If anecdotes are evidence, why aren’t you drinking paint thinner?

I want to begin this post by doing something atypical for me. I want to tell you about an amazing cure-all that I that was recently introduced to: turpentine (aka paint thinner). According to the vast wealth of knowledge available on the internet, most (if not all) diseases are actually caused by parasites, fungal infections (particularly Candida), and even modern medicine itself. Don’t worry, however, because all of these can be cured by drinking turpentine (or sometimes kerosene or even gasoline). Now, you may think that sounds crazy, but have no fear, because this treatment is totally natural (turpentine is made from distilled tree resin). Also, it has been used for nearly two centuries, and several brave doctors have bucked the medical establishment and are promoting it (e.g., Jennifer Daniels). You may think that is pretty flimsy evidence, but don’t worry, I also have multiple blogs, alternative health websites, and Youtube videos explaining why this is the cheap trick doctors don’t want you to know. Best of all, I have tons of anecdotes. There are countless success stories of people who tried traditional medicines to no avail, but as soon as they started drinking turpentine, their symptoms went away and they could just tell that they were healthier. Take, for example, this person who wrote the following after taking turpentine, “My energy level is so much better, lungs feel cleaner. Can’t tell me this stuff doesn’t work.” With confidence like that, how could they be wrong? Finally, you may be wondering why there aren’t a lot of scientific studies supporting turpentine as a treatment, as well as why there are lots of health recommendations against taking it. The answer is simple: big pharma only cares about profits, so they are suppressing the truth of this amazing treatment.

As most of you have hopefully guessed, the paragraph above is facetious, and I’m not going to try to induct you into a pyramid scheme, but I wanted to open this post that way to illustrate a very important point. Namely, most of the people reading this probably spotted the flaws in my arguments for turpentine. The idea that drinking paint thinner could cure all diseases is so outlandish that you probably realized that blogs, Youtube videos, and anecdotes aren’t sufficient evidence. You probably realized that the fact that something is natural or ancient doesn’t mean it’s safe or effective (appeal to nature and appeal to antiquity fallacies). You probably realized that the fact that I found a handful of doctors that support drinking turpentine doesn’t mean that it works (appeal to authority fallacy), and you probably scoffed at the notion that safety warnings on turpentine were actually part of a conspiracy by “Big Pharma.”

Nevertheless, despite all of that, a large portion of you probably use identical reasoning to support your favorite alternative remedy. Based on what I see in the comments, most of you probably have some “cure for the common cold” or other pseudoscientific practice that you cling to dearly, and if I asked you for your evidence, you would respond with the exact same type of reasoning. Most prominently, you would give me anecdotes and cite blogs and Youtube videos.

Further, on the off chance that someone reading this believes in the magic powers of turpentine, there is still almost certainly some other alternative practice that you think is nuts, even though it is supported by the exact same evidence base. For example, I used to know someone who believed in all manner of nonsense, from crystal healing to anti-vaccine conspiracy theories, but they drew the line at homeopathy. As I tried to explain to them, however, that doesn’t make sense because homeopathy has the same evidence base as things like crystal healing. In other words, when I asked them to give me evidence of crystal healing, they replied with blogs, Youtube videos, and anecdotes, yet they rejected homeopathy even though homeopathy is also “supported” by countless blogs, Youtube videos, and anecdotes. To try to make them grasp this paradox, I once asked them, “If homeopathy doesn’t work, then why do so many people claim to feel better after taking it?” They very correctly responded that those reports could be from placebo effects, total coincidences, regression to the mean [technically a type of placebo effect], other medications, etc. In other words, when it wasn’t their pet belief, they had no problem seeing the flaws in the line of reasoning, but when it was their personal views at stake, suddenly cognitive biases clouded their vision and inhibited their ability to think logically.

The point that I’m trying to make here is that your reasoning has to be consistent. Either anecdotes can establish causation or they can’t. You don’t get to pick and choose when you think that they work. In other words, if an anecdote, or even a collection of anecdotes, is actually sufficient grounds for saying that cannabis cures cancer, acupuncture works, vaccines cause autism, etc. then it must also be sufficient grounds for the effectiveness of homeopathy, miracle mineral solution, bleach enemas, turpentine, kerosene, gasoline, crystal healing, bloodletting, leaches, sacrificing to the sun god, and every other form of woo that has ever been proposed, because they all have anecdotes. If anecdotes actually can establish causation, then you have to believe in all of them. They can’t only establish causation when you want them to. That’s not how evidence works.

