Tag Archives: second-order preferences

Economic innumeracy in public health, with an emphasis on tobacco harm reduction

by Carl V Phillips

I recently had the opportunity to give a talk at what was basically the wake for the end of the quarter-century run of the wonderful Robert Wood Johnson Foundation Scholars in Health Policy Research program at the University of Michigan. I chose to put together some themes from my work as a tribute to one of the goals of that program, bringing the thinking of serious social scientists into health policy arenas where it is desperately lacking. Alas, most of my fellow alumni focus on engineering a better medical system or medical financing, with few choosing to try to deal with public health (let alone “public health”). Medical practice is obviously extremely important, but not so desperately in need of imported thinkers. Well, at least you have me.

I got some great feedback on this talk making that alone well worth my effort. (Thanks to all my colleagues. And it was great seeing you. We’ll be in touch.) But I wanted to also share what I created more broadly here. The following are my slides from the talk, with some text to explain what is not fully contained in the slides, along with a bit of extra material that was not in the talk. Continue reading

Utter innumeracy: six impossible claims about tobacco most “public health” people believe before breakfast

by Carl V Phillips

As anyone with a modest understanding of the science knows, tobacco controllers and other “public health” people make countless statements that are utterly false. The tobacco control industry depends on making claims that flatly contradict what the science shows. But there is a special class of claims that are not wrong just because they contradict particular empirical evidence; rather, everyone should know they are wrong based merely on understanding some basics of how the world works. Many such claims are constantly repeated as if they were self-evidently true even though they are actually self-evidently false. I was having trouble defining the category until I recalled the quote from Alice in Wonderland alluded to in the title. Continue reading

Second-order preferences as addiction, cont. (part 4 of “what is addiction” dialog tree)

by Carl V Phillips

This post continues immediately from the previous one, which you will need to read to make sense of this.  That, in turn, is part of the “dialog tree” of how to talk to someone about what they think addiction is (starting here), which in turn follows from this post.  Got all that?

In the previous post, I pointed out that one thing that people often mean when they say “addiction” is that someone chooses smoking over not smoking, but “wants to quit”.  I pointed out that this almost certainly means they have a second-order preference for wanting to prefer not smoking over smoking, but that is not their actual preference.

I emphasized one critical point that is often overlooked when this concept is implicitly invoked as an excuse for torturing smokers into quitting:  Most everyone has been tricked into believing that every smoker who manages to quit will find himself just as happy/alert/clear-headed/etc. when he abstinent as he was when he smoked.  While this will be true for some people who give up tobacco/nicotine, it is definitely not true for many.  In particular, it is almost certainly not true for those who “want to quit” but start again after a period of abstinence.

In a hundred different ways, this lie permeates discussions about use and cessation (“all you need is just a little help getting over that hurdle”).  The lie dooms countless smokers — literally millions of them — to a repeated cycle of temporary abstinence, expecting to find that they no longer want to smoke after being abstinent for a while, followed by starting to smoke again because they still prefer smoking to abstinence.  But it is also one of the fundamental anti-THR lies that I have written about before:  Everyone would be happy being abstinent, and so they just need to get there; thus, we do not need THR.  (Actually, this claim is wrong for other reasons too, actually, but I will stick to the one theme.)

Circling back to the dialog tree of what to say to a non-ANTZ who invokes the old “…but they are still addicted” claim as a reason for opposing THR, when someone defines “addiction” with answer #6, how should we respond?  I took so long getting to this because I think it is necessary to really understand the implications of second-order preferences in order to answer well.  With the background in mind, the script for responding should begin with:

If someone says they want to quit but never does, or they quit for a while and start again, it shows that they actually prefer smoking to abstinence when those are the only two choices.  But there is a good chance they would act on their stated desire to quit, and stick with it, if given a way to make non-smoking better — that is, to keep some of the advantages of smoking.  That makes such a smoker the perfect candidate for THR.  Indeed, it is difficult to think of any more compelling reason to support THR than this notion of “addiction”.

If you can get to this point in a conversation with any thinking and feeling person (i.e., most any non-ANTZ), I venture to say that you will almost certainly win them over.

This argument renders the point about whether someone is “still addicted” rather moot.  But should the conversation continue to focus on addiction, the argument only becomes stronger:

Not only is THR a good way, and perhaps the only way, to help such an individual fulfill their stated desire to quit smoking, but it is the only way to actually overcome the addiction.  The abstinent ex-smoker whose preferences fit this pattern is still addicted by this definition, because his preferences have not changed — he still would prefer to smoke even though he is forcing himself to act like someone who prefers to not smoke.  However, if he is satisfied with THR, then not only is he no longer a smoker, but he is no longer “addicted”.

Just pause for a moment and let that sink in, because I think that these points might well be the strongest arguments for THR that are almost never made.

