Philosophy

Why You Should Care About The Philosophy Of Probability & Statistics

A RARE case where the probability is easy to compute.

“Why, Dr Briggs,” asks the earnest Student, “Why must we study all this gobbledegook about philosophy when it’s more important to learn all the neat methods of handling data? Why can’t we just get right to it like in all those other textbooks and classes?”

Well, I’m glad you asked, merry Student. Reason you can’t get right to playing with data is because you don’t understand the point of playing with data. And what is that point? The answer is a philosophical one. The philosophy of uncertainty and epistemology.

And what’s uncertainty? I can tell you what it isn’t: a mathematical formula. Nor is it a picture on a computer screen, nor a table of data.

Let’s take an example (kindly provided by reader Rob Ryan). We want the truth of the proposition, “Living near high voltage towers causes cancer.” The first step in any logical argument (the first act of the mind) is to come to terms with what we mean. The proposition is in several parts—“living near”, “high voltage towers”, “causes cancer”—each of which must be clearly defined.

There are different kinds of “high voltage towers” in various condition, so it would be best if we can identify those elements of the towers which, given other knowledge we have, can plausibly cause cancer. Some sort of dose measurement of the electrical field would be superior to “distance” to nearest tower. Distance can be ambiguous, and because of background effects two people at the same distances in difference background could have widely different exposure.

Now regardless of our eventual evidence (defined below), the probability the proposition “causes breast cancer or lung cancer or skin cancer or toe cancer or brain cancer or …” must be larger than “causes breast cancer” alone because the expanded proposition has more opportunity to be true, so to speak. Which version are we using?

“Living near”, as already hinted, can mean many things. Such as “Having a mailing address within 10 miles”, “Dose of electricity at airport nearest person’s zip code”, “Having a place of work which uses large amounts of electricity”, “Living, from just moving in yesterday to having been there fifty years, within X yards” (where X is allowed to vary at will), “Spending time near”, and on and on. All of these, even the ridiculous ones, have been used by epidemiologists in similar studies, the differences between them being dismissed as not terribly important.

Whatever we have settled on, an indicator of “living near” or a dose estimate, and whatever cancer is of interest, any or specific, we now have to collect data. From where? From the region near a home where a man who swears the power lines gave him bone joint cancer? From areas where there were no reports? Just in one state? Several? Just the USA? From what dates? Historical? Just now? How long into the future? People come and go so this isn’t trivial.

https://twitter.com/mattstat/status/439410623919583232

Then there are the people. Who should we use? Those over 50, which gives them time to develop cancer? Everybody? All “incomes”. Smokers? Democrats and those who engage in other harmful behaviors? Potato chip eaters? What about those folks who have cancer but never have it diagnosed? And what of those who don’t have it but are falsely diagnosed. Or what about the nascent cases whose cancer is caused by electricity but where it doesn’t manifest until after the data collection?

Either all that is sorted out or it is ignored. If ignored, all those still considerations matter, but now we’re flying blind. This is because as each new piece of data is added to the list we can update the probability our (now fixed meaning) proposition is true. Suppose we settled on just one county in one state, Utah, for example. Now whether all those other things are spelled out, it is still the case that the probability we figure is conditional on all those things. The probability is thus valid only for those combinations of characteristics, such as “for this county in Utah, for those who live in this location, over 50, non-smokers, etc., etc.”

Whether or not the probability is valid for other characteristics is not something we can say using the data we have. Hold up. Let me repeat this most important point. Whether or not the probability is valid for other characteristics is not something we can say using the data we have.

At this point enters the statistical analysis (we could choose one of several procedures), which strips away all the most important and interesting uncertainties and focuses on a few numbers. The problem becomes the vast simplification “Cancer with Exposure?”. In order of most to least useless, a p-value will be produced, or a posterior, or maybe even what we really want, the probability our proposition is true.

But it will invariably be the wrong, misleading, blinding answer because we ignored the most important aspects of uncertainty.

