Posts by James Green
Last ←Newer Page 1 2 3 4 5 Older→ First
-
when I saw an author called....Student
For a couple of years I assumed it was because it was an uncomplicated test, therefore used mostly by students.
However, the real story is much better. William Sealey Gosset worked for Guinness, but was prohibited from publishing by his employer, so used the pseudonym A. Student. Guinness used statistics both for beer production, but also barley production. Actually, it's rather disturbing how much of statistics is based around agricultural sampling in a field. I used to exhort my students to have a Guinness every day they used the t-test.
-
The book so good, he's writing it twice? (Or, 1.996 times, final number may decrease as Jolisa finds more free time).
I heard a rumour that this number approximates a Z score, which means that, because it exceeds 1.98, we can be 95% confident that the book is his. Luckily he's an academic not a student though. For if we had to use Student's t distribution, then we might be forced to accept the null hypothesis. <too early in the week for obscure stats jokes?>
-
it's probably my last chance for this decade
Eeek. I'm not sure that's a thought I needed!!
<rocks quietly> -
I plead distraction. My boy was installing a new power supply, motherboard, quad-core CPU etc in his computer. I was concerned that he didn't set the house on fire.
He didn't.
It's pretty hard with these new fangled computers. The last time I set one on fire was 2006. And that was because the floppy drive(!) I was installing didn't have any housing around the outside of the the socket, so I inadvertantly put it on 3 out of the 4 pins, and presumably shorted something, because the plastic on the wires burnt off. But that was a 10 year old floppy drive. Everything is pretty easy now; or at least I haven't started any fires.
-
The closest I ever got was a crowd of about thirty listening to him play at The Gathering. He started by berating us for shelling out $150 to listen to other people's CD collections, but went on to treat us with the goods.
-
I took delivery of one on Monday. The larger 6-cell battery gives it some junk in the trunk so it's quite a bit more than 1.35kg and sits high at the rear, but it's still super portable. Seems quite snappy and the speakers are surprisingly good. Multitouch works beautifully - a little too well in fact as it renders the clickable buttons practically useless. But I'll live with it - I bought a slightly (*cough* £580) cheaper laptop to leave budget for a nice 24" Dell ultrasharp LCD and good wireless mouse. Best of both worlds. I'll be requesting a Windows 7 upgrade as soon as Dell sorts out their site; it's not accepting my service tag right now.
Hey Zach! Sounds awesome. Unfortunately, Dell are treating this as a 'home' offering, and I want the computer for work, which requires it to be able to join a domain, which means not home premium. I currently have the Dell XPS M1210, and see the 11z as like a slightly smaller meaner version.
As to dell and reliability. The VGA port on mine died a few weeks ago. This means I now have a horrible inelegant solution to running my 22" monitor. Bring on HDMI ports on laptops I say (as Dell in fact are). Oh, and every now and then the wifi craps out (sometimes after 5 minutes, sometimes after days, requiring a restart). I do really love the keyboard action on my XPS.
I also got a Dell for a family member. It ate 2 hard-drives before they decided it was motherboardFail, but sent a service tech to their house in the country to replace. So can't fault their service.
-
@Phil Bahahaha. I give up. Too close to beer o'clock. Although on my 3rd answer I now reckon 0, and I think I've got a better pattern going. My ineptitude reflects a lack of practice, among other things.
-
Given that they require it of international applicants, though, they're loading the scales a little. They don't exactly say "we are testing your knowledge of maths and English and the intricacies of our antiquated and useless measurement system."
Come now. If you're using numerical indices of reliability, then having a whole load of questions that are failed by a group of people indicates they're good quality questions. You'd have to be concerned about validity to consider that it might be problematic to have a group of people from a different cultural background systematically scoring lower on a group of questions.
And why worry about that type of validity, when you could simply look at how well people's test scores predict their graduate school performance? Next you'll be suggesting that this is a poor measure of validity because people who score highly will go to more prestigious schools with better resources, and that this will somehow make them achieve better. Pssh. Free thinker.</sarcasm>
Bwahahahahaha. It took me a moment to realise that while we're talking about the GRE, I could just as easily be mocking the new standards system.
-
(The question I remember went something like "Congress is to gubernatorial as senate is to ...?". And James thinks this is an IQ test?)
That's crystallised intelligence, as opposed to fluid intelligence. And speaking of straws to hold onto. Your crystallised intelligence keeps increasing often well into your 60s or 70s, whereas fluid intelligence starts to decrease by the late 20s.
NB: Theoretically, IQ tests are not supposed to be culturally bound, which is, like most other things about IQ tests, fail. Although the GRE as a test is bound to the US, so I guess it's not too outrageous to have US knowledge questions.
-
When somone can explain with precision what intelligence and is not I will entertain the idea of measuring it.
There are lots of annoyingly abstract things in science that are difficult to measure, and psychology is blessed with many such. Uh, happiness, depression, intelligence ...
What makes it amusing, is that there's a big difference between a test showing some validity on average over individuals, and one that is predictive for an individual. So higher scores on a test might be associated with a higher salary, but that doesn't mean that a person with a high score will get a high salary. Or whatever.
They are, as someone already pointed out, a laziness tool. It's really hard to come up with a valid way of measuring something like intelligence, and a really good measure is likely to appear subjective and easy to criticise. Wouldn't it be much easier and cheaper to come up with something much more reliable that appears nice and objective, but that has much lower validity. It's why things like impact factors and the h-index are so popular. Interestingly, I'd suggest that PBRF is trying to head down the more valid/less reliable direction; but whether it has win is debatable. Plus it's Friday!