[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index][Subject Index][Author Index]
crossed purposes and weariness
I'm beginning to feel like B'rer Rabbit enmeshed with Tar Baby, so this
will be my last communication for a while on these particular points.
If brooding pythons are by definition endotherms, we have certainly been
talking across purposes. My definition puts them well outside of
endothermy, brooding or not.
I have already tried to point out how body size (above a certain point) has
precious little to do with walking speeds. Perhaps if I asked "Why doesn't
a Norway rat walk at <1 kph while a green iguana does?" or "Why doesn't a
6-foot dinosaur walk at <1 kph while a 10-foot salt water crocodile does?"
I might get a non-evasive response. But I am weary of trying.
Boy, wasn't I off base in suggesting that this correlation/causation thing
has assumed the status of a religious doctrine? Just look at the
open-minded responses I got. Here as I understand it are the
well-thought-out, mutually contradictory responses:
1) Every correlation is <100%, therefore there is some probability that it
could have occurred by chance. Thus correlation does not imply causation.
2) Any two variables that both tend to increase are 100% correlated,
regardless of any causal link. Thus correlation does not imply causation.
Since any measured correlation is <100%, there is always the probability
that it could have occurred by chance, whether it is in a controlled
experiment or not. This has no bearing on the correlation/causation
question. #2 is not even worth responding to. I do not intend to go more
deeply into it. It is somewhat off topic, and I only brought it up because
it seems to be frequently thrown out as a denigration of descriptive as
opposed to experimental science, and as a criticism of dinosaurology in
general. I will not suggest that you tattoo anything to anything. I
invite you to think for yourself. Find out for yourself what correlation
means, what causation means, what a controlled experiment does, and what
statistical significance means. Ask yourself how it could possibly be that
two variables simply correlate out of thin air, and pursue the answer to
your own satisfaction. John Bois' example illustrates my point. A
correlation between corporal punishment and out-of-control adults doesn't
indicate that one causes another. He argues that instead both are
reflecting an UNDERLYING cause, "out-of-control" kids, which leads to both
corporal punishment and out-of-control adults. This kind of example is
exactly what statisticians use when they speak of "spurious" correlations.
Why does he feel the need to posit any connection at all between the two?
Surely it could just be the covariance fairy at work.
Finally, a truly objective jury is always out on any subject. My ideas are
nothing more nor less than working hypotheses. My world will not crumble
if Bachiosaurus brancai turns out to be an ectotherm. Statements I have
made about "painfully obvious" conclusions notwithstanding, I am never
closed to questioning my own ideas. I have no need for certainty, and if I
did I would not look to science to provide it. I invite everyone to bring
open-minded skepticism to every idea, authoritative or heretical.