Magnifying Partisan Research with a Partisan Spin

Photo credit: Washington Post

Social scientific studies are increasingly infected by ideological bias and a crisis of unreplicable results. Compound that with the ideological bias of the mass media, which spin findings to advance their own partisan narratives, and you get articles like this one from the Washington Post: “Trump’s presidency may be making Latinos sick.

Trump’s presidency may be making some people sick, a growing number of studies suggest. Researchers have begun to identify correlations between Trump’s election and worsening cardiovascular health, sleep problems, anxiety and stress, especially among Latinos in the United States. A study published Friday using data from the Centers for Disease Control and Prevention found the risk of premature birth was higher than expected among Latina women following Trump’s election.

This is the same kind of junk reporting of tendentious science that we see increasingly in Virginia, where newspapers report on “studies” showing “correlations” that supposedly demonstrate the existence of systemic institutional racism. It’s not impossible that some of the studies are valid. But they need to be subjected to much closer scrutiny before being accepted and propagated widely, as they invariably are.

The WaPo’s William Wan and Lindsey Bever briefly acknowledge that the “some of the research” regarding a “Trump effect” on the health of Hispanics has been “inconclusive,” and they do quote one academic expert as cautioning against drawing hasty conclusions. Otherwise, the article proceeds with the supposition that the “Trump effect” is real. “Taken altogether, some researchers say, the growing body of work suggests Trump’s short-term strategy of fanning immigration fears for political gains could be causing health problems.”

Let’s take a closer look at the original study cited by the Post: “Association of Preterm Births Births Among US Latina Women With the 2016 Presidential Election.” And let us accept the key statistical finding that the number of preterm male births among Hispanics increased from 10.2% before the election to 11.0% after, while the number of preterm female births increased from 9.2% to 10.2%. (I am not qualified to appraise the merits of their statistical methodology.)

The authors offer numerous caveats and qualifiers not noted in the Post article. The U.S. presidential election, the study authors write, “appears” to have been associated with the uptick in preterm births. The researchers “proposed” that anti-immigration rhetoric and policies were to blame, but acknowledge that “future research should evaluate the association of these actions with public health.”

The study says that the 2016 election “may have” acutely stressed Latino immigrants and their U.S.-born coethnic family members. Preterm birth “appear” to have distinct etiological linkages with material psychosocial stress, although “the biological mechanisms underlying this association remain unclear.”

The study authors acknowledge, but do not explore, an alternative explanation: “Birth outcomes may also be affected by changes in health-seeking behavior; a recent study documented increases in inadequate prenatal care among US nonnative Latina women coincident with anti-immigration rhetoric.”

Furthermore, due to data limitations, the analysis does not differentiate between native and nonnative Latina women, although the authors write they “anticipated” that, had the data been available, the effect would have been most pronounced among nonnative women.

The authors also acknowledge an issue related to their statistical methodology: They “measured gestational age based on last menstrual period rather than the preferred measure based on obstetric estimate.”

Finally, they concede that, due to a lack of data, they did not study other groups, such as persons of Middle Eastern and North African heritage, targeted by anti-immigration rhetoric during the presidential inauguration.

Most notably — a fact that the authors never mention — the study does not compare pre-term birth trends for Hispanics to the trends for whites, blacks, and Asians. Do other ethnic groups, adjusted for socioeconomic status and access to health care, exhibit the same uptick in pre-term births? For conjectures of a “Trump effect” to hold water, it is necessary to demonstrate that the trend for Hispanics differs from the trends for other racial/ethnic groups. Yet, for whatever reason, the study did not do so.

That is a lot of caveats and qualifiers.

I can’t help suspecting that the study authors found what they wanted to find, but I do give them credit for acknowledging the pitfalls and limitations of their data. As a journalist, I find the greatest fault with the Washington Post reporting of the study, which, with the exception of one partially dissenting remark, used it to advance a partisan narrative.

The Post says, “Democracy dies in darkness.” Yes, and the Post, which makes every public policy controversy a racial issue in order to advance the inflammatory Narrative of Endemic Racism, is the one smothering it with a pillow.