The Pennsylvania v. Trump PI and an example of biased assimilation

I just had the chance to read EDPA Judge Beetlestone's opinion explaining the grant of preliminary injunctive relief to Pennsylvania in the Commonwealth's challenge to the overdue conscience protections afforded employers by Interim Final Regulations implementing the ACA and the RFRA. There are problems from beginning to end. But for now I'll focus on a particularly striking example of motivated reasoning in the opinion. 

Students of judicial opinions are familiar with the concept of motivated reasoning. The example here, a type known as "biased assimilation," might perhaps even more accurately be described as a motivated _lack of_ reasoning. This is the "tendency to interpret information in a way that supports a desired conclusion. Supporting facts may seem overwhelmingly strong and negating facts may seem automatically weak."

To support a legal conclusion about the harm from unintended pregnancies that will result in the absence of a preliminary injunction, Judge Beetlestone credited an incredible statistic that can be shown to be false for anyone with some curiosity, an internet connection, and the ability to read footnotes. 

 

Here's the sentence that sent me looking: "Eighty five percent of women who do not use any form of contraceptive services and who do not want to become pregnant, become pregnant in one year." Can that really be true? At a minimum, mustn't there be a population constraint of some sort, such as "sexually active women between ages __ and __"?  As written, the sentence just can't be right.

The cited source is p. 106 of the Institute of Medicine Report. Sure enough, the table at that page does appear to support Judge Beetlestone's proposition. Table 5.3 is titled "Percentage of U.S. Women Experiencing an Unintended Pregnancy During First Year of Typical Use and Typical Year of Perfect Use, by Contraceptive Method." The first "method" is none, and this carries an 85% chance of "experiencing unintended pregnancy in first year" of both "typical use" and "perfect use." 

But isn't there something strange going on here? The table is about "unintended pregnancies" and "contraceptive method," but the statistic is about the use of no method at all. Unfortunately, the citation in the IoM Report is not particularly helpful. It just says: "SOURCE: © 2007 by Contraceptive Technology Communications Reprinted by permission of Ardent Media, Inc." But Google steps in where the IoM authors fell short. I pasted the source material into a google search box and went to the first hit: http://www.contraceptivetechnology.org/. From there I clicked on "The Book," which brought me to a drop-down menu. The first choice was "Take a Peek > Contraceptive Efficacy." That sounded like what I was looking for, so I did take a peek. And I found a table very similar to the one in the IoM Report.

Among other things, this version has footnotes not included in the IoM version. The most important is footnote 4, which is the footnote to the "None" "method" yielding an unintended pregnancy rate of 85% in a year. It turns out that the number is pretty much the opposite of what Judge Beetlestone cited it for. Here's what the footnote says:

The percentages becoming pregnant in columns (2) and (3) are based on data from populations where contraception is not used and from women who cease using contraception in order to become pregnant. Among such populations, about 89% become pregnant within 1 year. This estimate was lowered slightly (to 85%) to represent the percentage who would become pregnant within 1 year among women now relying on reversible methods of contraception if they abandoned contraception altogether. (emphasis added)

The 85% figure is not about unintended pregnancies. The population includes "women who cease using contraception in order to become pregnant."

There are still a lot of unanswered questions about the population at issue (age range, sexual activity, and so on). But there's no need to go any further at this point. The district court's statistic was facially incredible, and a little digging would have easily uncovered the mistake. 

It's just one example, to be sure, and just about one part of the opinion. But it has the virtue of being an unarguable error. The best explanation is biased assimilation.