To put that another way, if for any one of the thousands of alternative treatments that have ever existed, you are content to say, “the anecdotes could easily be from placebo effects or other factors,” then you must say that for all of the treatments. In other words, by acknowledging even once that the fact that someone took a treatment then got better is not good evidence that the treatment actually works, you have just universally acknowledged that anecdotes can’t establish causation. The logical syllogism, “someone took X, then got better, therefore X works” either works all the time or it never works. It can’t magically work when you want it to, then not work when you don’t want it to.

The same thing is true for the admissibility of blogs and Youtube videos as evidence. If you asked me for evidence that turpentine is a cure-all, and I responded with an unsubstantiated Youtube video, you would very correctly demand actual data. It is inherently obvious that any crackpot can make a Youtube video and say whatever they want in it. To be clear, there are some Youtube videos, blogs, etc. that are packed with non-cherry-picked citations to the original peer-reviewed literature, and there is nothing wrong with linking to a source like that and saying, “this video gives a good explanation and cites the relevant literature.” That is, however, almost never what I see when it comes to conspiracy theories and alternative medicine. The sources that I see people use as evidence are nearly always just someone spouting nonsense as if they were stating facts, and citations to original studies are either non-existent or horrible cherry-picked.

Finally, I want to contrast this type of inconsistency with a science-based view of reality. To put it simply, you can convince me (and scientists in general) of anything if you have sufficient evidence, and by evidence, I mean multiple independent studies that used large sample sizes, adequate controls, and rigorous analyses. If you can show me a consistent body of scientific evidence demonstrating that drinking turpentine is safe and effective, I’ll accept it. If you can show me a consistent body of evidence demonstrating that vaccines cause autism, I’ll accept it. If you can show me a consistent body of evidence demonstrating that crystal healing actually works, I’ll even accept that. Do you see the difference between that and cherry-picking when you do and do not want to accept anecdotes as evidence? Science has consistent criteria for what is and is not evidence, whereas there is no constancy in pseudoscience.

The take home that I want you to get from this is that you need to ensure that your reasoning is consistent. A great way to do this is by trying to think of situations where you would not accept the conclusion that results from your current line of reasoning. For example, if you are using an anecdote to claim that a particular alternative treatment works, stop and try to think of situations where you would not accept anecdotes as evidence. In other words, if you can think of a situation where you wouldn’t accept that X caused Y, even though someone took X then Y happened, then you have just demonstrated that your line of reasoning is flawed, and anecdotes are not sufficient evidence of causation.

Note: To be clear, I am not arguing that the existence of anecdotes is evidence that something doesn’t work (that would be a fallacy fallacy). In other words, when I said things like, “the logical syllogism, ‘someone took X, then got better, therefore X works’ either works all the time or it never works,” it is the syllogism itself that is the problem, not its conclusion. To put that another way, there will always be anecdotes for things that actually do work. The problem is simply using those anecdotes as evidence that it works.

Note: Inevitably when I start talking about anecdotes, pedants get all bent out of shape and argue that anecdotes do have value because they indicate that something may be worth studying. I agree, and never said anything to the contrary. That argument does not, however, in any way shape or form negate my point that anecdotes are not valid evidence of causation. So please spare me your pointless pedantry.

 Related posts

This entry was posted in Vaccines/Alternative Medicine and tagged , , , , , , , , . Bookmark the permalink.