Having let it sink in, I have to finish with one caveat.  Arguably it is necessary to add something about the impact of past use to answer #6, since some reference to past use is part of most notions of what addiction might mean (recall definition #5 in particular, and also #4).  Thus it might be that the full version needs to be something like:

6a. Someone is addicted when he prefers taking particular action even though he would prefer to prefer to quit, and the reason he prefers the action is because of past consumption.

That is, someone is “addicted” to smoking if he has that pattern of preference and second-order preference, and because he ended up with that pattern because of past smoking.  It should be immediately apparent that this is of no practical consequence if we are talking about someone who already fits this description.  There is no changing the past.  The only change required is to add the phrase “since he is already in that situation” to the above arguments.

However, if this describes someone’s belief about tobacco/nicotine use (which I believe is the case for many who are anti-tobacco but not ANTZ, and who are thoughtful enough to figure out what they are really claiming), and they believe that the creation of the second-order preference disconnect is truly a terrible thing in itself, then this concept of “addiction” remains an argument for preventing anyone from initiating use of even a low-risk product.  I do not actually think that this is a good argument, but I will acknowledge that it could be defended because, unlike the ANTZ, I am not just spewing propaganda, but am seeking honest understanding of these issues.

The reasons I think it is a very weak argument start with the observation that, as noted in the previous post, there is nothing inherently terrible or unusual about having a second-order preference to have a different preference about some choice.  I used to drink orange juice that I mixed from those little cans of frozen concentrate, but then I got in the habit of buying the jugs of “premium, never frozen” juice, and now strongly prefer that.  But I would prefer to still like the frozen concentrate just as much, since it is cheaper and easier to carry and store, but my experience changed my tastes.  But that experience also left me happier, since the premium stuff really is better.  My indulgence caused a second-order preference disconnect, but did not make me worse off.

The typical response to that observation would be that tobacco use is different because it causes not merely the discovery you like something and an evolution of tastes, but rather it causes some strange massive reprogramming of your brain.  It is not clear why that even matters, though, and moreover  I am not entirely convinced by these claims.  Most of them come out of the notoriously dishonest ANTZ “research” literature or the brain porn research in neurochemistry that is currently in vogue but is widely derided for not actually supporting the claims that are made.  Yes, tobacco/nicotine use changes our brains, but so does acquiring a taste for orange juice or reading a book (if you remember what you read, it changed your brain).  The bottom line is that there is a possible argument against tobacco to be found here, but even the factual claims are far from definitive, and the ethical and economic conclusions are much shakier still.  After all, there is also a valid case to be made that someone’s exposure to tobacco/nicotine can improve their life by making them susceptible to the benefits, so long as they can get those benefits without the high cost of smoking.  So perhaps I should complain that I was denied that opportunity to reprogram my brain at an early age, just as I wish I had learned to speak Mandarin when I was still young enough to do it easily.

But I digress.  That is a topic for another day.  At the end of this day I will sum up the thread by saying that if you can get an open-minded person to discuss their concerns about “addiction”, and you can offer the responses presented in this series, there is a very good chance that you can show them that the concept  of addiction offers much stronger arguments in favor of THR than it does against it.

Second-order preferences as addiction (part 3 of “what is addiction” dialog tree)

This post continues the series (started here) on “addiction”.  While the discussion of second-order preferences mostly stands on its own as an independent point, some bits will make more sense if you have read the series.

The last of the likely answers to your question “what do you mean by addiction” invokes the concept of second-order preferences (though presumably without using that jargon):

6. Someone is addicted when he keeps choosing to take a particular action even though he wants to quit.

I have thought and written quite a bit about the concept of second-order preferences as it relates to smoking and THR, much of it in collaboration with my THR.o colleague, Catherine Nissen.  The concept is that someone may prefer a particular choice (their preference), but would prefer to not prefer that choice (their preference about preferences, aka a second-order preference).

It is commonly claimed that most smokers want to quit.  The surveys that support this are actually quite suspect, since smokers know that they are supposed to say that, and thus often just give that answer as cheap talk.  But while this explains a large portion of the responses, there are definitely some people who sincerely assert that they want to not smoke, even as their actions show that they are choosing to smoke.  But what can this obvious contradiction possibly mean?  It almost certainly means, in most cases, that their second-order preference is to be someone who wants to not smoke, even though the reality is that they are someone who really wants to smoke.

This, unlike most of the previous notions of what “addiction” means, is a very real and common phenomenon with practical implications.  But those implications are not what those who use “addiction” as a negative epithet might think.

There is nothing horrible, or even the slightest bit unusual, about this second-order preference pattern.  We all have countless preferences for different preferences.  I would prefer to like going to the gym as much as I like playing computer games, and I would prefer to like unsweetened iced tea as much as I like Coke.

It is sometimes perfectly ethical and otherwise reasonable to implement public policies that help people achieve their second-order preferences in spite of their actual preferences.  But justifying this is complicated.  It clearly cannot just be taken as a given that such policies are good, as is sometimes implied.  In particular, there is no obvious reason why we (as a society) should favor someone’s second-order preferences over their actual preferences, let alone an obvious justification for policies that inflict costs in order to push that choice.

