I'm running out of cowgirl pictures that aren't family-friendly...
Me on BV
1. Perhaps you have forgotten that China is big. If so, I am here to remind you.
2. Why I'm not worried about all the papers about data mining and p-hacking.
3. Niall Ferguson continues not to be a serious or trustworthy public intellectual on matters of monetary economics.
From Around the Econ Blogosphere
1. Scott Sumner lavishes praise on Matt Yglesias. Sumner is a passionate man - he loves as he hates, without reservation.
2. Paul Krugman says important and necessary things about the ideological war against infrastructure investment, but fails to mention America's huge cost problem.
3. Here's why we use log returns, courtesy of Ren & Stimpy.
4. Alex "The Rock" Tabarrok has good news on the U.S. patent[ly f***ed up] system.
5. The left-right political axis is actually just a made-up piece of B.S.! Astonishing. But you're still all a bunch of right-wing nutjobs, FYI.
6. Cool Campbell Harvey paper about backtesting trading strategies. God, I hope there aren't people out there trading on strategies with a backtest t-statistic of 2.01...OK I'm kidding, no amount of silliness in the finance industry would ever surprise me.
7. Tony Yates reminds us that all of the models where we assume some simple passive behavior for (fiscal, monetary) policy and then examine the effects of (monetary, fiscal) policy are basically missing the point.
8. Jim Hamilton says that Saudi Arabia isn't likely to cut oil production to boost prices.
9. John Cochrane has taken up the Neo-Fisherian banner! Low interest rates cause deflation!
You're running out of cowgirl pictures that are family-friendly, right? Or is this blog about to get a whole lot more racy?
ReplyDelete33% selective reasoning in support of glibertarian handwaving, must have been a slow week.
ReplyDeleteRe. p-values:
ReplyDeleteCalling a result with a p-value > 0.05 a “negative result” is wrong -- it is a failure to find strong evidence for a hypothesis, possibly because the sample was small. The actual effect might be both large and life-or-death important.
Calling a result with a p-value < 0.05 a “significant result” can (even without data mining and selection effects) be grossly misleading to the typical reader. It may be the result of a statistically powerful study that has identified what by most standards is a negligible effect on something of small importance, like a 1% greater relative risk of hangnails.
According to an article in Nature several years ago, alleged scientists not uncommonly regard a result with p > 0.05 as rejecting the positive hypothesis, which is just insane.
Noah, today isn't Thursday, I think you mean "that are family-friendly", and it's Halloween, not April fools.
ReplyDeleteFair points, all.
DeleteI'm leaving this comment up for teh lulz.
ReplyDeleteThis comment seems weirdly out of place on an economics blog.
ReplyDeleteWhat exactly did you mean by "America's huge cost problem"? I think it refers to the environmental, regulatory, and time requirements regarding getting many infrastructure projects going, but I'm not sure. Any reading suggestions on the topic(s)?
ReplyDelete"John Cochrane has taken up the Neo-Fisherian banner! Low interest rates cause deflation!"
ReplyDeleteMeaning that U Chic is now OK with 'correlation = causation'