The academic year has finished and maybe now is the time for some reflection. It has not been a good year for my academic discipline, social psychology. Since last August there have been three academic fraud cases detected, leading to the resignation of three professors, two in the Netherlands and one in the U.S.
After the high profile case of social psychologist Diederik Stapel in August 2011, which I discussed in a previous blog, two further cases were recently brought to light. The Belgian social psychologist Dirk Smeesters who worked as a Professor of Marketing at Erasmus University, the Netherlands, and, most recently Lawrence Sanna, an academic psychologist at the University of Michigan. Although details about the latter case are lacking, the two most recent cases were brought to light via the efforts of “data detective” Uri Simonsohn at Wharton Business School who has found a statistical method to detect suspicious data patterns. He tested this method successfully on data from Stapel’s papers and has since used it to hunt down other suspicious cases such as Smeesters, Sanna, and a fourth (not yet named) researcher. A paper containing details of Simonsohn's methods will be offered for publication soon (who dares to reject this paper?)
Is it sheer coincidence that these three high profile fraud cases all concern experimental social psychologists? It probably is. Dr. Simonsohn’s research interests are in social psychology and consumer behavior so it may not be surprising that his attention was drawn to papers within this field. To my knowledge, there have not been any cases of fraud in other branches of psychology this year. Yet this could all change if Simonsohn’s method is applied more widely.
Nevertheless, this might be a good time to reflect as somewhat of an insider upon the field of social psychology—which is a very strong and productive academic discipline—and its current academic practices to examine how they could be updated to develop the field in the future.
My first suggestion is to look very carefully at the role of deception in social psychology studies. Deception is relatively common in social psychology and the practice dates back to the origins of the field (Milgram, Asch). Previously, I suggested that this practice might lead to a more relaxed attitude towards unethical behavior. I have been criticized for making this connection which is fair enough because I don't have the data to show that there is a relationship between deception and data fraud. Yet what seems obvious is that the subjects who participate in our studies expect to be deceived and this affects the quality of data. On these practical grounds and on ethical grounds deception is not tolerated in related fields such as behavioral economics and cognitive psychology. So why should we?
My second suggestion is that we as social psychologists must work together in groups more. The strength of social psychology is also its weakness: Data can be gathered relatively easily by single individuals. Many social psychology studies do not require the same level of team effort as in social neuroscience or in behavioral medicine. That means that one person can gather and analyze data without having various other eyes watching. My understanding from the three cases is that they were gathering and analyzing data on their own, without the help of lab or research assistants or Ph.D.’s. In my view (and that of many others), it is not the task of relatively well paid full professors to collect their research data.1
My third suggestion is to improve the integration of our knowledge with that from other disciplines. To proof its existence, social psychology has evolved slightly independently from other fields in psychology and the other behavioral sciences like anthropology, biology, and sociology. As a result, there are very successful research programs in social psychology that have little bearing on what other fields are saying about the essentially the same phenomenon. Sometimes we make claims that do not make any sense when viewed through the lens of another discipline. An example is a study conducted by Stapel showing (supposedly) that eating meat makes you selfish. In this study (which has been discredited as the data were made up), the researchers primed participants with either a picture of a meat dish or a vegetable dish and then they asked people to fill out a questionnaire concerning their willingness to give money to a stranger. Those who saw the meat pictures were allegedly more selfish in their choices. Alarm bells would go off in other disciplines if you would hypothesize that meat makes you more selfish. Anthropologists (and primatologists) have shown that hunting and providing meat are highly social activities involving lot of sharing and generosity. As social psychologists we need to do a better job in connecting our findings to that of other fields to fully understand a phenomenon.
My fourth suggestion has to do with our desperate search for Black Swans. Social psychology seems to reward fnding exceptional phenomena. The social psychological literature on group decision making is an illustration. If you look at any textbook in social psychology, one would easily get the impression that humans are very poor decision makers in groups. We study group think, group polarization, risky shifts, information failures, and brainstorming problems as if they are the norm. Yet, a cursory review of the scientific literature on groups in other fields suggest that humans in fact make most important decisions within groups and do so highly efficiently. Maybe because social psychology is a relatively young, and densely packed field, there is a strong desire to predict something surprising, and make rather bold claims. It is extremely difficult to find support for very surprising hypotheses dealing with irregularities in human behavior. And, it is even harder to find evidence twice.
This brings me to the final issue about replication. Some journals in social psychology only publish papers with several replication studies. This practice is quite unlike what I have encountered with high impact journals in other disciplines like biology, economics, cognitive or evolutionary psychology that are keen to publish good single experiment papers. From personal experiences as an author (and as an editor, on the other side), it is very frustrating to be told to do a replication of what looks like a decent experiment with an interesting finding. The value of a replication study is questionable anyway if it comes from the same lab and the same researcher(s). My suggestion: Let other researchers carry out the replication after the interesting new finding has been published!
So, after a year of bad press, social psychology has lost perhaps a little bit of its very good academic reputation. For the field to bounce back from this and develop its considerable strengths we may want to review some current academic practices. No doubt this is already being planned by the various professional societies in social psychology (like SESP and SPSP) where researchers work tirelessly to improve the field and its public image.
1. Interview with Simonsohn in Nature News by Ed Yong (3 July 2012).
2. Wilson, D. S. et al. (2004). Cogntive cooperation: When the going gets tough think as a group. Human Nature.