by Carl V Phillips
I was not going to post today, but there is so much hilarious chatter about this new press release from San Diego State University, and their FDA-funded “research” on e-cigarettes that I could not resist. This simplistic research about web searches related to e-cigarettes deserves a paragraph-by-paragraph dissection.
Oh, and of course there is a new journal paper that goes with it. But, seriously, who cares? Academic “public health” practice has descended to the point that a journal paper is just an excuse to write an even more misleading press release. It is time to stop pretending otherwise and just peer-review the press release. I am sure if I dissected the paper itself I could identify numerous problems that are not evident from just the press release — that seems to always be the case — but, again, who cares? It is not as if anyone in public health pays any attention to the quality of the science. When the paper is cited, those citing it will effectively just be citing the press release.
It is worth starting with the last bit, to see who shares the “credit” here:
The study was funded by 5R01CA169189-02, RCA173299A, and T32CA009492 from the National Cancer Institute and U.S. Food and Drug Administration Center for Tobacco Products. The content is solely the responsibility of the authors and does not represent the official views of the funders. The funders had no role in the design, conduct, or interpretation of the study nor the preparation, review, or approval of the manuscript. Additional collaborators on this study included: Benjamin Althouse of the Santa Fe Institute; Jon-Patrick Allem of the University of Southern California; Eric C. Leas of the University of California, San Diego; and Mark Dredze of Johns Hopkins University.
About San Diego State University: San Diego State University is a major public research institution that provides transformative experiences, both inside and outside of the classroom…blah, blah, blah….
Ok, on to the silliness, which begins with:
The Oxford Dictionaries selected “vape”–as in, to smoke from an electronic cigarette–as word of the year in 2014. It turns out that Internet users’ search behavior tells a similar story.
Between 2009 and 2015, the number of people in the United States seeking information online about vaping rose dramatically, according to a recent study co-led by San Diego State University Internet health expert John W. Ayers and University of North Carolina tobacco control expert Rebecca S. Williams as a part of the Internet Tobacco Vendors Study.
Yes, that really is the start of the press release. Normally cutesy hooks like this are not worth thinking about, but this is ostensibly a scientific press release, so let’s think like scientists for a moment. What could “tells a similar story” possibly mean? What story? The reader is presented with no story that offers any similarity to anything, let alone to search histories.
If the reader has the knowledge to fill in what the first sentence represents, then this becomes even sillier. The “word of the year” is not a measure of prevalence, or the first derivative of prevalence, of a word’s use. Those play some role, obviously, but the choice is ultimately a decision about what word seems to be a good representative of the changing zeitgeist, as evidenced by the many nominees and winners that are not particularly common. You would think that most scholars already know this, and those who did not know would invest the three minutes’ of research it would take to discover it before invoking the inaccurate analogy. But we are talking about public health activists, not scholars.
Oh, and of course, “to smoke from an electronic cigarette” would be a fairly stupid thing to do. Of course, don’t blame the linguistically careful folks at Oxford Dictionaries for that error; their definition makes clear they know that there is no smoke involved. That error is the fault of the team of crack team of FDA-funded researchers.
Moving on with the study results, we learn the shocking news that searches for a product category increase dramatically when it becomes fairly popular, as compared to when it was barely known. In other news, searches for “greek yogurt”, “is Pluto a planet”, and “Bernie Sanders” increased dramatically between 2009 and 2015.
E-cigarettes and other hand-held vaporizers began appearing on American shelves in the mid-2000s. Since then, they’ve quickly risen in popularity while regulators have been slow to adapt smoking legislation to account for these devices.
It is not clear which “other hand-held vaporizers” are not e-cigarettes, but never mind that. Also, we will just gloss over the odd use of either “mid-” or “shelves” (it should either say “began appearing via internet sales” or “late-2000s”; but, hey, it is just so much to ask to get simple background facts right). Focus instead on the last bit, in which these people who are supposedly issuing a press release about research lead off with an unsupported normative claim (that regulators should adapt smoking legislation to “account for” [sic] e-cigarettes). Notice that they slip this in as if it goes without saying, when it is actually a far more significant claim than anything in the research results they actually report. That, of course, is SOP in “public health”, where research is done primarily as an excuse to express unsupported political opinions.
“Big Tobacco has largely taken over the e-cigarette industry. Alongside unchecked marketing and advertising, e-cigarettes have exploded online,” Ayers said.