16 Responses to If anecdotes are evidence, why aren’t you drinking paint thinner?

  1. mvschulze says:

    You might want to add God to the chart. (not to be argumentative or even controversial – just pondering here…) God, would arguably garnish an “X”. Further it could be said that the power of belief achieves positive well-being – mental well-being. Accordingly, If one believes that an ingested substance that is actually inert or even detrimental, produces mental well being, then it may, in that person’s perspective, be considered a benefit, – it promotes (causes?) well-being. Perhaps the saying “Mind over matter” would apply here. Back to God… is that not one of the most powerful examples (in the human experience,) of unsubstantiated belief?
    I personally advocate science, not conventional religious belief. But it seems the majority of our 7+ billion living inhabitants still find solace (well-being) in unsubstantiated beliefs.
    Excellent post. Thank You. M 🙂


  2. horvendile2 says:

    There is one element you didn’t address, inherent plausibility. Turpentine though ridiculous is more plausible than crystals or homeopathy as the proposed mechanism, chemistry has been shown to exist where there is no known mechanism for crystals or homeopathy that would involve revising our basic understanding of the laws of physics.


    • Fallacy Man says:

      There’s an extent to which I agree with you, because prior probability is important and, as the axiom goes, extraordinary claims require extraordinary evidence. Nevertheless, the central point stands that the reasoning behind using anecdotes to establish causation is flawed. In other words, either the syllogism, “x happened before y, therefore x caused y” works or it doesn’t. So if other explanations for y (like placebo effects) negate the syllogism for crystal healing, then must also do so for chemical based treatments like turpentine.


  3. Girl with the red shoes says:

    A great post. It horrifies me that people believe information from unqualified people on Youtube and blogs that are not supported with any credible sources. There’s a lot to be said for having even a basic understanding of science and scientific reasoning, and applying said reasoning consistently.


  4. Chris Herrington says:

    So… are we going to ignore the countless good, peer reviewed studies on acupuncture and chiropractic?


    • Fallacy Man says:

      There are very, very few good studies on them that found positive results. There are tons of crap studies with inadequate controls, small sample sizes, and shoddy statistics, but good studies with a positive result are few and far between. Further, you have to always keep in mind that sometimes you get false positives just by chance. Therefore, you have to look at the entire body of literature, not just the handful of studies that support your beliefs, and when you do that, the evidence against acupuncture and chiropractic is pretty overwhelming. Tons of studies have found that they are no better than a placebo.

      It’s also worth mentioning that many of the positive studies come from countries that are known to have massive publication biases, thus making this their results suspect.

      This post gives a good overview of the evidence against acupuncture with lots of links to the peer-reviewed literature.

      Here are two particularly important ones (they are systematic reviews of systematic reviews)

      Similarly, for chiropractic, here is a systematic review of systematic reviews that concluded,
      “Collectively these data do not demonstrate that spinal manipulation is an effective intervention for any condition. Given the possibility of adverse effects, this review does not suggest that spinal manipulation is a recommendable treatment.”

      Also, here is an important review of subluxation (the concept on which chiropractic is based)


      • Chris Herrington says:

        Obviously if we ignore bias-ness, then there’s the inherent problem of cognitive bias that also has to be addressed with the fact that most extensive researchers doing longitudinal studies, are not well versed in the musculoskeletal realm. Double blind, randomized trials can’t be conducted with any manual therapy, because of the existence of therapeutic touch. Humans typically respond positively when someone is touching them with the intent to help them. This compounded with placebo, does in fact make studies difficult to assess, and the strictest of researchers will consistently point this out, stating it invalidates all manual interventions.

        This is where cognitive bias comes into play, because these researchers are typically NOT clinicians, and are not in the business of caring for people, simply pointing out trends and academic flaws. They only care what is on paper, not what is standing before them. If as a researcher you are trained in allopathic medicine and it’s benefits, then manual therapies will be completely insane to you. Such as many manual therapists don’t understand complicated pharmacological. Humans can only perceive what they are trained in, and as such, have inherent ignorances they must acknowledge.

        The last inherent problem I feel I must point out, is the academic ego flaunting that takes place in scholarly articles, currently. For any researcher who publishes good results, some other researcher is going to now inherently publish results to counter the initial. This is one of the biggest problems with manual therapy, as the hardest/strictest of researchers wants to see a cause/effect relationship, when the mechanism that is being studied is not that simple.

