 Imagine a hypothetical friend of mine. We'll call him Bob. Suppose Bob goes tomorrow and buys a scratch-off lottery ticket. He wins $100 on a $1 ticket. Bob goes back the next day and again he wins $100. The action and the result is the same on the third day. In that moment, Bob decides that he is going to quit his job, sell his house, and become a professional scratch-off lottery player. Based on his experience, he should be able to make thousands of dollars a day. How do you think this will play out for Bob? Will he be able to buy a mansion and a yacht with all his winnings by the end of the first year? Probably not. But how can you be so confident that his plan is foolish? What knowledge prevents us from making the same conclusion that Bob does? If we calculated the chance of Bob winning three scratch-offs in a row, we would find it to be quite low, but it did indeed happen for him. From his perspective, the historical chances of winning the first three games was one-in-one. If he then extrapolates that probability to all future chances of winning, he makes the wrong decision to become a professional lottery player. In the trivial sense, Bob has committed the fallacy of the hasty generalization, also called the fallacy of insufficient statistics. Bob's taken too small a sample and generalized it prematurely. Over time, he's most likely going to spend his last dollar trying to replicate this unlikely outcome. In fact, all gambling is based on our inability to comprehend statistics and sample size. Over the very long term, the house always wins, which is why casino owners tend to be very wealthy and powerful. The only legal way to have any positive outcome for the player is to quit after a low probability favorable outcome. But there's a deeper fallacy I want to address here. It's about the value we place on anecdotes. I'll freely confess that I'm quite guilty of this fallacy about some topics, and I see from the comments on some of my other videos that some of you are too. An anecdote is a self-contained story about an event. I'm specifically referring to anecdotes about alternative medicine or paranormal or supernatural phenomenon. For example, the one I heard a lot on my video about chiropractic was, it may not be scientific, but it cured my headaches or back pain, neuralgia, whatever, when the doctors couldn't. Or the ever-popular anecdote that goes, I knew this woman who used homeopathy or whatever, and it cured her cancer after the doctors had sent her home to die. Why isn't that a compelling argument for an alternative therapy? For the same reason that Bob is never going to get rich scratching off lottery tickets. The sample size isn't sufficient to generalize from. Yes, one patient may have subjectively responded to a therapy in a positive way, but that doesn't tell us very much about how the next patient will respond. It's only when the sample size becomes significant that we can begin to draw conclusions. That's why science-based medicine values large population studies more than case reports or amazing anecdotes of miraculous cures. And yet, when we view alternative medicine claims, they often rely heavily on anecdotal evidence and studies in very small populations. Case studies are likewise of limited utility. They do give us a starting point for more rigorous investigations, but they don't possess the power to tell us about risks or efficacy. If you're in a debate with someone who is citing papers to support their views, one thing to check is the size of the populations used, usually indicated by the letter N as in N equals 100. If groups are matched to each other, we usually call them cohorts. The control group, and there needs to be a control group if we're eliminating placebo effect, and the treatment group should be roughly the same in size and composition. How big the cohorts are determines the statistical power of the test. A clinical drug trial may be evaluated on cohorts of thousands or even tens of thousands. Meanwhile, alternative medicine is often conducted on very small groups, less than 50. If they ever want to be taken seriously by those of us who value reason and logic, they need to correct this methodological flaw. A homeopath referred me to a systematic review paper to cite its effectiveness, apparently without having really understood the outcome of the study. The homeopath authors concluded that homeopathy could be demonstrated to be effective only in very small populations. For the eight largest trials, there was no detectable difference between placebo and homeopathy. Completely unrelated to the previous point is another fallacy I want to deal with while I'm on the topic. This is the treatment X isn't possible to evaluate in the clinical study. The argument being made is that a treatment only works for patients who believe in the therapy or who receive proper indoctrination or training. This is a claim I hear a lot for holistic medicine, Reiki, Chi, and other energy healing modalities. They assert that a receptive attitude is necessary for the treatments to be effective. I've even seen outcome studies stratified, that is, separated out by the patient's self-reported faith in the treatment. Of course, we have a name for treatments that rely on the beliefs of the patients to be effective. The term is placebo. This is a big enough topic for its own video. I'll briefly say that I respect the medical value of a placebo. It has a powerful biological effect. However, I don't think it's ethical to give a seriously ill patient a placebo in place of effective therapy. I know I'd rather have a realistic view of the outcome of any treatment I'm receiving. And I'd expect an ethical clinician to use placebo only in cases of self-limiting disorders or where there is no effective treatment. This leads me to my last point. Alternative medicine is often touted by proponents as having fewer side effects than conventional medicine. This is sometimes true. Homeopathy is just water in most cases, so it's not likely to cause allergic reactions or kidney damage. But the point is that it has the same effectiveness as giving sugar pills, but more expensive to the patient. And in some cases, it displaces the more effective evidence-based therapy, which is its own kind of side effect. We don't choose our treatments on their side effects alone. We choose them for their ability to treat the disorder, and we balance that against the possible risks. This is something we do in any human endeavor. It's a logical consequence that anything we choose that has an actual biological effect might have multiple effects, some undesirable. The argument being employed here is called the perfect solution fallacy, a form of the false dilemma. It proposes that we reject workable solutions because they don't completely eliminate the problem, or they have some negative consequence. One favorite statistic is the number of people who die from drug interactions, or medical mistakes, as though this validates the non-effective treatment of homeopathy. I've been accused by some of being naive. I find this very amusing, as though my personal evidence-based standard of what therapy I choose is somehow an unreachable goal. I've also been chided for demonstrating the weaknesses of alternative medicine, but not the weaknesses of conventional medicine. To do so would be a false balance. Of course, we should continuously strive for better medicine, to reduce side effects, to find out the best solution for the patient. But the methods we need to employ to make that improvement are evidence-based methods. I don't attack alternative medicine for being mystical. I attack it for not being demonstrated to work. So, to summarize, the basis for gambling and alternative medicine have an element in common, and exploit the same flaw in our thinking. They rely on a hasty generalization from a small sample size. So the next time you're tempted to try some pseudoscientific therapy because a trusted friend relates a positive anecdote, remember Bob, the professional lottery player, and look for the outcomes of large population studies instead. I'll leave you with two favorite quotes on the topic from Roger Brenner. The plural of anecdote is not data. And from that great scientist that kept us laughing, Richard Feynman, science is what we have learned about how to keep from fooling ourselves. As always, the best advice comes from a trusted physician. Stay safe and be smart. Thanks for watching.