Harry Jaffa RIP; Statistical Inference

View from Chaos Manor, Thursday, January 22, 2015

I would remind you that extremism in the defense of liberty is no vice! And let me remind you also that moderation in the pursuit of justice is no virtue!

Barry Goldwater, 1964, in a speech written by Harry Jaffa


Harry Jaffa, RIP. We had not corresponded regularly since I left academia to become a full time writer, but we were close enough before that. He was a guest lecturer to my political philosophy classes, as I was to his at Claremont, and we attended several conferences together. At that time I was mostly associated with Russell Kirk and Stefan Possony, and getting more into actual politics; Harry was the intellectual inspiration of the Goldwater movement (not that the Senator was not his own man, and although he was not primarily an intellectual he certainly understood the issues.) In 1969 I was co-director of Barry Goldwater Jr.’s successful Congressional campaign and could have gone to Washington as a Congressional staffer, but I did not like the political game; Harry was one of those who advised me not to get into the political game.

Harry’s inclusion of the Cicero quote in Goldwater’s speech came as a surprise to everyone. In those days Johnson ran on a platform of being moderate, and made Goldwater look like a raving maniac eager to nuke everyone. Johnson was successful, but he also used divisive and deceitful advertising, such as the little girl and the countdown to atomic explosions. When Goldwater suggested that we ought to bomb the Ho Chi Minh trail in Laos, Johnson said that was the most trigger-happy thing he had ever heard. As Johnson said that I was looking at strike photos of US strikes in Laos, but of course neither I nor Harry nor Senator Goldwater could say so. I had clearance and I suppose Harry did also (mine was from being an intelligence analyst.) I used to wonder who we were keeping it a secret from. The North Vietnamese surely knew they were being bombed, and presumably told their Russian allies… the only people it was secret from were the American people.

Over the years we had less and less contact, which is a shame. It was always a delight to discuss the Federalist Papers, which he knew intimately as I was learning about those vital documents. I thought I knew them when I got my doctorate, but not so.

We did not agree on the importance of Strauss, and with Harry’s death one of the most informed and articulate followers of Leo Strauss is gone.

The world will miss Harry, even those who never heard of him.


I went for a walk outside the house today. The sidewalk cracks are hard to ignore when you are in a walker.  Lunch time now and my schedule is not my own, more later…

It was a pleasant walk, and I look forward to many more, and longer. Tiring but that is good.



Senate Rejects Climate Measures NYT


The Senate on Wednesday twice rejected measures declaring that humans are causing climate change <http://topics.nytimes.com/top/news/science/topics/globalwarming/index.html?inline=nyt-classifier> . But in the course of those votes, 15 Republicans, including Senator Rand Paul of Kentucky, voted yes. Mr. Paul, who is considered a likely presidential contender next year, was joined by Senators Bob Corker and Lamar Alexander of Tennessee, Kelly Ayotte of New Hampshire, Lisa Murkowski of Alaska and Jeff Flake of Arizona, among others. Two other potential Republican presidential candidates, Senators Marco Rubio of Florida and Ted Cruz of Texas, voted no. It was the first time in years that senators had voted on a climate change measure, and it came in the course of a debate on a bill forcing approval of the Keystone XL <http://topics.nytimes.com/top/reference/timestopics/subjects/k/keystone_pipeline/index.html?inline=nyt-classifier> oil pipeline. President Obama is expected to veto the bill if it passes, as is likely, but lawmakers are using it to send their own political messages. Democrats had hoped to force Republicans on the record on the issue of climate change by introducing the two amendments.

And so the politics continue without noticeable actual science


On climate models and 0.1degree accuracies

Lies, damned lies, etc.


Mike of course has it correct. Measurements made by hundreds of instruments, not calibrated to a consistent standard, representing different times of day at different locations over years of time, cannot be averaged in any meaningful fashion to create a statistically meaningful grand ensemble. And of course the temperature measurements are not independent – the temperature of the water an hour later and 20 nautical miles distant is highly correlated with the temperature of the original water sample. Finally, of course, we are estimating systematic air temperature from a variety of metrics including seawater temperature for comparison with modern air temperature measurements.

Admittedly one can make corrections for all of those factors, but said corrections are of course model-dependent and hence subjective – and such corrections are one of the criticisms of the GW people in their consistently manipulating data to make the past appear cooler so as to exaggerate the effects of warming.

And of course extrapolations of such data into the future, even under the simplest of assumptions (the linear data hypothesis with uncorrelated random errors) diverge hyperbolically from the midpoint (in time as determined by the weighting of the measurements if available, assuming that the measurements are uniform in weighted error across the period), so that the forecast error diverges as (elapsed time)^2 measured from the midpoint time.


Dear Jerry Pournelle PhD,

Your blog entry today (or maybe yesterday your time) with Mike clarifying the necessity of having a "set of data that is not in statistical control" made me understand the root of the problem. Your blog is truly a source of knowledge, and more importantly a gentlemanly discussion of difficult matters. From now on I am proud to be a platinum subscriber.


Rune Aaslid PhD

Most social science departments offer their own statistics courses because the math dept. statistics courses are too hard or require prerequisite courses. Of course this is because real statistics is hard, and requires real math. Stat in the Education Dept. or in Psychology is really cookbook stat on calculating means, and standard deviations, but has little to nothing about distributions, assumptions, or requirements for valid inferences, which is why so many “experiments” cannot be repeated even though they are “significant to the 10% level” etc.  They often mean nothing.  Alas this is true in some “hard” sciences,too. The worst of it is that many scientists who know much about physics know little about statistical inference and the assumtions in their models.  I was fortunate in that Paul Horst required me to go to the math dept for probability and statistics, which led me to operations research  which turned out to be more valuable than psych.

One reason for this journal is to encourage rational debate. I don’t presume to know everything even if I sometimes appear to pretend to, but I have many readers with great expertise.





Freedom is not free. Free men are not equal. Equal men are not free.




Bookmark the permalink.

Comments are closed.