        Modern medicine wants everything to be cut and dry, black and white. Treating the body, without chemicals or surgery inherently cannot be this simple, as the free flowing biochemical reaction that we call “living” changes by the millisecond. There are trends we can follow, but as of yet, there is no level of technology that allows us to demonstrate the instantaneous changes in biochemistry in vitro. As such any treatment that didn’t treat doesn’t permanently deform or alter the body will be seen by some as “useless” or “ineffective”. Physically changing the mechanics of the body via manipulation, does alter biochemical interactions and immediately changes inflammatory chemicals and endorphin release. This much has been quantified and extensively studied. It can’t be said that manipulations do nothing, but they certainly don’t cure anything. This is the problem most researchers have with the fields. They want definite answers in an indefinite mechanism.

        As for research: you cite that meta-analysis demonstrates manual therapies as ineffective, yet this demonstrates the opposite: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC3997823/

        Manual therapy and placebo:

        This isn’t to say this is any “end all be all” treatment. It’s good for what it’s meant to do, aid in the resolving of musculoskeletal problems that are not severe enough to warrant surgical intervention. In the US, chronic pain is an epidemic, and modern pharmacologics aren’t doing anything to treat it. Movement based problems can only be solved by restoring movement. Cutting doesn’t fix that.

        Is manipulation the only fix? Absolutely not. Is it a viable treatment? For some, yes. Physical therapy also works, and is extensively researched.

        The biggest hurdle is resolving pain and restoring activity levels naturally. Regardless of how science says is the best way, the result is the answer. If the result is attained, then the means worked.

        No healthcare intervention should be continued long term (in th absence of disease) Seeing a chiropractor, physical therapist or acupuncturist weekly (or even monthly) is a choice, and should not be a mandate. Just like taking pain medication shouldn’t be continued long term. The best practitioners are getting their patients better, educating them on how to sustain themselves, and discharging them from care. They exist in all fields, but are diamonds amoungst piles of coal. Every field has bad members, every field has good.

        If chiropractic/acupuncture/physical therapy/etc were utterly pointless and fraudulent, they wouldn’t be federally licensed and regulated, nor would they be allowed to have multiple accredited colleges.


        • Chris Herrington says:

          Side note: I purposely didn’t address subluxation (vertebral, not the partial dislocation that is quantified radiologically), because I agree, it has never been quantified or proven. It’s a sham that needs to be lose to the sands of time.


        • Fallacy Man says:

          Our agreement seems to be a fundamental one about the very nature of evidence and conclusions.

          For example, you said, “This is one of the biggest problems with manual therapy, as the hardest/strictest of researchers wants to see a cause/effect relationship, when the mechanism that is being studied is not that simple.”

          This is a cop-out that believers in woo love to use. They assert (as you seem to be) that their favorite pseudoscience simply isn’t conducive to the practices of science. So, according to them, the negative results aren’t actually evidence against the treatment. That response is ad hoc and it presents a non-falsifiable situation, which, by definition, makes it pseudoscience. Of course we want a cause and effect relationships. You have to demonstrate causation before you can say that a treatment works. That is a fundamental requirement of science-based medicine.

          Similarly, you said, “Regardless of how science says is the best way, the result is the answer.” here again, we can only the know the result via science. In other words, if we can’t see the result via carefully controlled studies, then we have no reason to be confident in it. It puts us back into the pre-science realm of basing practices off anecdotes and conjecture.

          Regarding the studies you cited, the review of reviews most definitely did not “demonstrate the opposite.” To quote, “Most ‘inconclusive’ or ‘moderate’ evidence ratings of the UK evidence report were confirmed. Evidence ratings changed in a positive direction from inconclusive to moderate evidence ratings in only three cases.” One of those three cases also involved exercise, so it seems silly to the manual therapy all the credit. Further, by the paper’s own definitions, the evidence for those three was “moderate.” That’s hardly compelling. This screams of a type 1 error. Whenever you have a treatment that is claimed to do a broad range of things, and you test all of those things, one or two will nearly always come up with weak positives just by chance. That’s just a by-product of the statistics that we use. So when you have a handful of weak positives amidst a sea of negatives, it’s silly to latch onto those weak positives as evidence.

          The second link is a single study. Again, we have to look at the whole body of literature, not cherry-picked studies.

