The WSJ has this editorial:
Academics claim to revere open debate but often recoil when they see the genuine article. Witness the campaign some scholars--loosely defined--are waging against Heather Mac Donald for challenging university pieties about a recent surge in violent crime.* * *The American Society of Criminology claims to pursue "scholarly, scientific, and professional knowledge," but a better description of its priorities is one-sided inquiry and activist politics.
Earlier this year former ASC president Joanne Belknap of the University of Colorado at Boulder published a call for "criminology activism" among crime experts, citing their responsibility to "advocate for social and legal justice." The society's journal has since shifted toward papers that downplay particular crime trends and emphasize a policy agenda that opposes broken-windows policing or much attention to black-on-black violence. In the process it has abandoned its old role as a forum for a healthy clash of ideas.
We suppose this intellectual panic over Ms. Mac Donald is a tribute to the power of her persuasion and, we hope, of the Journal's editorial platform. But it's also a shame to see an academic group that ought to be a forum for scholarly debate descend into hackery.It is deeply corrosive to science for researchers to be motivated by a desire to advance a particular result rather than by a quest for knowledge as such. It is even worse if nearly all the researchers in a field favor the same result. It is much too easy, particularly in soft sciences such as criminology, to structure a study to produce the result one wants. Many decisions must be made along the way in a study, and if those choices are result-driven, the outcome cannot be trusted.
To take an extreme example, suppose there are 15 ways to crunch the numbers in a particular study. A result-driven researcher may crunch them all 15 ways, publish the one that supports his predetermined result and trash the 14 that do not.
How do we protect against this? Well, one thing we can do is require all researchers receiving government funds (which is practically all) to make available their raw data and not just their summary results. See this post by Neuroskeptic. Another thing we need is viewpoint diversity in academia. When schools are rated, that diversity should be examined, and severe one-sidedness should count as a strong negative factor.
Leave a comment