« My email to Brad DeLong | Main | China book fact countercyclical asset of the day »
Assorted links
1. Michael Spence defends the Geithner plan.
2. Chimpanzee markets in everything; research is here.
3. Via Michelle Dawson, more on the fMRI voodoo correlations debate; relevant papers can be found here.
4. A tauntaun sleeping bag: do you get the reference?
Posted by Tyler Cowen on April 8, 2009 at 01:56 PM in Web/Tech | Permalink
Comments
The tauntaun bag (from Empire Strikes Back, natch) is an April Fools, I believe. Sadly. If they had one for a 6'1 man I'd buy it.
http://www.thinkgeek.com/stuff/looflirpa/2009.shtml
Posted by: Luke G. at Apr 8, 2009 2:03:57 PM
Who cut the tauntaun? It smells in here.
Posted by: jurinaturalist at Apr 8, 2009 2:28:55 PM
What an insulting question! Who wouldn't get that reference?
Posted by: Robert at Apr 8, 2009 2:37:59 PM
Yes, the tauntaun sleeping bag was an April Fool gag, but the response was so overwhelming that ThinkGeek is trying to get the license from LucasFilm to make them for real.
Personally, I'd rather have the Squeez Bacon: http://www.thinkgeek.com/stuff/41/squeez-bacon.html
Posted by: Franklin Harris at Apr 8, 2009 2:54:37 PM
April Fools, I believeLooks like they are trying to bring it to life, though:
ATTN Tauntaun Fanatics! Due to an overwhelming tsunami of requests from YOU THE PEOPLE, we have decided to TRY and bring this to life. We have no clue if the suits at Lucasfilms will grant little ThinkGeek a license, nor do we know how much it would ultimately retail for. But if you are interested in ever owning one of these, click the link below and we'll try!
And didn't this happen to ThinkGeek last year as well? ...
Yes. And, apparently, the year before...
Posted by: Bob Montgomery at Apr 8, 2009 2:56:58 PM
I read Spence's article, and I still do not see why the loans should be non-recourse, other than as a back-door subsidy to private equity and the banks. If the plan is simply "designed to restart frozen markets", why do the private equity folks get to keep their winning bets and stick the taxpayer with the losers?
Posted by: Nemo at Apr 8, 2009 3:04:35 PM
Hells yes. That's worth it just to bother the mother-in-law.
Posted by: Andrew at Apr 8, 2009 3:09:54 PM
The chimp article is fascinating and makes a lot of sense. It does make me wonder why so many people today want to be vegetarians. Funny enough, I am typing this comment right next to an ad titled "Meet Your Meat."
Posted by: Yan at Apr 8, 2009 3:23:45 PM
1. All other things being equal, the plan is bad in that it finagles price without addressing value and separating the mortgage paper renegotiation one more step from the tradeable security.
Posted by: Andrew at Apr 8, 2009 4:59:49 PM
*Blinks at fMRI controversy.*
This is one of the most epic failures of statistics I have ever encountered. I shall not be trusting future statistics about fMRI correlations unless I know how they were obtained.
Posted by: Eliezer Yudkowsky at Apr 8, 2009 5:21:10 PM
Isn't the fMRI "Voodoo Correlations" issue just an in-sample / out of sample problem? That is, the brain regions in question were not specified in advance and then found to be correlated with personality, but were first found to be correlated?
Posted by: David at Apr 8, 2009 5:38:25 PM
References mimic the fundamental comedic mechanism w/o delivering the goods
Posted by: Fenn at Apr 8, 2009 5:59:04 PM
Chimpanzees exchange meat for sex
Chimpanzee markets in everything
Not really. It's not a market transaction; it's closer to a date than prostitution.
And a potlatch isn't a company.
Posted by: Jason Malloy at Apr 8, 2009 6:11:05 PM
David: Essentially yes. It's an old and pretty simple error. The problem is when you search the whole brain for areas where activity is significantly correlated with personality, using conservative significance testing. Any areas which pass such tests must be strongly correlated by definition (that's the only way they could pass the tests.) But people have been reporting the correlation values and holding up their high values as evidence that they are somehow especially exciting.
Eliezer: It's pretty epic, but it's not the first time this error has occured - in fact Vul et. al. point out several other recent cases at the end of their response to critics. And to be fair to the fMRI researchers concerned, it's one of those things that's obvious once you've been told about it, but not before.
Posted by: Neuroskeptic at Apr 8, 2009 7:00:18 PM
Wait, do the male monkeys get sex because they share the meat or do they share the meat because they get sex? The articles note that there are monkeys who don't share meat and don't get sex. Are they not interested in sex, or do they refuse to share because they know they won't get any sex anyway?
Posted by: Doug at Apr 8, 2009 8:20:40 PM
voodoo:
This is a great opportunity to study publication bias. If fMRI publications were pure noise, selected for statistical significance, or were satisficing fraud, the papers making the non-independence error should have similar results to those that don't.
Posted by: Douglas Knight at Apr 8, 2009 9:23:20 PM
You would have to be incredibly uncultured not to get the reference.
Posted by: Garrett at Apr 8, 2009 9:59:00 PM
DK: Not necessarily, because the non-independence error could be the only way to generate extremely "good" results (in the subset of fMRI papers in question, which is quite a narrow one). In which case just publishing significant noise without the error would still not generate the same results.
In fact Vul et. al. did compare the erroneous papers to those which did not make the error and found that they reported "better" correlations, although the difference was fairly modest.
Posted by: Neuroskeptic at Apr 9, 2009 5:16:41 AM
Isn't the fMRI "Voodoo Correlations" issue just an ordinary grant-seeking/tenure-seeking/fame-and-fortune-seeking/is-this-what-science-has-descended-to problem?
Posted by: dearieme at Apr 9, 2009 6:26:37 AM
Not really. Although the paper's original title of "Voodoo Correlations in Social Neuroscience" was a bit cheeky.
Posted by: Neuroskeptic at Apr 9, 2009 7:53:57 AM
Iceland has collapsed so thoroughly that at this point, it's only economically viable export may very well be an internet spaceship game, and that internet spaceship game's króna is for all intents and purposes a more real and valid and valuable currency than the actual country's actual money.
http://crisper.livejournal.com/198722.html
Posted by: davidc at Apr 9, 2009 10:28:03 AM
re: the Tauntaun bag: We're guessing that even if they get past Lucasfilms, their bigger problem is going to be the CPSC and compliance with CPSIA. They have targeted it to kids 7-13 years, which clearly overlaps the 12 and under crowd. On the other hand, since NRDC doesn't think CPSIA applies to their onesies (whose very name is a trademark violation), why should it apply to Thinkgeek?
Posted by: Eric H at Apr 12, 2009 11:35:25 AM