          The third link simply points out that there are a lot of low-quality studies in this field, which was one of the points that I made previously, so I’m not sure why you think it helps your case.


          • Chris Herrington says:

            Then the only plausible explanation is that what is happening is magic. If science and tech can’t explain it, and the results are reproducible, then it can’t be rationally understood.

            The logical downside to such hard-lined science opinions as your own, when applied to healthcare, is that humans aren’t that well understood. Our methods of explaining the world aren’t always accurate, and when researchers with hard-line, unforgiving science stances try to understand healthcare, they always leave out the fact that humans are being hurt by the very ideas the researcher wants to ignore.
            (Fallacy alert, but honest questions)

            You want the world to be black and white, but what about all the gray areas that definitely exist. If a person in pain was to come to you and say they aren’t a surgical candidate, and are allergic to NSAIDs, have joint problems from decades of steroidal use, and are too obese to exercise, which thus leads to them to over eat due to stress and self medicating with foods. Nutritional consults don’t work here. What study are you going to direct them to? On paper, they are “bleeding the system” with their inability to take care of them-self, but regardless of what paper and policy says, they are still a human and deserve someone to care for them.

            We don’t have studies to prove a lot of things. Medicines are routinely used off label for things that doctors found they work for. I’ve never seen anyone write an article about how horrible that is, and that it needs to be stopped. What about all the elective surgeries that occur in the nation daily? There’s no reason for them to occur other than poor diagnostic abilities, but there aren’t enough hours in the day, nor skilled clinicians to catch everything.

            Alternative medicine has a substantial problem with charlatans and snake oil con-artists. No doubt. However, that is a small, yet loud population.

            Instead of pointing out how bad the research is for these fields, instead of saying they are pointless, why not dedicate a portion of your superior research knowledge to create a framework for their fields to conduct research in a proper way? If they are fraudulent, let them fail on their own accord, rather than on the basis of cognitive bias or lack of research?


            • Chris Herrington says:

              Then the only plausible explanation is that what is happening is magic.

              That’s a bizarre leap of logic that the author has neither made nor implied. In fact, on the original blog post he lists several explanations for why people may think something works when it does not work.

              Chris Herrington says:

              If science and tech can’t explain it, and the results are reproducible, then it can’t be ratiobnally understood.

              You are confusing various things and creating a contradistinction that does not exist. In our context, science can be about determining mechanistic explanations for a phenomenon or it can be about observation of causes and effects without looking at underlying mechanisms (reproducibility).

              In the studies we are talking about, which relate to therapeutic applications, we are interested primarily in testing hypotheses regarding a particular treatment having a therapeutic effect and only secondarily interested in the underlying mechanisms of a potential effect. Because this is applied science and we are interested in achieving therapeutic effects (that is the application), reproducibility here is what science is all about. Of course, knowledge of mechanisms of action can used to (and often is used to) generate useful hypothesis about something having potential particular applications but this need not be the case and has often not been the case in the past. That is, we have had plenty of therapies without first knowing how they work (and we still do). They can be and are tested for safety and effectiveness without knowing how they work.

              In other words, the kind of reproducibility we are interested in here is something we assess using science so there is no opposition between science and reproducibility at all. There is also the kind of reproducibility where we “know” anecdotally that something works because a lot of people do it and report good results (this is the kind of “reproducibility” that tells us that blood letting and homeopathy work). This is less interesting to us and has limited value.


            • Fallacy Man says:

              Your arguments aren’t making any sense.

              You said, “Then the only plausible explanation is that what is happening is magic.”
              What? Placebo effects, regression to the mean, and countless other explanations are way more plausible then magic.

              You said, “If science and tech can’t explain it, and the results are reproducible, then it can’t be rationally understood.”
              I’m not actually sure what you are asserting here. The “benefits” of alternative treatments like acupuncture aren’t reproducible. The citations in my first comment demonstrate that. Or, at the very least, the notion that they are better than a placebo isn’t reproducible. People certainly do report feeling better after taking them, but science has explained the reason for that (again, placebo effects, regression to the mean, etc.). Further, don’t mistake “science can’t explain” and “science hasn’t explained yet.” It is entirely possible for something to pass clinical testing and be shown to work before scientists have figured out exactly why it works. That is not, however, the same thing as science being unable to explain it. Further, once again, the problem here is that alternative treatments don’t pass clinical tests. They fail to consistently produce better results than a placebo.