Yet because the statistical procedure produces a number, and numbers are science, that number becomes everything, the only thing. It is all anybody remembers. Strike that. That number will be transformed to either “Living near power lines does causes cancer” or, very rarely, “It might not”.

And that is why there is a massive epidemic of “scientific” over-certainty.

And that is why the philosophy is more important than a few computer tricks.

Categories: Philosophy, Statistics

54 replies »

  1. Reblogged on “Watching those who watch the Deniers”.

    An excellent explanation of what goes into statitics rather than just those cool data manipulations that are so loved by much of science. Thank youl.

  2. Pshaw!

    Next you’ll be telling me that reports of white swans don’t confirm “All swans are white”…

    Although, I’d consider arguments that purported one black swan doesn’t vitiate “All swans…” That’s just me.

    We need to return to -and resolve- the ‘problem’ of induction. Start with Hume; then go back to Aristotle. (Please skip Descartes entirely…)

    It shouldn’t be that hard. Not for the Statistician to the Stars…

  3. Understanding the physics of whatever is being analyzed is crucial.

    Something like ‘distance from a high voltage tower vs. cancer” must be analyzed in terms of objective facts — like how much electromagnetic flux is present where people are exposed (not all towers are equal in output).

    One need not understand the “philosophy” at all if they get the proper physics correct.

    Conversely, they can get the philosophy right and if they don’t get the quantitative physics factored in the analysis is useless. It may be logically unassailable, but still useless.

  4. most to lease*

    The goal in the games of apartment building ownership and Monopoly.

    * as in In order of most to lease useless …

  5. One need not understand the “philosophy” at all if they get the proper physics correct.

    The best thing physics will give you is an approximate answer to how much exposure each individual could have had provided they weren’t shielded in some way. Still, there will always be unknowns and complicating factors such as:

    o should driving by one in a car twice a day count?
    o what about exposure to the 1200v conduit in your office building that runs through the wall just inches behind you back?

    Remember, the question doesn’t involve just physics but physiology.

    Frankly, studies of this nature are just plain pointless. Why not just expose lab animals to fields of varying strength at 60 Hz. or whatever then go around measuring the field at various locations (vs. flat out guessing)? That, too, has its own problems.

  6. Ken & DAV, the main thing to take account of is that it is non-ionizing radiation and thus can not break chemical bonds.

  7. Scotian,

    Re: non-ionizing radiation. True, true. But the field can heat and heat can do things. Enough energy? Who knows?

  8. Ken & DAV, the main thing to take account of is that it is non-ionizing radiation and thus can not break chemical bonds.

    That doesn’t mean it can’t have an effect. Nobody really knows what causes cancer despite the many theories.

  9. Briggs “But the field can heat and heat can do things”. It can keep you warm on a cold winter’s night, it can boil water, and cook your meals. Stray low frequency e/m fields, not so much but I would be careful not to put your miniature poodle in the microwave to dry.

    DAV, that’s what it means.

  10. I would be careful not to put your miniature poodle in the microwave to dry.

    Don’t see why not. Poodles look like mops with four legs. Should dry the microwave just fine.

    BTW: the electric field potential can cause ionization at 0 Hz. Ever seen a “static” discharge? It’s around 20-30Kv per inch for dry air. What’s the breakdown voltage of chromosome bonds? And that’s assuming they would have to break to cause cancer. Maybe the interference alone can do so and then maybe only if other factors are present.

  11. DAV: Which is precisely why I own a small dog named “Dusty” (as in dust mop) and a very, very small microwave! Eliminates temptation! 🙂

  12. DAV, you can add: don’t play golf in a thunder storm to the list. In any case this is not e/m radiation but the flow of high current (electrons) through delicate tissue. The discharge flash includes a range of frequencies including the ultraviolet which can break bonds. You have a lot of maybes.

    I’d be careful not to get the dog lovers riled up. Those miniature poodles can be vicious.

  13. not e/m radiation but the flow of high current (electrons) through delicate tissue.

    So? Can you say for certain that entering/exiting/moving within the field doesn’t cause damaging electron flow? How do you know it has to be high (whatever that is) current?

