Replication: Is it worth it?

Milgram Generator

Stanley Milgram in front of the famous “shock generator” used in his obedience experiments

 

 

 

 

 

 

 

 

 

 

I think just about every college psychology major wonders what would happen if someone replicated the original Milgram Obedience Experiment or any of Milgram’s explorations of obedience from the 1960s to the 1970s. Would I get the same results today if I encouraged a participant to punish a failing learner by applying electric shocks? Are there enough differences between the current generation and those individuals who participated in Milgram’s experiments that we’d see more or less disobedience? Is it even possible to get an Institutional Review Board (IRB) to allow a replication of Milgram, especially considering that Milgram’s obedience studies are often cited as a reason why government regulation of the research process is necessary (e.g., Jerry Burger’s APS IRB OpEd)? For decades, no one dared try. The closest anyone came was using a “virtual learner” (Slater et al., 2006) rather than the actual human confederate needed for a true replication (for a summary of other non-exact replication examples, see Miller, 2009).

In 2009, Jerry Burger published a partial replication of Milgram’s original work. He describes in a 2007 APS Observer article that he thought it would be impossible to get through an IRB. However, with some careful planning and extensive screening of participants, he was able to gain approval for his replication of Milgram’s 5th experiment. I think that most of us were surprised, nay “shocked” (e.g., Miller, 2009), upon discovering such an important replication was given the ethical “go-ahead” by an IRB. Perhaps more importantly, Burger’s (2009) work challenged what is traditionally seen as the role of scientific investigation, at least in one sense. “The most important characteristic of any new study or conceptual analysis is the degree to which it extends the line of inquiry, for example, by testing theoretical ideas, by suggesting new methods, or by posing further questions for research (or stimulating others to do so)” (Miller, 2009, p. 21). By replicating Milgram’s work, Burger (2007; 2009) has opened the metaphorical door to other researchers to explore obedience again, allowing decades-old questions and explanations to be revisited.  This is wonderful!

Perhaps more importantly than replication alone, Burger’s (2009) successful replication, at least in part, brought to prominence the idea that replication is important. Usually, replications are incorporated into existing work to establish that that work and it’s products are valid. However, considering that null results are often not published (Winerman, 2013), even many successful replications do not mean a study is valid (Francis, 2012). As long as we keep in mind that repeated successful replication (reliability) is not the same thing as validity, as my old colleague Rolf Zwaan suggests, most psychologists are behind, or even excited about, the idea of replication. As evidence of this, the Association for Psychological Science has initiated a replication initiative that offers funding to conduct registered replications.

This may be an excellent opportunity for both psychologists and the undergraduate students they supervise! Consider approaching your research mentor about potentially participating in a registered replication.

References

Burger, J. M. (2007). Replicating Milgram.  http://www.psychologicalscience.org/index.php/publications/observer/2007/december-07/replicating-milgram.html

Burger, J. M. (2009). Replicating Milgram: Would people still obey today? American Psychologist, 64, 1–11.

Francis, G. (2012). The psychology of replication and replication in psychology. http://pps.sagepub.com/content/7/6/585.full.pdf+html

Slater et al. (2006): http://www.plosone.org/article/info%3Adoi%2F10.1371%2Fjournal.pone.0000039

Winerman, L. (2013). Interesting results: Can they be replicated?  http://www.apa.org/monitor/2013/02/results.aspx

Zwaan, R. (2013). How valid are our replication attempts? http://rolfzwaan.blogspot.com/2013/06/how-valid-are-our-replication-attempts.html