              You said, “The logical downside to such hard-lined science opinions as your own, when applied to healthcare, is that humans aren’t that well understood.”
              First, humans are pretty well-understood. Second, there certainly are some things we don’t understand, but the only way that we will ever understand them is through science! You can’t say, “science is imperfect, therefore my random unsubstantiated crap is fine.”

              You said, “they are still a human and deserve someone to care for them.”
              Yes, but that care needs to be based on actual evidence, not anecdotes and conjecture.

              You said, “Medicines are routinely used off label for things that doctors found they work for.”
              Do you have a citation for this? That seems like a pretty serious breach of ethical conduct. If it is actually occurring, then I (and indeed most skeptics) would be strongly opposed to it.

              You said, “Instead of pointing out how bad the research is for these fields, instead of saying they are pointless, why not dedicate a portion of your superior research knowledge to create a framework for their fields to conduct research in a proper way?”
              Because proper studies have been done and these methods have failed. Again, read the reviews I cited earlier.

              This is going nowhere and you are constantly shifting the goal posts, so this will probably be my last comment, but I’d like to end it by going back to where all of this started. Your initial assertion was that there was high quality scientific evidence for chiropractic and acupuncture. That claim is blatantly false, and the sources I provided demonstrate that. Following that, you shifted the goal posts to argue that we shouldn’t rely solely on scientific studies, to which I say, read the original article to see why that doesn’t work. As I said in the article itself, if you can show my an actual consistent body of high quality scientific evidence supporting acupuncture, homeopathy, etc. I’ll accept it, but unless you can do that, there is no reason to accept them as valid treatments (and given that they have failed countless trials, there is very good reason to reject them).

              Liked by 2 people

  5. I want to begin this post by doing something atypical for me. I want to tell you about an amazing cure-all that I that was recently introduced to: turpentine (aka paint thinner). According to the vast wealth of knowledge available on the internet, most (if not all) diseases are actually caused by parasites, fungal infections (particularly Candida), and even modern medicine itself. Don’t worry, however, because all of these can be cured by drinking turpentine…

    Unfortunately, this post is based on a false premise. If turpentine ingestion is associated with improved outcomes it is only because it addresses an underlying condition of turpentine deficiency.


    • Vegard says:

      You mean the same way antibiotics treat infections by removing the underlying condition of antibiotics deficiency, or chemotherapy cure cancer by treating chemotherapy deficiency?

      Liked by 1 person

  6. While I applaud the correctness of your argument, I think the real problem is our bias towards narratives (related to confirmation bias but not exactly the same thing). Hence “natural” is better because God and Nature work together to offer cures for all possible problems. Paint thinner and gasoline may be just as natural, but they occupy a very different niche in day to day life than say, oil of oregano. It just seems that paint thinner must be bad because if there’s any paint in your body, it should not be thinned.

    To offer a strong example in the other direction, many chemotherapy agents are related to mustard gas, which simply must be bad for you. Just a glance at some old photographs offers horrific proof. The repulsion is visceral. It takes conscious and abstract reasoning to work out that the same powerful poison might be useful if used carefully.

    Liked by 1 person

  7. xaspierudegirlx says:

    My favorite remedy that circled for sometime was drinking bleach. xD ‘Miracle Mineral Solution” Not only did it not have much research or evidence to back it up and there were unknowns etc, what person thinks drinking a form of bleach is a good idea for anything? Like sure it may cure you of whatever you have, but does drinking bleach sound healthy?! Like sure my child is so called cured of his autism (Autism is not curable nor does it need to be. Lol) but should we be concerned with what they consumed? What about the other health risks caused by digesting bleach? Another one I am seeing circle again is Chelation Therapy which is still being studied heavily but shares many risks and unknowns, one of the side effects is still in fact possible death. I am all for finding cures and remedies for certain things, but with things like death even being a possible side effect, how desperate are you?! People are nuts!

    Liked by 1 person

Comments are closed.