More suggestive than conclusive

Uh oh. From The Role of Evidence in Politics: Motivated Reasoning and Persuasion among Politicians by Martin Baekgaard, et al. From the abstract:

Does evidence help politicians make informed decisions even if it is at odds with their prior beliefs? And does providing more evidence increase the likelihood that politicians will be enlightened by the information? Based on the literature on motivated political reasoning and the theory about affective tipping points, this article hypothesizes that politicians tend to reject evidence that contradicts their prior attitudes, but that increasing the amount of evidence will reduce the impact of prior attitudes and strengthen their ability to interpret the information correctly. These hypotheses are examined using randomized survey experiments with responses from 954 Danish politicians, and results from this sample are compared to responses from similar survey experiments with Danish citizens. The experimental findings strongly support the hypothesis that politicians are biased by prior attitudes when interpreting information. However, in contrast to expectations, the findings show that the impact of prior attitudes increases when more evidence is provided.

While they are focused on politicians, I suspect the finding is not restricted to politicians. It would be interesting to know the results for the control group of regular citizens.

The implication of the research is interesting though. At what point do people become epistemically closed? Is it by age? By consistency of information up to a point of decision? By motivation (countervailing incentives to believe something as opposed to simple cognitive inertia)?

I have the impression that there are a material number of things about which, with changed information, I have changed my assessment and changed my consequent behaviors. I would have assumed most people do. Am I being self-deceptive? Or is there something definitional and contextual going on?

As an example, for years I would have argued with reasonably good evidence that parental influence, and not just genes, were instrumental in childhood outcomes. I accept, with some asterisks, the evidence that genes are a stronger predictor of childhood outcomes than we used to accord. But many take a strong position that all childhood outcomes are due to genes and non-shared environment (outside the home); parents don't matter.

I natively want to believe otherwise but I can see the evidence as it exists. However, I suspect we are not defining the boundaries between shared environment and non-shared environment with great enough specificity and we are not paying enough attention to developmental sequencing.

More specifically, I suspect parents actually do exercise more substantial influence over their children's life outcomes, not via the time spent in the home (shared environment), but rather in all the attendant decisions about a child's environment. Where they live, the neighbors to whom they are exposed, childhood friendships that are encouraged or discouraged, school attended, church community involvement, sports, etc. Parents create, consciously or unconsciously, a developmental ecosystem around a child, most of which parents may not be, and usually are not, involved in directly. My suspicion is that by developmental ecosystem choices, parents are in fact shaping their children's life outcomes but doing so indirectly.

I think the level of our research sophistication is at this point still undeveloped and that when we begin taking into account indirect parental influence, we will find, counter to the current research, that in fact, much of the variance in life outcomes will be attributable to parental decision-making.

Is that motivated reasoning or is it rationalizing away from the nascent evidence. No way to know at the moment but we will know in time.

So with the Danish politicians. Is more evidence against their prior position simply causing a n instinctive digging in of the heels? Is it that, depending on what the sample issues were or the quality of the evidence being offered, that there is a design flaw in the study? Is it that there are other motivations in play (a man more easily discerns which his bread is buttered than the quality of evidence to an abstract issue)? Who knows.

The study is interesting but more suggestive than conclusive.