Wednesday, September 24, 2014

The Christian and the Statistician Walked into a Bar-Graph

Facebook, love it or hate it, gives me a cross section of my life. I joined it only to see photos of my grandsons. But quickly I met old friends from virtually each epoch of my life. There are people from high school, several from my evangelical days and then friends I'm met more recently.


I don't post that often. However, I have been bombarded by postings from old Christian friends. Some of them post about five times per day. I've had to "un-friend" many. The thing I notice the most evangelical posts (and I suspect that this is the backbone of the pop-side of the Internet) is the dissemination of "information" with quotes intended. This "information" is usually concerning political issues, such as "the lie of global warming." Recently it has been about the evil people in Gaza or the Soviet invasion of the Ukraine is proof that Jesus is coming soon (eyes roll here).  Regarding the situation in Gaza, if I've heard it once, I have heard it . . . well a hundred times at least, "Arabs believe only in death, Christians and Jews in life."  This is propaganda at its best.  But I digress.

Of course of late everything is about Isis. Their extreme violence makes them an easy "evil empire."  No doubt that they are evil and there is no excuse for that extreme evil, but the problems is, in the Christian domain at least, the discussion becomes very simplistic, dualistic and somehow points to the end times. We, the good guys are opposed to the bad guys.  No discussion of the political-social situation that breeds such groups as Isis.  To even bring up US foreign policy, as if were not perfect, would be an outrage. It did bother me that the greatest outrage of Isis' evil from my Christian friends came when they started killing Christians as if the tens of thousands of non-Christians of Isis' heinous crimes don't matter.

The topic that I'm the most expert in is medicine. Constantly cures are promoted on FB along with warnings against vaccinations, pasteurized milk, poisons in our food or the evils of prescription pills.  It seems that this "non-western medicine" has not only gripped the Post Modern society, but the Evangelicals as well and maybe more so. I remember when I was on staff with an Evangelical group that our insurance rates went sky high (like everyone's but more so) because the employees of the organization were extremely high-uses or non-western medicine.  If 20% of your company is going to the chiropractor twice a week it can be very expensive for insurance companies.

The evangelicals of course don't have the insight that they are just one small part of the present post-post modern culture. I work in what the lay people call "Western Medicine."  This is not an accurate title because most of the medical practitioners in China, India or any country practice the same type of medicine that I do. So the much better descriptive term is "Evidence-based Medicine."  This is what separates us out from the other side, which doesn't do studies or look for evidence but follow particular philosophical views about the body and health. It is a Gnostic knowledge, meaning in this case a spontaneous knowledge without supporting research.

I am not here to argue all the merits of Evidence-based Medicine because, like all things in this fallen world, it too is messed up.  I wish I had the time to tell the stories of how, what is best for the patient, is at the lowest end of the priorities in our present system. The top end is the war between insurance companies and providers and the constant fear of being sued. But I digress again but there are legitimate ethical questions that need to be asked about our system as well.

In Evidence-based medicine, as the name implies, the real dividing point is simple statistics and acknowledgement of the placebo (as well as the nocebo, imaginary negative effects) of treatments. In our form of treatment, we can not (or at least should not) promote a treatment unless it has been proven, statistically, to be superior to a placebo.  Our placebo is our gold standard because the placebo negates the physiological influence of the imaginary.

The backbone of evidence-based medicine is the well developed statistical method.  This is where this argument becomes philosophical as all mathematical arguments eventually do when you take them to their ultimate point. Statistics have worked out carefully the laws of probability and real effect. Math is the language of nature, meaning of all that is there in the physical world. I also am convinced that math explores the fabric upon which God as perched reality. I do believe that math is a "god-thing" to borrow an overused term from the evangelical. I've debated in my mind if I can say that math is infallible . . . and I think it is. It is my humble opinion that mathematics is spared from the Fall of Adam . . . but NOT the mathematician. So even in evidence-based medicine, the "mathematician" can get it wrong, either by poor calculations or by intent . . . in order to make something look better than it is.  As I understand, Mark Twain said that there are "Lies, Damn Likes and Statistics."  While I do believe, unlike Aristotle, that human reason is fallen and can't be trusted, the reason I believe that math is perfect, is because it is the echo of the creator Himself.

So, I do say that I believe that evidence-based medicine is more godly than that which is not because of the science behind it. It is based on statistical proof of effect and not just emotional placebo. I think you should expand this topic to the general way that Evangelicals approach truth. They don't consider the math  and they don't respect to how we can so easily deceive ourselves psychologically.