So in public health land, “largely taken over” means having less than half the market share (which is itself divided amongst several competing companies), and “unchecked” means “subject to the countless restrictions on all marketing communication, as well as specific restrictions such as not being able to tell the truth about the health risks.” Um, yeah. So that would mean that Apple has largely taken over the smartphone industry, and is free to lie to consumers about their products with impunity.
Internet users’ search history bears this out.
Wait, what? (Note that I am not skipping any text.) How can any data about search history “bear out” any of the claims in the previous sentence? The closest it could come to any of them would be that an increase in searches would probably be associated with that “exploded online” thing, though the latter seems to be a claim about the supply of information while the searches reflect demand for information.
Ayers, Williams, and a team of colleagues from across the country examined search history from Google Trends, which includes statistics on what specific words people searched for, the search term’s popularity relative to all other concurrent searches in a specified time, date and geographic location. From this data, the researchers can find patterns that point to Internet searchers’ apparent preferences and attitudes.
Ok, that is a nice high-school science project or term-paper-level exercise. One can learn something from that. Let’s see if they did….
When they looked at searches related to e-cigarettes starting in 2009, they found a sharply rising trend through 2015 with no end in sight.
I am not sure if that is LOL funny for everyone reading it, or if it is just me. I actually had to pause for a couple of minutes before resuming. As already noted, obviously there is going to be a sharply rising trend from 2009. But it is the “no end in sight” that really got me, at several levels. Why would anyone in their right mind even think to mention that? Internet use is increasing and e-cigarettes are increasing in popularity; what possible end are they even talking about?
Moreover, at a deeper scientific level, we are talking about a social phenomenon that could end, and quite abruptly, whatever the historical trend is. This is not like claiming there is no end in sight for the warming of the planet, which could be based on what we know about atmospheric chemistry, which is not affected by social proclivities that can change abruptly. Social trends can change abruptly. I am sure that the data for searches of “Bernie Sanders” follows a very similar trend to searches for e-cigarettes, but any real social scientist can foresee an end to that trend.
For example, in 2014 there were about 8.5 million e-cigarette-related Google searches. For 2015, their model forecasts an increase in these searchers of about 62-percent. Looking at geographic data, they found that e-cigarette searches have diffused across the nation, suggesting that e-cigarettes have become a widespread cultural phenomenon in every U.S. state. Over the same time period, searches for e-cigarettes far outpaced other “smoking alternatives” such as snus (smokeless tobacco) or nicotine gum or patches.
I am not sure what that first sentence means, but it obviously does not mean what it says. I would guess that readers of this blog alone conducted most of 8.5 million “e-cigarette-related” Google searches in 2014. Obviously researchers can choose to study whatever specific phenomena they want to, narrowing what they are counting up, but they need to say what they are doing. The fact that someone would put out an obviously incorrect number like this and that the press would dutifully report it without thinking it through speaks volumes about what a joke public health discourse has become.
The results about states is equally useful information. I mean, who would have guessed that the primitive tribes of Tennessee and the transcendent life forms in Oregon would have similar internet search behavior to the rest of the country? It is not surprising that the searches exceeded those for other low-risk alternatives (though I have no idea what their scare quotes are supposed to mean — presumably it is innuendo that alternatives to smoking are not really alternatives to smoking). Though I have to wonder if their methodology missed most of the searches for smokeless tobacco, which probably used established brand names rather than the word “snus”.
The researchers published their findings today in the American Journal of Preventive Medicine. (Note: this URL will be active after the embargo lifts)
Oh, look, I don’t have to wonder. I could go read the paper and learn their exact methodology. Just kidding. I have little doubt that I would still not know, given the poor quality of methods reporting in public health research. (Also I would have to go search for it, since the aforementioned URL does not actually exist on that page — the smallest of the many errors that appear.) Since there is really no chance that anything useful will come of that effort, I am skipping it. Anyone who actually bothers to read the paper can use the comments to backfill anything good I might have missed.
What most concerns the researchers, though, is that when people search for e-cigarette information, they’re using search terms like: “best e-cig,” “buy vapes” or “shop vaping.”
Why should anyone care about what most concerns these researchers? They should not. Being able to run simple statistics on Google searches implies absolutely no moral authority to opine about, let alone ability to analyze, what is better for society. (Incidentally, what would most concern me would be if the phrases “buy vapes” or “shop vaping” were more common than grammatically sensible phrases. I assume they actually were not, and that they were intentionally cherrypicked to try to ridicule consumers.)