Most people who support such policies have clearly never thought through the ethical questions, and frankly appear mostly to be lying about their stated motives.  There are those who argue that such policies help make everyone “better” people, but that is only by their own measure.  These are usually busybodies who would push the particular choice even if it were not someone’s second-order preference.  How many active supporters of anti-smoking efforts ever say, “if someone does not want to quit, then we should avoid interfering with that choice”?  Their real motives are wanting to force particular actions, not to help people fulfill their own second-order preferences.

A particular clear justification for some types of interventions can be found in how I phrased my personal second-order preferences.  Notice that I do not want diminish my enjoyment of computer games to the point that I prefer the gym; I want to raise my enjoyment of gym-going to that of computer games.  If there were an action that accomplished that, making gym attendance better without diminishing the value of any alternative, there would be no question that this was an improvement in the world.  But, of course, for most actions, particularly those where there is profit to be had by improving quality, there is no such magical cost-free improvement that no one has bothered to implement.  You can sweeten and flavor the iced tea, but only at the cost of making it nutritionally about the same as Coke.  My gym options are much nicer than the university weight rooms I frequented for a few decades, thanks to the wonders of the free market, but those wonders are still not enough to change the fact that staying home is nicer still, and there is no reason to expect magical further improvement.

Many anti-smoking efforts can be characterized as attempts to cause people’s preferences to align with their second-order preferences.  But this is done by lowering the net benefits of smoking until quitting is the preferred option.  That is ethically justified if (and those of us who believe in individual liberty would argue, only if) those who are affected genuinely want an external actor to do that to them, and they understand that the action is not going to make being abstinent as good as smoking, but is going to make smoking as bad as being abstinent.  So, a drug like Chantix that makes someone not want to smoke is fine (ignoring, in this analysis, the side effects), so long as someone understands that it probably will not make the experience during long-term abstinence as enjoyable/productive/etc. as life when they smoked, but is merely going to lower the benefits of smoking in the short run to help achieve a new habit of abstinence that is in line with the second-order preference that the individual cannot bring herself to act upon.

Yes, there are a lot of conditionals and caveats in that ethical analysis.  Notice that those conditions are clearly not met for policies that try to make smokers miserable through punitive taxation, forcing them out of pubs, and otherwise increasing the costs of smoking.  Many of those who suffer the inflicted costs do not actually want to quit, and the rest have not consented to someone making one of their choices less pleasant in order to try to try to align their preferences.  Moreover, even many of those who would consent to have their preferences forced into alignment, as they do with voluntary drug therapy, have been tricked into believing that the result will be that they will be just as happy or happier once they are abstinent.

This last bit is critical and rather more subtle than the other conditions that are more often discussed.  Most people (smokers and nonsmokers alike) have been tricked into believing that every smoker will be just as happy as a nonsmoker, once they get over some hurdle (which is often called “addiction”, but is really better labeled “dependence” or even just “withdrawal” — see the discussion under #4 in the previous post).  Some ex-smokers are happier, certainly, but these are probably the ones who just decide to quit and end up staying abstinent.  Those who are seeking extra help because just deciding to quit does not work are typically misled into believing that once those tools cause them to quit, they will be happy about being abstinent, but that is often just not so.

When the ANTZ defend policies that make smoking more costly and less pleasant with language like “helping smokers quit”, a standard response is that this is utter crap:  Intentionally inflicting pain on someone to try to modify their behavior is called “torture”, not “help”, and it violates all accepted modern Western rules of ethical behavior.  But while this response is solid, and the ANTZ position is indeed crap, there is a bit more to it.  The missing nuance is important not just for getting the argument completely right, but for understanding why non-ANTZ might buy into the ANTZ position (and thus, coming back to the theme of this series, how to respond).

When the ANTZ claim that torture is help, they are invoking the implicit (and clearly false) claim that everyone affected by the torture has second-order preferences for quitting, and also invoking the implicit (and clearly false) claim that they want someone to use force to align their preferences with their second-order preferences by inflicting torture until quitting is preferable to smoking.  In addition, the ANTZ are making the implicit (and clearly false) claim that all of these smokers who want to be forced really understand that when they are abstinent they are likely to be less happy/productive/etc. compared to when they were smoking.

If all three of those implicit claims were actually true, then there would be a legitimate argument that the current anti-smoking approaches are ethical, and perhaps even an argument that anti-THR is justified (though anti-THR lies are never justified).  To be sure, under those conditions the position would be defensible, but there would still be some strong ethical arguments against it (which I will not address in this series).  But a lot of people, having never thought seriously about the ethics, have a gut notion of ethics that says that under those three conditions, that position is Right.  The biggest problem is that they have been tricked into believing that those conditions are met, and that is optimal point of response.

And with that, I have to leave the issue of how to try to un-trick them until the next post, having already passed our “no post longer than…” limit.