The key support of non-evidence based medicine is the testimony. From a mathematical side, and understanding human psychology (and economics) this is a very weak body of evidence.  back to the case of medicine, the testing where 1,000 patients are treated for y disease with x treatment and another 1,000 are treated for y disease with a placebo and neither the patients nor the investigators know who got the real, x, (double blind) can statistics be fulling trusted  . . .  but at that point it can be fully trusted due to the perfection of pure math.

Sorry, once again I only had time to type once and not proof-read. I dream of a day when I can think and type and proof-read what I type.




4 comments:

abmo said...

I am sorry to ask here, but what is your e-mail address? I would like to ask you a question? Since Michael died (imonk), your blog has given me sanity in our crazy beautiful world :-)
Mine is nestus.venter@gmail.com

Thank You
Nestus Venter

NOTAL said...

I think math, like all deductive logic, might be "infallible" in the sense of being internally consistent. However, all deductive methods are absolutely impotent to describe the real world without quality axioms and assumptions. And axioms and assumptions must come from outside of deductive logic or math. So by necessity the conclusions of math are just as fallible as the inductive, abductive, or arbitrary selection of axioms.

This was well illustrated by Euclid, who based his system of mathematics upon five axioms that he thought were very safe and self-evident, one of which turned out to be wrong. Euclid's fifth postulate stated that only parallel lines never touch, but it turns out that we live in a curved universe where that's not the case.

The statistical methods used in modern science are at least as problematic. In addition to being based upon supposedly "self-evident" axioms, they are also built upon unverifiable assumptions and blatant value judgement. Take for example the notion of significant findings having p<0.05. This is a value judgement. There is no reason for this except that people have arbitrarily decided that 1 in 20 false positives is good threshold. Why not 1 in 10 or 1 in 100? The math alone cannot decide that something is significant, but only when the math is built upon a specific system of values and assumptions.

I think the failure to understand the fallibility of this necessary basis of math and science was part of the arrogance of modernism that led to tragedies like the eugenics movement and Soviet style Communism. The postmodern rejection of that arrogance is healthy. I think postmodernism goes too far when it switches from epistemological humility to epistemological relativism, but epistemological humility is a good place to be.

Maybe that natural rain forest sloth poop actually helps prevent cancer, stranger things have happened (like the most toxic poison known to man being able to cure headaches). I don't know, but I'd be skeptical until I see a study demonstrating it with a p<0.05 that it works.

...well maybe p<0.1

j. Michael Jones said...

abmo, I sent you a note.

NOTAL, you make several good points. Questions for you, are parallel lines a concept or a mechanical construct? If the former, then parallel lines can exist because they don't lay within the 3-d universem cured or straight.

Regarding modernism vs post-modern, I think you are right. The problem with humans is that we always migrate to the extremes and we can't find balance. So either reason is infallible and will always lead to utopia (so we think) or it is worthless and all truth is relative and personal (so we think in the other extreme). There should be healthy skepticism. Science and reason could not solve the moral problems of people (where as people like B F Skinner thought it could).

Yes, of course sloth poop might have the cure for cancer. I have no problem with that. But, me saying it cures cancer doesn't make it cure cancer. Nor does me saying it cured my cancer make it true. The latter statement could raise eyebrows at best. If then another person says it cured their cancer, then the eyebrows should go higher . . . then comes the double blind and statistics to clear out the placebo effect.

I agree that p<0.05 is arbitrary, but reflects our comfortableness with "certainty" and practical efforts to reach that "certainty."

NOTAL said...

Hey pops,
The parallel lines postulate is a concept, but it was believed by Euclid and pretty much everyone for a long time after him that it was self evident that this concept corresponded to the real world. Euclidean geometry is still self consistent, and is still useful as a close approximation of the real world, but it's describing a system that is fundamentally different than the world we live in. If you tried to make a GPS system work using Euclidean geometry, you would have major problems even though the math proves that it should work.

I think my point was just that a mathematical proof is only as good as the inputs you provide. If you input flawed assumptions, you end up mathematically proving flawed results. Likewise if your values (e.g. your comfort with certainty) differ from mine, we can use math to prove different results. Responsible use of statistics makes it clear what assumptions are being made, so that a reader can decide whether or not they agree with those assumptions, but it can be challenging as a researcher to actually understand what assumptions you are actually making, and how legitimate those assumptions actually are.