There have been some often quoted studies (e.g. McCabe & Castel, 2008) that showed that people will believe a scientific statement more easily when you include neuroscience. I wrote earlier this post about it. These studies were redone lately and the effect wasn’t found during the direct replication.
A series of highly-cited experiments published in 2008 demonstrated a biasing effect of neuroimages on lay perceptions of scientific research. More recent work, however, has questioned this bias, particularly within legal contexts in which neuroscientific evidence is proffered by one of the parties. The present research moves away from the legal framework and describes five experiments that re-examine this effect. Experiments 1 through 4 present conceptual and direct replications of some of the original 2008 experiments, and find no evidence of a neuroimage bias. A fifth experiment is reported that confirms that, when laypeople are allowed multiple points of reference (e.g., when directly comparing neuroimagery to other graphical depictions of neurological data), a neuroimage bias can be observed. Together these results suggest that, under the right conditions, a neuroimage might be able to bias judgments of scientific information, but the scope of this effect may be limited to certain contexts.
I always wondered if the non-replication was because of the research or because times had changed. Something this research also mentions:
“Although we were unable to replicate McCabe and Castel’s (2008) findings either conceptually or directly, this does not necessarily indicate that the original study was flawed; however, it does suggest that the biasing impact of neuroimagery is far less than was originally thought and, in fact, may be negligible. One potential explanation for these failures to replicate is that, as people have become more exposed to neuroscience over the past 5 years, the persuasive punch of neuroimagery has dulled, and what might have been legitimate effects in 2008 simply no longer exist. Unfortunately, testing this hypothesis would be quite difficult.”