Wednesday, 5 February 2020
Blood-glucose levels influence judges’ decisions less than we think
Science is based on empirical facts, such as data gathered in experiments. But it is also based on the interpretation of these facts, what they mean, what may have made them possible. (In scientific articles, the data—the facts—are presented in the “Results” section and interpreted in the “Discussion” section.) It should be no surprise that in any given field of science, some scientists may disagree with the way that their colleagues have interpreted certain results. Such disagreements arise in all scientific disciplines, especially in psychology, and especially when the data are very clear-cut or the correlations are very strong. And it’s just that kind of a case that I’d like to discuss in this post today (maybe in some future post, I will discuss the broader results-reproducibility crisis that has been shaking psychology for some years now).
The case I’m talking about is a famous experiment that has received a great deal of coverage in reputable scientific media: “Extraneous factors in judicial decisions”, published by Danziger, Levav and Avnaim-Pesso in 2011. This study presented some surprising data about the cognitive abilities of judges who make decisions about paroling prisoners. As the above graph shows. the percentage of decisions granting parole was about 65% at the start of the day, but curiously fell to almost zero before the judges’ first meal break. After the break, the percentage rose right back up to about 65%, but fell again to nearly zero a few hours later, just before the judges took another meal break, after which the pattern repeated.
Because many past studies had found that lower blood glucose levels might interfere with rational judgment associated with activity in the prefrontal cortex, Danziger and his co-authors interpreted these findings as evidence of what psychologists call ego depletion, meaning that the lower glucose levels in the judges’ brains impaired their judgmental abilities, and that rather than risk making bad decisions and paroling people who were dangerous, they erred on the side of caution and left more people in prison.
This interpretation that these authors suggested to explain their striking findings has been criticized in many ways since their study was published. First of all, the theory of ego depletion has itself been called into question in many more recent studies. Second, other studies have re-analyzed the approach taken by Danziger et al. and have proposed alternative explanations.
One such study, entitled “Overlooked factors in the analysis of parole decisions”, by Keren Weinshall-Margel and John Shapard, also was published in 2011. After interviewing corrections personnel, Weinshall-Margel and Shapard criticized certain aspects of the methodology of Danziger et al., such as their having failed to distinguish cases where parole was refused from cases that were deferred to a later date, which Weinshall-Margel and Shapard regard as two very different things. They also pointed out other aspects of the way parole hearings are held that might have contributed to the observed drop in parole approvals. For example, they suspected that attorneys who present several cases at the same hearing present their best cases first and save their weakest cases for last. (Danziger et al. did, however, address these points in a response to Weinshall-Margel and Shapard’s critique.)
Another article criticizing the Danziger study was published by Andreas Glöckner in 2016. Its title was “The irrational hungry judge effect revisited: Simulations reveal that the magnitude of the effect is overestimated”. After investigating and conducting various simulations, Glöckner too found that some aspects of the parole-decision procedure may have had a non-negligible influence on the observed effects. For example, to manage their workload within a given block of time, judges tend to start with the cases that involve the most material and hence contain the most elements favourable to the parole applicants. But even Glöckner admits that these factors intrinsic to the parole-decision process can explain only part of the sharp decline in favourable decisions observed by Danziger et al. Although for Glöckner it is clear that the contribution of the decline in blood glucose in the judges’ brains was overestimated in the original study, he doesn’t discount the possibility that it may have contributed to the observed effect, albeit to a lesser extent.
Lastly, in a 2017 blog post, Daniël Lakens says that the cognitive effect observed in the Danziger study was simply too large to depend on biopsychological factors alone. He argues that if such potentially catastrophic effects could be attributed to hunger alone, they would already have been observed just before mealtimes in all professions. But this is not really the case, and it is rather sleepiness after a big meal that impairs thinking, because the body is too busy digesting—a phenomenon with which all of us are familiar, as we are with many other manifestations of mental fatigue (due to having to perform the same task repeatedly, or a very difficult task, or a task in the presence of continual distractions, for example). But just like any psychological manifestation, these are complex phenomena with multiple causes, and trying to explain them with a single cause is always very risky and hard to defend within the scientific community (even if it is hard to shake the widespread belief that justice often depends on “what the judge had to eat this morning”).
From Thought to Language | No comments
