated the purpose of the study. Even in deception research, however, participants are generally told about the goals of the research project afterward during debriefing. The debriefing session serves an important educational purpose: It informs participants about the broad questions being addressed, the specific hypotheses being tested, and the potential social value of the results. Debriefing can also be helpful to the investigators by letting them know whether participants are interpreting the stimuli used in the experiment as the investigator intended.

When asked their opinion about what they were put through, deceived participants generally understand the reasons and often say they learned more, and enjoyed the study more, than subjects who were not deceived or made uncomfortable (S. S. Smith & Richardson, 1983). For example, participants in the insult condition of the Cohen study actually reported that they learned more and had a better time than participants in the control condition.

When participants have been deceived or made uncomfortable, experimenters owe them a full accounting of what was done, what aspects of the procedure involved deception, why they were made uncomfortable, what the experiment was intended to examine, and what potential value to society the research might provide.