    It’s being near the lines that is the question. For whatever reason. How do you know it’s not the radiation as well?

    you can add: don’t play golf in a thunder storm

    Good point. Some golfers have died from cancer. Maybe you’re on to something.OTOH, maybe it’s just playing golf that was a contributing cause in their demise.

  14. Those miniature poodles can be vicious.

    Never been bitten by a dog. I think the 411 in Doggy World is I don’t taste all that good.

  15. DAV: That would be tough since cats are not allowed in my house! I think the microwave would be too small for a cat, also. I suppose I could toss in the salamander or gecko…..

    (I have been bitten by a dog–my own vicious Yorkie. Don’t know about poodles, but Yorkies are nasty biters. Don’t know if I tasted good or not, but he never did twice. And no, I didn’t toss him the microwave as punishment. 🙂 )

  16. I suppose I could toss in the salamander or gecko

    Maybe but the Geico gecko looks like it would just push the water around without absorbing any.

    Been bitten by a cat. Two of them in fact. The second was raised by the first so it might have just been a family thing. Still prefer cats though. Low maintenance. Dogs need more.

  17. He’s a desert gecko

    Yum! Gecko dessert: a la mode with whipped cream and a cherry on top!
    Oh, wait.

  18. DAV, “Can you say for certain that entering/exiting/moving within the field doesn’t cause damaging electron flow?” This is called heating and so we have come full circle. The external electric field required to sustain the polarization field you are referring to inside a human body, which is conducting, would be immense and orders of magnitude out of the range of the domestic fields that we encounter. I’m not sure what your question about current is?

    “I don’t taste all that good.” If you have never been bitten how would they know? By smell maybe? Don’t show fear. Cats bite just to keep their hand in; it’s part of the job description. Dogs are good for exercise.

  19. Scotian,

    Strange that oncologists aren’t beating a path to your front door. Ain’t life ironic?

    I’m not sure what your question about current is?

    Maybe because I don’t have one.

    Don’t show fear.

    Maybe. Years ago, when I was selling home improvement, I had an appointment with a guy who let his viscous looking pit bull into the yard. I guess he had second thoughts about the appointment. He was surprised when I knocked on the door and asked if the dog I had by the collar was his.

    If you have never been bitten how would they know?

    I think it’s on the cat’s Facebook page.

  20. DAV, “Strange that oncologists aren’t beating a path to your front door.” That is because they already know what I just explained to you. That and the fact that they treat the disease and are not so concerned with the causes. Maybe you are thinking of epidemiologists and we all know what short shift they get on this website.

  21. “Strange that oncologists aren’t beating a path to your front door.”

    Sweetie, that was in reference to

    The external electric field required to sustain the polarization field you are referring to inside a human body, which is conducting, would be immense and orders of magnitude out of the range of the domestic fields

    and that you know the order of magnitude required at the subcellular level to cause cancer when the knowledge of the causes of cancer is small to nonextant.

  22. DAV, What I know is that the effects that you are talking about are background noise. The amount of heating due to any but the most extreme environments (i.e. a lightning strike or sitting inside a microwave oven) are less than that experienced from walking from a cool to a slightly heated room in your house. But no one worries about the latter because it is silly. What they don’t see is that it is all silly until you get to ionizing radiation and we haven’t even mentioned hormesis.

    Also a patronizing tone does not become you. You are no where near smart enough for that.

  23. <i?What I know is that the effects that you are talking about are background noise.

    Again you claim to know the levels required. Yeah, sure.

    Also a patronizing tone does not become you. You are no where near smart enough for that.

    Calling me stupid isn’t patronizing? Hmmmm…. Then I guess it wouldn’t be patronizing should I pointed out your own stupidity, yes?

    In any case, please do Babylon, No. 5. I hope you don’t mind the levels of Chanel I deem necessary to apply, No. 5. More than a scotian amount for sure. Say, are you related to 7 0f 9 by any chance?

    If you’re just looking for a consoling pat on the head, here it is:
    There, there, now!

