|
|
|
|
Hawks
SFN Regular
Canada
1383 Posts |
 Posted - 05/31/2012 : 11:57:38
|
The Discovery Institute's "Evolution news and views" has a new contributor (Stephen A. Batzer) and judging by his most recent post, he will be a clueless one. He starts out by stating the following:
| If you've followed the ID vs. Darwinism debate at all, you've probably come across the term "Bayesian analysis." This technique is the skeptic's friend and it can actually be very simple if informally used. |
(Emphasis added. By skeptic, Stephen means people skeptical about evolution)
Yes, Bayesian analysis is a powerful tool one can use to compare competing hypotheses. But the key word I highlighted in the quote above is "informally". Because, in order for Bayesinism to be an ID supporters friend, one has to discard that pesky little Bayes theorem that Bayesian analysis uses. I.e., instead of basing one's conclusion on a rigorous mathematical framework, one is supposed to rely on, it seems, one's own intuition.
In essence, what one has to be able to do under bayesian analysis is to update one's belief regarding a hypothesis as new observations are made. I.e. new observations can strengthen or weaken the likelihood of a hypothesis. In Bayesianism, the terms prior and posterior probabilities are used to described what one believed before and after the new observation(s) were made. (this much I think Stephen would agree with).
What Stephen fails to realize is that ID gives us no reason to measure the likelihood of any specific ID hypothesis rather than any other one. ID, famously, says NOTHING about the designer. It says NOTHING about what the designer would want to do or what it could do. Therefore, making any sort of hypothesis about what observations one should expect under ID is impossible. What follows from this is that one can't use Bayesian analysis on ID. This is a problem that Stephen has swept under the rug. Good work.
|
METHINKS IT IS LIKE A WEASEL
It's a small, off-duty czechoslovakian traffic warden! |
|
|
HalfMooner
Dingaling
Philippines
15831 Posts |
Posted - 05/31/2012 : 14:44:55 [Permalink]
|
Oh, the maths! Me, they hurt.
But I can at least understand why the Discoveroids prefer an "informal" type of Byanesian analysis to the kind that actually requires real evidential data with real numbers. Only by skipping the quantifiable parts and replacing those with their imaginations can they let their Designer have the room to freeball as he is accustomed. And they also themselves have freedom from being forced to know mathematics any better than, say, this old math cretin.
Of course, when you strip a mathematical analysis of its numbers, and substitute making-shit-up, there's that little problem about not doing science. But to the Discoveroids, that's an actual advantage, since doing that godless science stuff is the very last thing they want to try to explain to their pastors.
But they still get to use fancy terms like "Baynesian" and "analysis" to baffle and confound their ignorant followers with their erudition, thus making themselves sound more edumacated and sciency than the likes of Kent Hovind and Ken Ham. |
“Biology is just physics that has begun to smell bad.” —HalfMooner
Here's a link to Moonscape News, and one to its Archive. |
| Edited by - HalfMooner on 05/31/2012 14:57:53 |
 |
|
|
Machi4velli
SFN Regular
USA
854 Posts |
Posted - 06/06/2012 : 21:31:03 [Permalink]
|
Bayes' theorem is an often bastardized thing. For one, I don't even agree with the interpretation that your degree of belief in something is the same as the probability of it occurring (subjectivism in the statistical sense). The latter sometimes cannot even be quantified, which brings one to inconsistent conclusions at times. (This is the same why I think Dawkin's scale of belief in god is silly, for example, as we simply cannot quantify these probabilities with any basis unless we adopt subjectivist statistical interpretations.)
Although it's commonly written P(A|B) = P(B|A)P(A)/P(B), sometimes P(B) = 0, which renders the statement meaningless. From a different angle, I prefer to deduce P(B)P(A|B) = P(B|A)P(A) from the probability axioms to avoid the problem, but then if P(B) or P(A) = 0, it becomes tautological, and essentially meaningless. This makes it so that you can't really make useful empirical claims it if either probability is zero (e.g. if something's simply false, you can't gain evidence for it).
Sometimes you'll hear someone argue that the probability of god's existence is fixed. Now if I give you some evidence that "supports" this claim, my argument must increase the probability, and therefore, you should believe it's more likely. This gives them license to claim any "evidence" however silly shows something is more likely. (This in itself is a mistake even if you take the above interpretation, because it would be equal or more probable because the evidence may be irrelevant.) |
"Truth does not change because it is, or is not, believed by a majority of the people."
-Giordano Bruno
"The greatest enemy of knowledge is not ignorance, but the illusion of knowledge."
-Stephen Hawking
"Seeking what is true is not seeking what is desirable"
-Albert Camus |
|
|
|
Hawks
SFN Regular
Canada
1383 Posts |
Posted - 06/07/2012 : 09:55:11 [Permalink]
|
Yeah, bayesianism isn't as straightforward as Batzer would have it.
| For one, I don't even agree with the interpretation that your degree of belief in something is the same as the probability of it occurring (subjectivism in the statistical sense). |
What to remember is that bayesian testing is contrastive. The likelihood of one hypothesis (Pr(O|H)) tells you very little about it's probability (Pr(H|O)). However, comparing the likelihood of two or more hypotheses indicates which one is more probable (assuming one has formulated one's hypotheses correctly to start off with). |
METHINKS IT IS LIKE A WEASEL
It's a small, off-duty czechoslovakian traffic warden! |
|
|
|
 |
|
|
|
Bayesianism, the Evolution Skeptic's Friend...?
