Junk science

July 22, 2019

Whatever Happened To Science—Revisited

By Michael D. Shaw

The original article appeared in 2013, and attempted to explain the deterioration of science, compared to the great promise it held right after World War II. A big factor was the explosion in available research dollars. A follow-up piece came out a year later, and touched on the pernicious influence of politics in science. Also mentioned was the cynical use of advanced statistical software in the attempt to find correlations between countless measured parameters. A third piece, published in 2018, focused on faddish junk science, foisted on an unsuspecting public.

This time, we will examine blatantly political diatribe masquerading as science, and published in respected journals. Exhibit A is a newly released polemic entitled “Association of Preterm Births Among US Latina Women With the 2016 Presidential Election.”

As the title aptly indicates, the researchers are attempting to link the election of Donald Trump with an increase in premature births within this particular cohort. Indeed, the article concludes: “Given the rhetoric and policies promised under the Trump presidential campaign, the 2016 presidential election has been proposed as a significant stressor in the lives of US immigrants, their families, and their communities, with potentially uniquely acute effects on the US Latino population.”

The notion of stress causing premature birth is key to this article, and is supported by three references. The first begins as a keyword search which yielded 107 articles to be analyzed. All sorts of populations were involved in these studies, and no reference is made to any sort of political etiology. Interestingly, the review makes very positive mention of the influence of social support, which can offset these stressors.

The second reference is a comprehensive discussion about the contribution of maternal stress to preterm birth, but offers no robust support of this contention. Rather, it suggests areas for future study.

The third reference is another literature review that looks at the epidemiology and causes of preterm birth. However, out of the 141 sources cited, only six refer to stress.

In short, the authors of the newly released study (Gemmill et al.) do not present a sound case for stress being a major factor in premature births. Moreover, if stress were involved in these births, the authors make no attempt to link it specifically to the Trump election. Rather, they merely assume so because of the timing, and then underscore this with his supposed “anti-immigrant” (conveniently avoiding whether legal or illegal) bias. A more flagrant example of the post hoc fallacy could scarcely be proffered.

Fair enough, you might ask, but how significant was this increase in preterm births? According to their analysis, “[W]e observed approximately 3.2% to 3.6% more preterm births to Latina women above expected levels of preterm births had the election not occurred.” But, what does “had the election not occurred” mean? The election has to occur every four years, as it has since the first contest of 1788-89!

The implications here are fanciful indeed. Had there either been no election with Obama staying on, or an alternative universe in which Hillary Clinton won, the rate of preterm births would not have increased. But such a direct conclusion would require nothing less than these fantasies to be carried out as a control group. By the way, if you look at figure 1 in the study you will note that the number of premature Latino births was higher–on average–in the Obama years than in those following 2016.

The Gemmill et al. study is quite similar to an earlier work published in the BMJ entitled “Severe sociopolitical stressors and preterm births in New York City: 1 September 2015 to 31 August 2017.” Interestingly, the authors of this study are social scientists, while Gemmill et al. are affiliated with public health institutions.

Notwithstanding the lack of true scientific rigor in either article, beyond supporting the work with statistical analyses, one could possibly make an argument for studying the effects of politics on preterm births from a sociological point of view. However, what possible public health takeaway could be garnered? Avoid getting pregnant if you don’t like the current president? After all, stressors are always with us, so who’s to say that a politically-derived stress in worse than any other?

I suspect that the late Nobel laureate Richard Feynman would include both of these studies under his banner of pseudoscience.