In any case, in what world would it be at all concerning that most of the searches for a consumer product would be a combination of seeking review information (as “best e-cig” is presumably intended to do) and purchase options? It would certainly not surprise anyone.
“One of the most surprising findings of this study was that searches for where to buy e-cigarettes outpaced searches about health concerns or smoking cessation,” Williams said.
I stand corrected. Let me amend that to, “It certainly would not surprise anyone with a clue about how the world works.”
“Despite what the media and e-cigarette industry might have you believe, there is little research evidence to support the notion that e-cigarettes are safe or an effective tool to help smokers quit. Given that, we think it’s revealing that there were fewer searches about safety and cessation topics than about shopping.” In fact, she said, searches for e-cigarette safety concerns represented less than 1 percent of e-cigarette searches, and this number has declined over the past two years.
Set aside all the usual lies that are embedded in that, about what the evidence shows, the “safe” wordplay, misrepresentation of the predominant message in the media, and misrepresenting what is permitted in marketing. Just skip to the specific claim here, that he seems to think that those phenomena (even if they actually were true) would cause people to do more searches for background information than for product reviews and purchases. Seriously?
Even for a product where most of the background information you would find from a random search was not utter bullshit, as it is in this case, consumers are still going to mostly search regarding purchase plans. People who are seriously interested in that other information develop networks of trusted sources; they would get nowhere doing random searches. Anyone who is shopping for e-cigarettes has already acquired the information that e-cigarettes are worth shopping for, presumably knowing that they are a low-risk alternative to smoking. Why, exactly, would he want to search for that?
Frankly, I would be extremely disturbed to learn that many short-phrase searches about e-cigarettes were seeking scientific information. That would truly be a tragic commentary on people’s understanding of how to learn anything about controversial issues via the internet.
If these “researchers” actually had any expertise in the research they were conducting — which is to say, about consumer online search behavior, not about tobacco politics — we might have gotten some useful information. For example, how do these statistics (which are inevitably very weak, depending on their choices of phrases and how they were coded) compare to those for other products? What is the quality of information that someone would get were she to pursue such a blind search for information? This speaks to a common problem in “public health”, where research follows political interests rather than scientific skills. There is absolutely no reason why someone doing this particular research would need to know anything about e-cigarettes, other than some basic vocabulary. They should, however, know something about real-world consumer behavior.
A linguistic trend also emerged from the study. The term “vaping” has quickly overtaken “e-cigarettes” as the preferred nomenclature in the United States. That’s important for health officials and researchers to recognize, the team noted. Surveillance of smoking trends is done primarily through surveys and questionnaires, and knowing which terms people use can affect the accuracy of this data.
Wow, that is almost useful information. Of course most of us already knew that. We also know that “vaping” is an act whereas “e-cigarettes” are a product, and so they are not really commensurate, which is a rather important distinction for doing those surveys and questionnaires. (“Hey, guys, according to this new study, we should stop asking ‘have you used an e-cigarette’ and start asking ‘have you used a vaping’.”) Those of us who know how to do surveys, of course, already make a point to define what we are talking about, including offering the multiple popular terms if that is an issue.
Also, one of the major weapons anti-smoking advocacy groups have is counter-advertising. In the Internet age, advertisers look for specific keywords to target their advertisements. Knowing that more people use the term “vaping” than “e-cig” helps them be more targeted and effective, Ayers said.
So “counter-advertising” in the interests of “anti-smoking” should make sure to intercept all those searches for “vaping” in order to make sure the anti-vaping propaganda reaches its target audience of people who want to avoid smoking. Yup, they pretty much earned their FDA money with that one.
“Labels do matter,” Ayers said. “When you call it ‘vaping,’ you’re using a brand new word that doesn’t have the same historical baggage as ‘smoking’ or ‘cigarette.’ They’ve relabeled it. Health campaigns need to recognize this so they can keep up.”
“They’ve relabeled it”? Who is this “they”? Oh yeah, I remember, it is the Oxford Dictionary people. The bastards.
Of course what “health campaigns” really need to do to keep up is to learn something about health. They are not going to achieve that by doing random internet searches, by the way. Nor by reading press releases or papers by the tobacco control industry’s pet academics.