    Just wait until I get around to patronizing.

  24. DAV,
    It is amazing how the fates work, but just today Cecil Adams posted an article on a closely related topic.

    http://www.straightdope.com/columns/read/3151/anything-new-on-the-dangers-of-microwave-ovens

    You might wish to read it.

    “Again you claim to know the levels required.” Of course and so do you. For example do you feel warming when you walk past an electrical transmission tower? No, then there you go.

    The rest of your comment seems to be gibberish.

  25. All,

    What we see in this discussion is that coming to agreement on evidence and uncertainty is nearly impossible. Just as I set out to prove.

  26. “Again you claim to know the levels required.” Of course and so do you.

    No. I say I don’t. Only you seem to know otherwise how could you possibly know if any level is insignificant enough to be called “background”?

    For example do you feel warming when you walk past an electrical transmission tower? No, then there you go.

    Seems you’ve never had any ESD training. The amount of energy needed to damage a small area can itself be small enough to be generally unnoticed by the surroundings. I’ve seen circuit boards fried by someone picking up a Styrofoam coffee cup. It’s one of the reasons coffee cups are banned from ESD safe areas.

    The rest of your comment seems to be gibberish.

    Am I to be surprised by this?

  27. DAV,
    “Seems you’ve never had any ESD training.” You like to make assumptions. I actually teach electronics. The frying as you call it is the punching of a hole through a very thin oxide layer in a MOS-FET by an electrostatic charge. I fail to see what you are claiming the relationship is between this and low level and low frequency e/m radiation?

    Briggs, I think that you may be seeing something else.

  28. I actually teach electronics. The frying as you call it is the punching of a hole through a very thin oxide layer in a MOS-FET by an electrostatic charge. I fail to see what you are claiming the relationship is between this and low level and low frequency e/m radiation?

    OK. You didn’t have to prove you never got an ESD course. I meant the board was fried — between the traces — and not the active components.

    The point is a small area can be affected by a small amount of energy. Just what “The amount of energy needed to damage a small area can itself be small enough to be generally unnoticed by the surroundings.” means. Are you really claiming you can’t read?

    The subcellular level is pretty small.

  29. DAV,
    I am claiming that you have a great deal of trouble making yourself clear. Maybe the snark gets in the way. If I understand you the energy required to fry the board came from the power supply and not the electrostatic charge, possibly due to a short. There is no question that a small amount of energy can effect a small area but in the context of the original discussion this has to be ionizing radiation if bonds are to be broken but only non-ionizing radiation for an increase of temperature or internal energy in general. Speaking of reading, did you read the Straight Dope article.

  30. I am claiming that you have a great deal of trouble making yourself clear.

    You’re right. You’re either every bit as incompetent as you sound or you’re just trying to be annoying. I have better things to do.

  31. DAV, You must be continually frustrated at all the incompetent and annoying people that you meet every day. You know, the ones that don’t recognize your genius.

  32. “Why must we study all this gobbledegook about philosophy when it’s more important to learn all the neat methods of handling data?” (Dr. Briggs)

    Because you should never peek at the data until you have in mind a good model of what you are looking for. Then, ceteris paribus, “all else being equal” you should understand the differece between “probable” and “likely” and what it means to be genuinely probably approximately correct and if you are willing to allow none of foo, one of foo, or any number of foo, etc., etc., etc.

  33. Italian oncologist, Tulio Simoncini, identifies many cancers as “candida albicans”; he has had a notable cure rate using – wait for it – Bicarb. Soda solution (internally) and Tincture of Iodine (topically, e.g. on skin cancers).

    Believe it, or not! But his website is interesting.

  34. You must be continually frustrated at all the incompetent and annoying people that you meet every day.

    Heh. Only one other. I used to run into a TCO sycophant with, now that I think about it, similar style, phrase choice and sentence structure. Maybe living in the same proximity if your speech pattern is any indication although that’s becoming more unreliable as time goes on.

