A Path to Redemption? Benevolent Scientists Admit Research Flaws and Mistakes
Scientists may have a public perception problem.
A handful of recent studies suggest that scientists have a rather complicated public image, at least among some Americans. Overall, U.S. adults appear to respect science as a profession and view scientists as highly knowledgeable and competent in what they do (e.g. science and research). The problem comes when we ask these same people to evaluate scientists’ warmth, a broad trait linked to morality, honesty, helpfulness, sincerity and sociability. A 2014 study by Susan Fiske and Cydney Dupree concluded that scientists are viewed as coldly competent.
Some science communicators have suggested that scientists’ low rating on warmth as compared to competence means that scientists do not have the level of public trust that many of us (including me) assumed had remained relatively stable over the years. Some science communicators might, knowing this, question whether encouraging scientists to talk about their own research and science in the public sphere is the best thing to do.
To be transparent, I tend to be more on the optimistic side about the state of public trust in scientists.
“They [including scientists and researchers] earn respect but not trust. Being seen as competent but cold might not seem problematic until one recalls that communicator credibility requires not just status and expertise (competence) but also trustworthiness (warmth).”
Fiske and Dupree did acknowledge in their paper that “scientists whose job involves teaching and communicating may seem warmer and more trustworthy.” But how exactly does science communication relate to U.S. adults’ perceptions of scientists’ warmth, morality or honesty?
A study published in Studies in Communication Sciences in late 2016, titled "Disclose your flaws! Admission positively affects the perceived trustworthiness of an expert science blogger," offers some insights related to scientists’ perceived trustworthiness when they practice science communication online. The “online” piece is important, because the internet is now Americans’ top source of science news and information.
Psychology researchers at the University of Muenster in Germany (Hendriks, Kienhues and Bromme, 2016) investigated laypersons’ perceptions of the trustworthiness of a scientist who blogged about his* research findings. They specifically explored how roughly 100 high school students judged the expertise, integrity (related to honesty and openness in communication), and benevolence of a scientist who blogged about his research and admitted a flaw or limitation. They compared this to a case in which the scientist did not admit the flaw, but instead another expert pointed out the same flaw or limitation in a comment on the scientist’s own blog post.
*Note: My first issue with this paper is that the made-up scientist in the study was male, and that the authors did not account for potential gender differences. In other words, could a female scientist be judged more critically for admitting a study flaw or limitation? Based on some previous research, I wouldn't be surprised if this were the case.
Illustration/Photo by Philip Bond, Flickr.com
“Following [a] blog entry, a comment was displayed, introducing the flaw. In the comment, it was said, that the sample was too small to come to such optimist conclusions and more studies would be needed to assess the benefits and side effects of the drug. The source of the flaw’s disclosure was varied in two experimental conditions: In the admission condition, the expert blogger disclosed the flaw himself. In the critique condition, another scientist disclosed it. In the no-disclosure condition, no additional comment was added to the introductory blog entry.”
Hendriks, Kienhues and Bromme found something curious. Both admitting a flaw and having another expert point out the flaw did hurt the scientist’s perceived expertise, where expertise ratings were provided by high school students reading the made-up blog post. But admitting the flaw in his research made readers score the scientist higher on integrity and benevolence. In other words, readers rated the scientist’s behavior as being more moral and considerate when he admitted the flaw.
Note: In my opinion the study authors should have showed a table of statistical comparisons between all three conditions (no disclosure of the flow, self- admission of the flaw, and critique by another expert) with Bonferroni correction for multiple comparisons. It seems like comparing the admission and no-disclosure cases directly would be valuable. From simple t-tests, these comparisons appear to be significant for benevolence and integrity scores. In other words, it appears that admitting a flaw significantly boosted the scientist's perceived benevolence and integrity compared to not admitting the flaw at all. But I'm not sure how this comparison stands up to adjustment for multiple comparisons in the raw dataset.
There’s weight to the argument that scientists’ can improve public trust by talking about the uncertainties and limitations of scientific research, something that science communicators tended to avoid in the past. A study published in 2008 found that both scientists and journalists may be viewed as more trustworthy when they report study limitations within news coverage of scientific research, such as cancer research. Other research highlights the benefits of scientists talking about the uncertainty inherent in science and of doctors admitting medical errors directly to their patients as soon as these errors or flaws are discovered.
“In a survey study, Fisseni, Pentzek, and Abholz (2008) found that patients preferred to stay with their physicians when these disclosed a severe flaw themselves but not if the flaw had been disclosed by other physicians. Admission of flaws restored patients’ trust whereas critiques by unaffiliated experts demolished patients’ trust in their physicians.”
Is admitting research mistakes, flaws and study limitations in public a good idea for scientists? In terms of boosting laypersons' evaluation of how warm, honest and benevolent scientists are, the answer might be, yes! This idea makes me think about the trend of researchers sharing #overlyhonestmethods on Twitter.
An example of an #overlyhonestmethods tweet.
So scientists, disclose thy flaws?!
Have questions, comments or concerns? Leave them below or ping me on Twitter.