Yesterday, a US federal judge ruled that the gruesome graphic labels that the FDA wants to mandate that cigarette manufacturers print on their packages would likely not stand up to first amendment scrutiny, and issued an injunction agains the regulation until the full constitutional issue could be litigated. (The text of the ruling was online this morning but the link does not seem to work anymore; you can link to my saved copy.) Since the litigation will likely take years, the judge ruled, requiring the labels appear until they are ruled unconstitutional, if that was indeed the ruling, would do irreparable harm. Thus the decision must be to not impose the regulation until and unless it was ruled constitutional.
The judge in question, Richard J. Leon of the United States District Court in Washington, was the same one who saved e-cigarettes from being banned in the US, making him the man of the year for tobacco harm reduction and maybe public health in general, though that honor, for which he will never get his deserved nomination for a Nobel Prize, was presumably not his motivation. He showed similar wisdom in the new ruling and that was very well reported by Duff Wilson in the New York Times.
The core basis of Leon’s ruling was that the labels were not simply the communication of facts. That is, as I have pointed out a few times, they were not warning labels. Anyone who refers to them as such apparently has no idea what the word “warning” means. I have characterized them as “emotional violence”, intentionally infliction of distress, and pointed out that they are designed to manipulate behavior in a particular direction rather than give people the facts they need to make an informed autonomous decision. Leon did not use the word “violence”, but clearly argued that the graphics do not warn, but merely manipulate emotion; indeed, he emphasizes the legal evidence presented to him that this is what they were designed to do. He wrote, “they appear to be more about shocking and repelling than warning”.
It is gratifying to see this recognition of the true nature of these graphics. I wonder if we will see anyone stop incorrectly referring to them as “warnings” as a result. Not the anti-smoking zealots, of course, or the World Health Organization, but maybe some of the people who try to write about this topic honestly will get the memo.
Another key basis for the ruling, a somewhat more technical, and thus not the focus of popular press stories (at least not when written by health reporters — there might be some more thorough reports by law reporters out there), is that the large warnings, covering most of both sides of the package, clearly do not meet the standard for being the minimal infringement on free speech necessary to achieve the government’s purpose. That part of the ruing actually included the seemingly sarcastic parenthetical, “…purpose (whatever it might be)”. Indeed, the judge wrote about a previous ruling he was drawing upon and clearly agreed with:
the dimensions alone strongly suggest that the Rule was designed to achieve the very objective articulated by the Secretary ofHealth and Human Services: to “rebrand our cigarette packs,” treating (as the FDA Commissioner announced last year) “every single pack of cigarettes in our country” as a “mini-billboard.,,26 Mot. for PI at 6 (citing a June 2001 press briefing with Sec. Sebelius, and an FDA Tobacco Strategy Announcement). A “mini-billboard,” indeed, for its obvious anti-smoking agenda!
And, yes, that exclamation point was in the original — apparently you are allowed two of those in a 29 page judicial opinion. Interestingly, Wilson’s NYT article quoted the last few words and the punctuation, out of context, as well as another passage that characterizes the government agenda, without suggesting that the the agenda must be legitimate or even quoting some QUANGO activist endorsing the agenda. It was remarkable restraint and professionalism for a health reporter, since declaring fealty to the government agencies that feed them most of their stories is the standard practice. Wilson did include the mandatory QUANGO quote from Matt Myers, but chose a technical observation about how there will be an appeal of this ruling, rather than printing the self-appointed holy-man ranting that Myers no doubt also offered.
All in all, that was some pretty healthy health news. But I have to take a few points off for something — being a professor is a profession and state of mind, after all, not a matter of who writes you a paycheck, and I should stay true to the titular theme of this series. Both Leon and Wilson did blunder when they wrote about the FDA’s research on the effects of the graphic labels not addressing whether they would have a “statistically significant” effect on consumers’ awareness of the risks of smoking, and specifically the research not being designed to be able to answer that. Health reporters have no more business writing about statistical significance than they do writing about legal nuances of the First Amendment of the Constitution, and this is a great example of why. (Judges might also want to shy away from using jargon from highly technical topics except by quoting its use by experts.) Since it was obviously impossible to study the effect of the labeling before it happened, it is not clear how a study could be designed to test the effect, let alone designed to achieve statistical significance in so doing.
As a more general point about their error, statistical significance is a property of a dataset — or more precisely of a dataset and a particular hypothesis that is being tested — not a property of the world. It has to do with the chance of seeing a pattern in the data due to chance alone, and relates to how sure we should be about the results. What people care about, and what the law should care about, is whether there is a substantial effect. “Substantial” and “significant” (without the “statistical”) are rough synonyms in natural language. They have to be defined by context and basically mean “it matters”. By contrast statistical significance is precisely defined (although only a small minority of those who use the term could actually give you the correct definition) and does not necessarily matter.
So, credit for recognizing that the proposed graphic labels are not warnings, and that they are emotional manipulation, and for genuine objective reporting that (contrary to the usual tone of the NYT) made clear that a particular goal of some people in government is just some people’s preference, rather than some God-Given Correct Way. But points off for still, in spite of all that, not understanding some fundamental points about how science works.
[Update: The NYT undermines its good reporting with clueless editorializing — has both irony and humor.]