    One quirk you both have in common — which really stands out — is posting things like I’m not sure what your question about current is? . You’ve done similar more than once on various occasions so it likely isn’t a typo. Distinctly odd. Why is it framed as a question? You don’t know when you’re sure?

    You also share making a lot of one or two line posts, the lack of comprehension of the posts of others, and the tendency to project or take the offensive by accusing others of doing what you’ve just done or intend to do.

    This other had an affinity for using names starting with the letter ‘S’ and sometimes prepended “Dr.” for who knows why.

    Small world. Should have seen it earlier. Must be getting old.

  35. Allow me to introduce, “best projection reiterative truncated projected least squares (BP-RTPLS)” a technique for dealing with “omitted variables”. I only read the abstract because I was able to omit the rest of the paper using an analogous technique. This deals so successfully with Briggs’s problem of facts not included as evidence that I predict that fourth generation omitted variables compensation will allow statisticians to draw amazingly accurate conclusions without collecting any data at all!

  36. DAV,
    More assumptions and psycho-babble. I guess it is easier than debating the scientific basis of your beliefs about thermal radiation. I have to ask, what does TCO stand for?

    “the lack of comprehension of the posts of others”. Mostly you and that is because you have a tendency to speak in code laced with heavy sarcasm.

    “This other had an affinity for using names starting with the letter ‘S’ and sometimes prepended “Dr.” for who knows why.” Now this is just bizarre. Do you even know what it means?

    The funny thing is: my position is not even controversial. It is the standard scientific and medical position.

  37. Tulio Simoncini should be guest on Dr. Oz, probably on the same day as he has the psychic that speaks with the dead (which Dr. Oz has done). He’d fit right in. I would note that if it’s true that cancer and yeast are associated, women in America should have a cancer far above that of men and the rate should be very, very high based on the sales of products to cure yeast infections.

  38. debating the scientific basis of your beliefs about thermal radiation

    ???
    Didn’t think it necessary. Straying off-topic and straw-man argument seems more your thing; not mine.

    what does TCO stand for?

    Can’t say really. Mom? Apple pie? Freedom? Civil rights? Academic integrity? Military superiority? Not sure why you it should be relevant or why I would know. I would think you are in a better position to answer the question than I am. I’m not the person to ask in any case.

    This is your last feeding. Go back under your bridge and wait for someone else to come along.

  39. DAV,
    It is times like this that I wish that I had Willis Eschenbach’s way with words. But since it is becoming increasingly evident that you are simply incapable of civil discourse and possibly of following an extended line of argument, it is probably best that you are left to mumble to yourself.

  40. I have no expertise on the biophysics of EM radiation, but here are my uninformed first thoughts on the matter:

    A single photon of non-ionizing radiation can do no chemical damage.

    Microwaves correspond to vibrational and rotational frequencies of small molecules and so can pump energy into those modes which can either end up breaking the molecule or more likely getting converted to other modes of thermal motion – which may still end up causing chemical breakdown due to collisions between molecules. But the penetration is poor (as anyone who has tried to defrost a roast in the microwave will know) and in any case we have nerves capable of detecting that effect so it would have to be felt in order to be significant.

    Ultra low frequency radiation penetrates seawater sufficiently to be used for communication with submarines, and if it is not absorbed it can’t be doing much heating.

    BUT if it’s not absorbed then it can penetrate, and the body includes processes with a wide range of natural frequencies (it is amusing to note that although the Herz is actually named after Heinrich it also corresponds to the frequency of slow normal human heartbeat).
    So I don’t see why it is not conceivable that some important large molecule might have a resonant frequency around 60 Hz which could get pumped to unsafe levels by the radiation from electric power lines.

    And that does seem to leave us with a question for the statisticians.

  41. Alan Cooper,

    Maybe not a molecule All that would be needed is a resistor and capacitor combination with the proper time constant to form an RC oscillator. Presumably, the capacitors will be small so large resistances would be needed. The effect could be small enough that it isn’t felt but still be substantial enough to cause localized damage.

    The presence of a large electric field could lead to small current flows which could in turn might cause localized effects particularly over a period measured in years.

