POST OF THE DAY

[Arakawa]

Heh. Oops.

Will this work instead?

http://www.sandraandwoo.com/2009/01/01/0021-allowance-reverse-default-swaps/

_{Not quite 'funny enough to share', but I got nothing else at the moment.}

(Also click 'Next' for the punchline.)

[Brian]

Ah, nothing to apologize for.  I just thought it was interesting we both found it amusing enough to share. :D

[Music-chan]

http://www.masterle.org/

If you've seen that really amazing Iron Man costume (with the moving parts), this is the website for the guy who built it.

[Brian]

<Kuyou> ...I hate to be 'that guy'.
<Kuyou> But while he can make impressive costumes and websites.
<Kuyou> He doesn't seem to understand what an apostrophe is used for.

[Music-chan]

Yeah, his grammar is pretty shit but oh well. =p  Still interesting if you want to see his other cosplay projects.

[Brian]

Be warned: http://www.youtube.com/watch?feature=player_embedded&v=-G3RoBHMu-o

Edit: I've got and have read that guy's book.  Key advice: "Don't wear t-shirts with anti-robot slogans in robot neighborhoods."

[Dracos]

Yay Robot Uprisings :)

[Brian]

Caveat: Nakamabot will always be your ally.

[Brian]

This must have taken quite a bit of work, but I know Drac will 'love' the ending:

http://www.youtube.com/watch?v=QwNVQvygCNQ

[Dracos]

Noooooooooooooo~!

[Merc]

Drac doesn't sound like he's in love! =p

[Brian]

That's because Drac is always ahead of me.

* Brian sends out a friendly blue shell to cheer Drac up. >:3

[Arakawa]

http://www.newyorker.com/reporting/2010/12/13/101213fa_fact_lehrer?currentPage=all

1. The universe likes to violate the "laws of statistics" for the express purpose of embarrassing scientists.

2. The statistical study, that bread and butter of modern academic scientific research, is in fact an ungodly sausage factory which tends to produce useless nonsense far more often than anyone would care to admit. (Quote: "The disturbing implication of the Crabbe study is that a lot of extraordinary scientific data are nothing but noise.")

3. Something else.

Pick your hypothesis.

[KLSymph]

2. The statistical study, that bread and butter of modern academic scientific research, is in fact an ungodly sausage factory which tends to produce useless nonsense far more often than anyone would care to admit. (Quote: "The disturbing implication of the Crabbe study is that a lot of extraordinary scientific data are nothing but noise.")

First, yes. It's so, so easy to misapply statistical methods and end up with terrible conclusions. As they say, lies and damned lies....

But second, no. I don't think it's all that disturbing. You're not supposed to read some scientific finding in a journal and take that as some kind of gospel truth about the world. Scientific publications are a means of communicating and rigorously documenting data within a field, but the stuff like the conclusions aren't taken all that seriously. A non-scientist isn't intended to conclude that a statement is true because it's been published, any more than a student is intended to conclude that a statement is true because it's in a textbook. Sure, it happens, but that's not how it's meant to work.

A "significant" result is defined as any data point that would be produced by chance less than five per cent of the time.

If the writer of the article is going to jab scientists on statistics, he really shouldn't get this as wrong as he did. Statistical significance is not a statement about the probability whether a datapoint in an experiment occured by chance. It is a statement about the probability that, assuming the experiment measures chance occurrences, the datapoint would be at least as extreme as it was.

[Dracos]

The statement: "Statistical studies are a sausage factory" misunderstands that most things are.

Especially when dealing with minute genetic or psychological variation.  Medical science in particular is especially vulnerable as most studies include folks being interested and attentive to the individuals in question, something not considerably present for many people as a standard thing.

The article lists several things that are reasonable...but none of which really invalidates science or invention.  Yes, some (even a large amount) of scientists fall into confirmation bias, dislike data that nulls/disproves their previous tests, prefer to group success articles, yadda yadda.  Yes psychological, behavorial, physiological, and minute environmental is particularly vulnerable to this.

KL is right.  Publications aren't about conclusion.  It's about process and data.

*shrugs*

Nom, lunch on.