From the Stanford prison experiment to Milgram, U-M Department of Psychology on reflect controversy in the field

Tuesday, September 11, 2018 - 9:19pm

In 1971, Philip Zimbardo, a psychology professor at Stanford University, constructed a fake jail in the basement of the university. He filled the prison with 21 college-age male volunteers, 10 designated “prisoners” and 11 designated “guards.”

Zimbardo planned to run a jail simulation, acting as the warden of the prison, for two weeks to observe the impact of roles and labels in the environment. The experiment quickly deteriorated, however, and the mock prison descended into chaos. After only six days, the experiment was shut down when a visiting student reported the abusive behavior of the guards and the psychological distress of the prisoners.

Zimbardo’s investigation, dubbed the Stanford prison experiment, would come to be hailed as a classic psychology study revealing one dark facet of human nature.

However, in June 2018, an article published in Medium claimed the Stanford prison experiment lacks credibility. The article said new interviews revealed the guards had been coached, and raised other issues having to do with the methodology and replicability of the study.

The Stanford prison experiment is not the only example of canonical psychology research coming under fire. A number of studies –– including the Milgram experiment, which was meant to demonstrate the disturbing extent of human obedience –– have received criticism for lack of replicability or sloppy procedures. Many of these studies are considered essential to the field and appear in standard psychology textbooks.

According to Howard Kurtzman, acting executive director of the American Psychological Association’s Science Directorate, the problem of older psychology studies being re-evaluated is not new but part of a continual trend.

“I see this as a gradual process,” Kurtzman said. “There are some findings that we’ve thought were true that maybe aren’t, but more commonly the magnitudes of the effects may be smaller than we realized.”

Kurtzman said a major issue is that many older studies have been difficult for researchers to repeat due to a variety of factors including small sample sizes and careless use of statistical methods. He clarified replicability problems affect all scientific disciplines, not just psychology. The National Academies of Science, Engineering, and Medicine is currently conducting a study on reproducibility in science.

Over the past five to 10 years, Kurtzman said, the psychology community has been working to address the replicability crisis, focusing on increasing sample sizes, sharing data and pre-registering studies to promote transparency.

Because of the attention being paid to replicability issues, Kurtzman believes research quality and reliability will improve within the next 10 to 20 years with journals and scientific funding agencies increasing their standards.

He estimated the field’s focus on high-quality research will begin to change teaching practices for psychology over the next five to 10 years.

“I think going forward research that is published in journals will be more reliable,” Kurtzman said. “All that will filter into teaching. I’m sure it’s being discussed already at the graduate level and seminars, and that will filter down, I think, to undergraduate teaching and eventually into textbooks.”

The replicability crisis has also become an area of focus for the University of Michigan Department of Psychology. Patricia Reuter-Lorenz, chair of the Department, said psychology faculty are well aware of the problem.

“I know that there’s been a lot of attention paid in smaller groups of faculty and graduate students in particular to the replication crisis and improving the rigor of research methods,” Reuter-Lorenz said. “We have several department-wide initiatives that are organized to achieve those goals.”

Though the department cannot dictate what professors teach, Reuter-Lorenz added, faculty are committed to training students to become good scientists.

“I don’t tell faculty they have to do things a particular way, but I think there’s enough of an appreciation for the importance of this that we have as a shared value that we will be rigorous in the methodologies that we teach,” Reuter-Lorenz said.

In addition to the replicability crisis, Reuter-Lorenz noted the University has been affected by the fact that standard studies like the Stanford prison experiment and Milgram experiment have been facing scrutiny.

“I think there’s some classic studies that have been part of many of our introductory classes and gateway classes that have come under fire, that have come under question, because of methods and ethics and things like that,” Reuter-Lorenz said.

Still, Reuter-Lorenz and Kurtzman said introductory psychology textbooks have not necessarily been updated to reflect controversies surrounding older, quintessential psychology studies. Kurtzman said whether older studies like the Stanford prison experiment remain in textbooks depends on the textbook author’s goal and approach.

“If a textbook author does address that study, they would be well-advised to include coverage of the controversies surrounding its methods and to address what it teaches us about the importance of institutional review boards,” Kurtzman said.

According to Reuter-Lorenz, this re-evaluation of older psychology studies has elicited a response from University psychology professors. She said faculty are generally aware of the issues with older studies and make sure to address them in class.

“Our instructors are cognizant of the importance of staying up-to-date on developments in the field, that’s why they’re here at the University of Michigan,” Reuter-Lorenz said. “Science is a work in progress. There’s always new discoveries and it’s very important that our faculty stay on top of those.”

In fact, Reuter-Lorenz added, the Society for the Teaching of Psychology published two papers in 2016 outlining how professors should cover studies with dubious methodology. One paper discussed the Stanford prison experiment while the other touched on the Milgram experiment. Reuter-Lorenz said one introductory psychology professor at the University has been using both to inform her teaching.

LSA senior Melissa Hall, who studies psychology, said older studies including the Stanford prison experiment, the Bobo doll experiment and the Genie experiment are still examined in introductory courses. Though Hall said she has not taken a psychology course that addressed the re-evaluation of these studies, her professors and classmates do discuss controversies surrounding older research.

“Often people will raise their hands and try and contradict the information and explain how there’s been controversy,” Hall said. “The professors would address it then, but they still use them as examples.”

According to Hall, many older psychology studies used methods that would not meet modern norms for scientific methodology or ethics. Those generational differences are typically addressed in her University classes, Hall said, though these older studies are still treated as valid examples.

“Many of these studies obviously took place many years ago, so we just kind of address the differences in generation and why it’s not all applicable nowadays,” Hall said.

Reuter-Lorenz said students can learn from discussing studies that are now considered flawed.

“There are better methods that can be used than were used in the past and they are available, and they are representative of best practices in the field,” Reuter-Lorenz said. “These, I guess you could call them errors, in the past are teaching moments for the present.”