Prismatic wrote:Groktruth wrote:
Your use of the word improbability when you clearly meant probability indicates you are in unfamiliar territory, but it would be fun to hear your explanation of how the "improbability" was calculated for this event and what it means.
More later, perhaps.
Well, I got a PhD in biomathematics in 1968, studying probability and statistics.
Your trump card? I also have a Ph.D. in mathematics. In all my years in universities I cannot recall anyone speaking of an improbability of less than 1 in 10,000. It's just not the kind of expression someone familiar with probability would use. Perhaps you need a refresher course in those subjects?
The "correlation does not prove causation" argument only applies to correlations that are first found serendipitously, and then used to justify the plausibility of a cause and effect relationship. In this case, the correlation was predicted, actually prophesied, ahead of time.
No, not at all true. You really do need a considerable review here. Prediction has nothing to do with it.
Two series of events may have a very high correlation and yet neither causes the other because they are both the result of something else. You might well
predict that as the sale of swimming suits increases the number of cases of sunburn increases and conclude that selling swimming suits causes sunburn. Of course the truth is both occur more frequently in the summer months.
The fact remains that while correlation is a necessary condition for causation, it is, in and of itself, not sufficient.[/quote]
Interesting response. First, good on you for persevering to a doctorate in mathematics. May I draw on that authority with questions that presuppose such training?
Second, I am troubled by your focus. Here I have presented a correlation, quite strong, with, as the university or academic statisticians like, or used to like, to say, a significance level of less than .0001. This correlation deals with billion dollar disasters, often causing deaths, and always causing much hardship. The correlation offers a possible, easily testable, hope that such disasters can be reduced or eliminated.
So, why are you interested in my use in a public, popular forum, of the term "improbable" which is frequently used by the public, and which makes the point clearer? Granted, among professional mathematicians and statisticians, improbabilities are a subset of probabilities, being those regarded as being low. And, I grant that they may be reluctant to include such a subjective term. But, my opinion, which may of course be wrong, is that using this word helps non-statisticians get the point.
The correlation I note would appear by chance less often than once in 10.000 presidential terms. Something is going on here, and we have a substantial clue to direct our pursuit of finding out what it is. I hope your drawing attention to my lapse from professional rigor was not a subtle attempt to make an ad hominem argument.
Now, you do not appear to understand Bayesian modelling of the scientific method, in your failure to distinguish serendipitous and predicted unlikely correlations. Do you get it, that Bayes theory is a mathematical statement that refers to conditional probabilities? That it allows one, in the case of the scientific method, to calculate the posterior plausibility of a given hypothesis, conditional on the finding of a confirmed predicted bit of evidence, including a correlation? In the calculation, the less likely, more improbable, the evidence, the more it affects the conditional probability of the relevant hypothesis. A serendipitous correlation, once it has been found, is by induction rather likely to be found again. So, the probability of a reoccurrence is not well estimated by the statistical probability, and such evidence bears little weight evaluating the plausibility of an hypothesis. However, a new, predicted, and unlikely correlation, if found, greatly improves the plausibility of the hypothesis that predicted it.
In this case, we consider, before we do the calculations, the prior probability that we will find a correlation between serious disasters, and US policy regarding Israel. Most would assign a low value to this, accepting the probability by chance alone to be the best estimate that this correlation will be found. When this prior value (in this case, .0001) is inserted into the Bayesian formula, the posterior plausibility of the predicting hypothesis is substantially increased.
Here is how your swim-suit example misses the point. First, you had no hypothesis for which you hope to evaluate the plausibility. Second, the predicted correlation would be considered highly probable, prior to your making the prediction. So, even if you did have a hypothesis, finding the correlation would not make it more plausible.
However, if you were hypothesizing that sunburn was caused by exposure to the sun, and that swimsuits purchases reflected a personal inclination to indulge in such exposure, so that locales with more swim-suit purchases would have a greater number of sunburn cases, you would actually be doing science. Testing this prediction over months would not however, be regarded as a good test, the prediction being likely to be found with or without the hypothesis. If you predicted over counties in Pennsylvania, however, the prediction might be regarded as slightly improbable, and if found would raise the plausibility of a cause effect relationship between such purchases and sunburns. Rather likely anyway, however. I think, however, that in your mind, you had the hypothesis that sun shining on skin causes sunburn, which would be regarded as highly plausible. Perhaps you were thinking too that deciding to purchase a swimsuit, that would eventually result in more sun on skin, ought not be thought of as a cause of sunburn. But, medical counselors trying to understand why one county has more sunburn than another might be wiser looking at swimsuit purchases than shade-tree density, say. Might decide to put a warning on the swimsuits, or advice for sunblock.
The predicting hypothesis here is biblical theology, where it is postulated that there is an intelligent being living in some other dimension of our universe than the one we inhabit, who has and uses the power to influence weather and other possible disaster causing events in our dimension. This being asserts plainly that they are available to be "proved" or tested, within the boundaries of decency and respect that we normally apply to what is effectively psychological research into their behavior and thinking. Some who claim to know them, referring to a scriptural promise this being has made, predict that the correlation we looked for will be found. When that exceedingly unlikely prediction is confirmed, it increases the plausibility that the hypothesis that predicted it is true, by Bayesian calculations.
So, do the math. As always, doing the math overcomes subjective biases. We all know that you are strongly subjectively invested in biblical theology being mythology. So, you will have to do the math, to get over that. For humanitarian reasons, you need to do the math here. We might well be close to understanding disasters, and knowing how to prevent them. Obama, for example, apparently unaware of this research, casually tossed out a "return to pre-1967 borders" suggestion, which was immediately followed by the Joplin tornadoes that killed over a hundred people.
i do hope, in your interest in mythology, you have done serious research into the processes of denial, hallucination, suggestability, hypnotism, and other relevant psychological factors that influence the human tendency to produce myths. When I do so, it becomes clear that the idea of evolution is plainly myth-making, and the evolutionists' resistance to intelligent design a text-book example of myth-holders resistance to data and ideas that call their myth into question. Like the Catholic Church and Galileo.
If God is all that, almighty, the best, then how lonely is his duty. It's like in the movie "Cast Away", the guy became the master of the Iceland, God, creating fire and his own rules and goals,not answering to anybody, ruling everything as he pleased, but what a miserable years had he spent there.
- By Surabhi Rani
- By Gertie
- By Fried Egg