    Seems to me the starting point should be lab experiments but — hey! — who wants to wait years to publish?

  42. Alan Cooper,
    The wavelength of a 60 Hz e/m wave is 5,000 km which is considerably beyond the molecular scale. Also molecular bonds will not break simply due to the adsorption of microwave radiation or molecular collisions at room temperature. Chemical reactions will occur at high (cooking) temperatures or the food might burn (oxidize) at even higher temperatures. Molecular dissociation will require even higher temperatures. Since the microwave oven wavelengths are in the mm or cm range this does not correspond to the vibrational or rotational molecular modes but is a process called dielectric heating. The molecular modes are in the infrared range of the e/m spectrum, in the micrometer range. You see that the name microwave oven is somewhat of a misnomer. This is actually funny when you think about it since no one worries about the radiation from an electric space heater which is primarily in the infrared. But as I’ve said before unless you cook yourself, and you can get burned with a space heater, the danger lies in the ionizing radiation range, which starts in the ultraviolet. It is here that bonds can be broken.

    Thank you for your candid interest in the science. It is very refreshing.

  43. With regard to what the statisticians “should” be looking at, I agree with DAV and Ken that if one wants to answer the question about EM field exposure properly then the only appropriate tool is a controlled experiment with animal subjects and intense fields, but I also think that a very loose observational study might be useful for deciding whether or not there is any urgency (as opposed to “just” scientific interest) in finding out the answer. For that purpose I think it would be of interest if one were to find a data set showing any significant indication of a substantial connection between total cancer rate (of all cancers) and any other variable which one might reasonably expect to be highly correlated with lifetime exposure to the relevant kind of radiation (and which one had no other reason to associate with cancer – as one might, for example, if the proxy was also tied to PCBs or other kinds of possibly carcinogenic chemical exposure).

    Of course, if one is just interested in the overall risk (or not) of “living near” high voltage power lines (regardless of the mechanism of that risk), then it’s not really a question of “cause” if the risk can be mediated (eg by EM shielding, or checking for and removing residual pollutants from the power line construction process, or maybe even just by having suitable counselling sessions if the risk turns out to be psychosomatic).

  44. Alan

    The risk is that studies conducted out of the lab will follow the course of studies such as the dust exposure/ highway ones where the assumption (and expected answer) is the exposure does cause cancer. Unfortunately, since it is the expected answer someone will keep redoing the study until they get it “right”.

    I personally doubt EM radiation is a factor but those living under a power line are in a more intense electric field than if they were living elsewhere. PCB’s maybe if they live near a power station. Thing is, if you go look at the PCB studies there are a number of problems with those as well. In fact, there are problems with nearly all (if not all) cancer studies.

  45. Someone seems to have overlooked that the high power exposure studies were used as an example (Let’s take an example (kindly provided by reader Rob Ryan). We want the truth of the proposition, “Living near high voltage towers causes cancer.”) to illustrate some problems associated with these kinds of studies and not about any particular proposed studies. I doubt it was meant to be inclusive of all problems associated with the example.

    Seems there are those here who would be quick to propose a mechanism and then, because they deem the mechanism an unlikely cause, consider these studies completely meritless; dismissing the example study and similar out-of-hand. That’s the proper “scientific” approach in their view. Elsewhere, this might be called the straw man method.

    Also, we’ve been hearing a lot of panting coming from the northside of this blog. Sounds like someone having trouble keeping up or maybe close to an orgasm. I suppose there could be other reasons. Whatever. Whoever it is seems to have a bad case the clap if the sound of the condescending applause is any indication.

  46. Oh well, I guess the quack watch reference struck too close to home. It wasn’t intentional. That just happened to be the website that the article was posted on. The article is very good though and definitely worth reading. It will answer all your questions. By the way your RC model is just another way of saying dielectric properties of living tissue and doesn’t add anything to the discussion. It is too bad about your temper problem as I think that we could have had an interesting discussion on this topic. Maybe we will try again when you are in a better mood.

Leave a Reply

Your email address will not be published. Required fields are marked *