Statistics in NIFLA v. Becerra, the Crisis Pregnancy Center Disclosure Case

Check out how the majority and the dissent describe one aspect of the California disclosure rules.

National Institiute of Family Life Advocates v. Becerra is an interesting, important, and difficult case, and there's a lot to be said both for the majority and dissenting opinions. But here I just wanted to point out how they deal with one data point—the number of languages in which the compelled disclosures have to be made. This data point is potentially relevant because it bears on how much the compelled disclosure interferes with the centers' speech; the majority argues,

The application of the unlicensed notice to advertisements demonstrates just how burdensome it is. The notice applies to all "print and digital advertising materials" by an unlicensed covered facility. These materials must include a government-drafted statement that "[t]his facility is not licensed as a medical facility by the State of California and has no licensed medical provider who provides or directly supervises the provision of services." An unlicensed facility must call attention to the notice, instead of its own message, by some method such as larger text or contrasting type or color.

This scripted language must be posted in English and as many other languages as California chooses to require. As California conceded at oral argument, a billboard for an unlicensed facility that says "Choose Life" would have to surround that two-word statement with a 29-word statement from the government, in as many as 13 different languages. In this way, the unlicensed notice drowns out the facility's own message. More likely, the "detail required" by the unlicensed notice "effectively rules out" the possibility of having such a billboard in the first place.

Some might argue that the magnitude of the burden shouldn't matter, either because such compelled disclosures are unconstitutional regardless of how big they are, or because they are constitutional regardless of how big they are—but the majority did view it as at least potentially relevant to some readers.

But just how should we measure the magnitude of the burden, given that the required number of languages varies from county to county? California law requires disclosures (both by crisis pregnancy centers and by many other medical facilities) in a particular language if the primary speakers of the language in the county exceed a particular threshold, and the number of languages can vary sharply, from 2 to 13. Here's how the majority frames the data, even before it gets to the explicit argument about burden:

The notice must be in English and any additional languages identified by state law. In some counties, that means the notice must be spelled out in 13 different languages. See State of Cal., Dept. of Health Care Services, Frequency of Threshold Language Speakers in the Medi-Cal Population by County for Jan. 2015, pp. 4-5 (Sept. 2016) (identifying the required languages for Los Angeles County as English, Spanish, Armenian, Mandarin, Cantonese, Korean, Vietnamese, Farsi, Tagalog, Russian, Cambodian, Other Chinese, and Arabic).

L.A. County, though, is the only county that requires 13 different languages; this isn't happening in "some" counties, but only in one.

On the other hand, here's how the dissent frames the data:

[The Act] would require disclosure in no more than two languages—English and Spanish—in the vast majority of California's 58 counties. The exception is Los Angeles County, where, given the large number of different-language speaking groups, expression in many languages may prove necessary to communicate the message to those whom that message will help. Whether the requirement of 13 different languages goes too far and is unnecessarily burdensome in light of the need to secure the statutory objectives is a matter that concerns Los Angeles County alone, and it is a proper subject for a Los Angeles-based as applied challenge in light of whatever facts a plaintiff finds relevant. At most, such facts might show a need for fewer languages, not invalidation of the statute.

Yet while it's true that only two languages are required "in the vast majority of California's 58 counties," that vast majority of counties contains a minority of the state's population. If you review the data in Dept. of Health Care Services, Frequency of Threshold Language Speakers in the Medi-Cal Population by County for Jan. 2015, apparently the proper source (cited by the majority, and the dissent doesn't offer any rival source), and then cross-reference it with population statistics for the state (I used the 1/1/2018 data), you find that more than two languages are required in counties that represent a titch more than 2/3 of California's population; it seems likely, then, that about 2/3 of California clinics have to provide disclosure in more than two languages.

Indeed, over 57% of Californians live in counties where the disclosures have to be given in 6 or more languages—Alameda (6), Los Angeles (13), Orange (6), Sacramento (6), San Diego (6), San Francisco (7), and Santa Clara (7). You can also calculate the average number of languages per county, weighted by county population; that should roughly capture the average number of languages seen by an average Californian visiting a clinic, as well as the average number of languages required per clinic (assuming the clinics are distributed proportionally to the population). That weighted average is 6.3.

So the majority opinion stresses an outlier county (with 13 languages), and seems to err in describing "some counties" (rather than one) requiring the 13 languages. But the dissent, by focusing on the "vast majority of .. counties" in which only 2 languages are required, misses the picture for 2/3 of all Californians, who will see 3 or more languages, and for the majority of all Californians, who will see 6 or more languages.

This is a small matter, of course; but I thought it was worth pointing out. Law students, lawyers, and others: Be careful with how legal sources (well, really, any sources) describe data, even when the sources come from highly authoritative institutions.