How Low Will They Go? Surveys As Propaganda, Not Research
Since the Nuremberg war crimes trials and the resulting establishment of modern health ethics, it has been considered unethical to try to cause impacts on people under the guise of doing research. The extreme case that motivated this was the Germans torturing and killing people and calling it research. But it was realized that less extreme manipulation of research subjects has similar ethical implications, and that a bright-line prohibition was necessary. Some human subjects research is inherently about impacts on people, of course; comparing two surgical techniques involves performing surgery. But it would be unethical to use the research protocol to, say, encourage support for more NIH funding. Ethical researchers try to minimize incidental impact on their study subjects.
Tobacco controllers conduct a lot of shoddy research and lie about what it shows. That is obvious. The “how low will they go?” series reports on the hundreds of more subtle — but often even more troublesome — transgressions tobacco controllers commit. It is not just that they are seeking misleading results, but they are also violating the ethical rules about manipulating their study subjects.
It is disturbing, though perhaps not surprising, that WHO’s marching orders to tobacco controllers including manipulating the views of research subjects under the guise of doing surveys. In their most recent report on their tobacco control efforts and what more they demand be done, one of the reasons given for doing surveys is:
…the act of responding to a survey can raise awareness about health issues among the community being surveyed, with the effect strengthened by repeat surveys on the same issue.
Unlike surgery research, survey research should have almost no impact on subjects, whether dressed-up as “raising awareness” or otherwise. Ethical researchers try to minimize how much their questions affect their subjects’ beliefs. By contrast, WHO is telling tobacco controller researchers to try to affect study subjects’ beliefs.
It can be tricky to completely avoid affecting survey respondents. If a survey asks people if they can hear a nearby industrial facility at night and then asks about sleep disturbances, it telegraphs that there is concern about the noise disrupting sleep. It might even draw subjects’ attention to the noise and create a problem that did not previously exist. This is one reason that a proper survey would ask about multiple unrelated exposures and health outcomes, trying to hide the true focus and hypothesis. The other reason for doing this is to avoid biasing the results. When subjects figure out the hypothesis, they often bias answers to the key questions.
The WHO statement about surveys is in the context of surveillance of tobacco product use (what they call “monitoring”). But the statement is obviously not claiming “if you simply ask people if they smoke, it will make them think about health issues.” Rather, they are encouraging governments and other researchers to add leading questions about health effects, and probably also political positions like smoking bans, into what is supposed to be surveillance of product use.
Many tobacco control surveys are intentional junk science, with questions phrased to get the desired answers or arranged to prime subjects to give a particular answer. For example, smokers might be asked a series of “are you concerned about…?” questions about health risks or other negative effects. Then, with their negative feeling about smoking temporarily maximized, they are asked “Do you want to quit smoking?” Moreover, they are only offered answers of “yes” and “no”, with nothing like “I often think I should quit smoking, but I never really want to quit.” Most subjects who would have affirmed an answer like that just say “yes”. This is how we get ridiculous claims like “73 percent of people who voluntarily engage in this activity, day after day, actually do not want to.”
The purpose of such manipulation is to engineer the desired answer to a key question and report it. That is unethical, of course. But if part of the intention is to leave subjects more likely to quit smoking, it is more unethical still. Tobacco controllers have the right to engage in marketing to try to get people to quit smoking. But it is an abuse to do this marketing under the guise of research.
While short of Mengele-level, it is at the level of unethical (and often illegal) political “push polls”. A typical push poll might involve calling voters and asking them “how concerned are you that the governor does the following bad things….” The primary goal is not to get the answer (though it might still be collected and reported), but to get the subject to listen to the story (which is often not true) and thus potentially change their vote. Similarly, tobacco controllers will ask tobacco product users a series of questions about their health status, linking the questions to tobacco use. “Do you know that vaping can cause impotence? Have you ever suffered impotence as a result of your vaping?” The purpose is not, of course, to collect the answers; they are worthless. It is to plant the idea that it happens.
A particularly notorious example was the fake survey in which tobacco controllers invented the term “third-hand smoke” and erroneously claimed to offer evidence that it is harmful. Subjects were asked a series of questions designed to cause them to worry about (nonexistent) health threats from deposited smoke residue, and then asked if they were concerned about it. It was then reported that people were concerned, and this was then offered as the (only) evidence that there was something to be concerned about.
If a contact were honestly identified as marketing rather than a survey, the voter would be less likely to listen to and believe the story about the governor and the vaper would be less likely to believe false claims about impotence. But people tend to trust that researchers are not lying to them. By eroding this trust, push pollers and tobacco controllers damage all future research efforts.
Of course tobacco controllers do not care that they are making it more difficult to do legitimate survey research, just as they do not care that their junk claims make people doubt legitimate epidemiology. They do not care about research ethics and abusing human subjects. They care so little that they do not even notice they are saying the quiet parts out loud, openly calling for research ethics violations in a major public document.
Follow Dr. Phillips on Twitter