Throughout the year, we will be publishing essays from Professor Andrew Lyon‘s Honors 389 course “The Science Blender” . For the first paper this term, students were asked to investigate the idea of  “science gone wrong”.  Below is one student’s essay  on the topic.

It is all too easy to blindly trust scientific research, particularly when it is published in big-shot journals like Science and Nature. I get it. When I was growing up, my parents were always sending their research to these prominent journals and waiting for a reply, as if getting published would be like Rocky fist-pumping on top of the Philadelphia Museum of Art steps. It’s a pretty big deal.

So, recognition in the scientific community can become incentive to publish false data. In December 2014, Michael LaCour and Donald Green’s “When Contact Changes Minds: An Experiment on Transmission of Support for Gay Equality” was published in Science. In their paper, LaCour and Green attempted to show if gay or straight messengers have greater potential in effecting voters’ opinions toward same-sex marriage through active contact. The data collected shows that gay messengers were more effective at changing voters’ beliefs for a longer period of time. LaCour and Green concluded that a 20-minute face-to-face conversation between a gay messenger and a voter can change the voter’s attitude for up to 9 months.

LaCour and Green’s paper met high media coverage almost immediately after publication. It received even more attention when Broockman, Kalla and Aronow’s “Irregularities in LaCour” became public. These three graduate students found numerous anomalies in LaCour’s work and questioned the integrity of his data. In their paper, Broockman, Kalla and Aronow outlined a total of eight irregularities along with their own replication of Green and LaCour’s protocol. Additionally, when Green requested raw data from LaCour, LaCour did not reveal his records. Marcia McNutt, editor-in-chief of Science, included an editorial retraction with the concurrence of Green in journal’s June addition. McNutt wrote that Green and LaCour falsified sponsorship and could not authenticate claims with original findings.

On May 19, 2015, Green submitted an official retraction letter to Science where he wrote, “the…concerns noted…undermines the credibility of the findings. I am deeply embarrassed…and apologize to the editors, reviewers, and readers of Science.” LaCour still does not agree to this retraction.

Donald Green and Michael LaCour’s study is one of many instances where science goes wrong. Regardless of whether data is faked or simply inadvertently mistaken, the outcome remains the same: science is not always reliable. It is more the authors’ responsibility than the journals’ to thoroughly review works before publication. It is not always possible to tell when the peer review system fails, which is why we must be ever-vigilant to what we read in our world of instant knowledge. In the end, it’s up to the reader to weed out genuine knowledge that can help advance human knowledge and understanding of life itself.

About Honors 389: The Science Blender

Throughout Professor Lyon’s Science Blender course, students identify knowledge gaps that prevent simple solutions to grand challenges, and then they develop strategies to address those gaps. In this process, students also become more conversant in the languages of different disciplines and how team-based problem solving benefits scientific pursuits.