By Chris Bjornsen
Bjornsen is a professor of psychology at Longwood University in Farmville.
A landmark study published in the most recent issue of Science, one of world’s foremost scientific journals, revealed a shocking statistic: Less than half of a select group of published psychological studies could be replicated, showing the same results, by separate teams of scientists.
The study, led by Dr. Brian Nosek at the University of Virginia, involved more than 250 other social scientists and replicated or repeated 100 previously published studies in order to verify the results of the original research. What they uncovered speaks to a fundamental paradox present within the academic community, and one that we must all work to resolve.
A famous example of the replication process gave us valuable insights into what we call the self-fulfilling prophesy. In the early 1960s, Robert Rosenthal and colleagues told a number of elementary school teachers the names of a few students in their classes who were academically gifted, although these children had actually been chosen at random from the class. At the end of the year, the “gifted” children showed larger gains in IQ scores than their classmates, and teachers described them as happier, more curious, more interesting, and likely to be more successful in life than the other children.
This was a fascinating result, and demonstrated that teacher expectations for student abilities can be shaped in ways that actually leads children to become smarter. The results prompted other researchers to conduct the same or similar studies in other schools. The replication studies, however, did not all produce the same results. What they did reveal, however, was a stable, repeatable effect: If teachers’ beliefs about children were shaped before they had much contact with the students, the effects were basically the same as in the original study. By studying the same process over and over, social scientists demonstrated when and under what conditions we could rely on this Pygmalion effect.
Science is often a long, laborious process designed to improve our understanding of each other and our world, to address problems, and to improve our lives. Along with the replication process, other procedures are used to ensure the objectivity of research and the interpretations of the results.
This critical analysis of previous research, with students and with colleagues, is essential to discovering “the truth” in science. However, in doing so, we often unwittingly contribute to the dilemma highlighted by Nosek’s seminal work.
Science, almost by definition, is a forward-looking endeavor. To spend time doing the same work others have done can be viewed as a stagnant endeavor, rather than a dynamic, progressive contribution to the field. One might conduct or learn about a replication study, obtain results that mimic the original research, and think, “But we already knew that.”
This very same view of replication work is not uncommon among the editors and reviewers of our scientific journals, as well as university tenure committees and administrators, who tend to place higher value on original research. To underscore this point, according to a top database of publications in psychology and related fields, 100,405 empirical peer-reviewed studies were published in 2014. Only 475 studies — one-half of 1 percent — were also indexed as experimental replications.
Obtaining and keeping one’s job in academia is concretely tied to research productivity and publication; hence the motivations for researchers to steer clear of replicating the work of others.
The paradox is, unless previous work is replicated, we fall short in our attempt to know whether we have revealed the true nature of the human condition. What we always need to see is a pattern of results from similar studies or actual replications, and then still be open to the possibility that the next five or 10 studies that are conducted on the same topic might give us completely different results.
We communicate to students, and should make clear to the public, “one study never proves anything,” and we also need journal editors, tenure committees and university administrators to embrace our effort to demonstrate we are as certain as we can be about our explanation of the nature of our world.
Humans, and the contexts and cultures in which we live, are neither identical nor static; both are constantly changing. Objectivity and certainty in science, especially social science, are moving targets by nature, although we do have ways of keeping them more or less in our line of sight. We simply need to raise the value we place on being able to demonstrate we can hit those targets with more than just one throw.