eirias: (Default)
[personal profile] eirias
I want you all to go read a fantastic blog post by linguist Mark Liberman. It will explain in very clear terms, if you didn't already know, why mass-media reports of scientific findings about group differences are so frequently misleading in a way that engenders and supports stereotype. It is THE most important statistical fact I know and I never learned it in school even though I have a PhD. I learned it from Mark Liberman.

Go read it now.

(no subject)

Date: 2009-04-06 09:35 pm (UTC)
From: [identity profile] leora.livejournal.com
This article makes me want to cry.

It's not the stats. While I like the way it describes things and gives you a way to think about it, my reaction was - umm yes, that's what it means, don't people know that? I think I got an intensely good statistical education starting from a very young age. I was raised with books of puzzles and info that included a lot of stats tricks and explanations for them. I read books about how statistics work and how people use and misuse them. I think I just have a fairly good intuitive grasp of correct statistical use. Although I'm crap at remembering formulas for calculating them or getting some aspects of the terminology right.

No, what makes me want to cry is their average working memory size. The poor average was 8.5 items at a time.

I know I have issues with a small working memory. It causes me all sorts of problems and annoyances. But 8.5... geez... Mine is about 4.

I work around it. In school, I was a huge fan of pen and paper. I don't do math in my head. I take notes. I use memory tricks. But my working memory size seems to tend to be at around 4 items. I cannot remember phone numbers unless I can chunk them.

It obviously wasn't a crippling problem. I got into a good college and graduated with a good GPA and a double major. Like many weaknesses, it can be worked around. But 8.5 items... I've been going with The Magic Number 7 Plus or Minus 2 and thinking... well, I'm a little on the low end, but it's not that bad.

I wonder how I'd do on formal tests. I only have an informal idea of my working memory capacity.

(no subject)

Date: 2009-04-06 09:48 pm (UTC)
From: [identity profile] leora.livejournal.com
But I decided to share the link on Facebook, because I think you're right that more people should read it. And as I said there, even I found it gave me an easier way to think about and describe properties that I already knew about statistics.

On a side note, I recently skimmed back over my favorite stats book from one of my college stats classes. It's called Seeing Through Statistics by Jessica Utts. It's quite lovely in how it's full of examples and then asks insightful questions about what that really means and what problems might there be? It goes into lots of bit about experimental design and what the numbers you get actually represent. I've been really pleased with a lot of my statistics education. Also, one of my profs was great at tossing out random facts for us to think about and work out solutions to. The one I remember is that 90% of first-time mothers will hold their babies with their right hand with no prompting. We were asked to hypothesize why that was and come up with ways to test for it. It turned out to be surprisingly interesting and the tests run to try to verify the most likely reason were fascinating.

(no subject)

Date: 2009-04-06 10:58 pm (UTC)
From: [identity profile] cognative.livejournal.com
Is it because 90% are right handed?
Or did you mean correct?

(no subject)

Date: 2009-04-06 11:28 pm (UTC)
From: [identity profile] leora.livejournal.com
That was the very first guess that I came up with and that my class came up with. The stat matches so well and it's so obvious. Alas, it's not the case. I did mean right as in not the left hand. And roughly 90% of right-handed new mothers and roughly 90% of left-handed new mothers will pick up their child with their right hand.

They were asked why they did so. The right-handed mothers tended to say things like: well, it's my dominant hand and my baby is very important to me, so I want to hold my baby in my right hand.

The left-handed mothers tended to say things like: well, holding my baby doesn't require a lot of precision and this way my dominant hand is free for tasks like opening doors and such.

This is when you start to suspect that their brains are rationalizing justifications for behaviors that they do not know why they are driven to do.

(no subject)

Date: 2009-04-06 10:42 pm (UTC)
From: [identity profile] smarriveurr.livejournal.com
That's really fascinating. I'd never really thought about the way a layman's terms article almost automatically removes all the useful information. I mean, I knew that popular reporting on scientific studies was sensationalized and largely uninformative, but I never really grokked why.

(no subject)

Date: 2009-04-07 02:43 am (UTC)
paperkingdoms: (Default)
From: [personal profile] paperkingdoms
::nods:: It is. I hadn't made the connection, either, and I've taught lowest denominator stats.

(no subject)

Date: 2009-04-07 02:45 am (UTC)
From: [identity profile] smarriveurr.livejournal.com
I do at least feel better that people who work with this stuff didn't connect or get taught it. I have the excuse of only taking prob and stats in college and a quick stats course before then.

(no subject)

Date: 2009-04-07 04:27 am (UTC)
From: [identity profile] eirias.livejournal.com
I'm not sure that most people who teach stats get it, yeah. Most of them have got the logic of null hypothesis testing wrong, too.

(no subject)

Date: 2009-04-06 11:48 pm (UTC)
feuervogel: photo of the statue of Victory and her chariot on the Brandenburg Gate (Default)
From: [personal profile] feuervogel
I think you should add "And if you're not subscribed to the Language Log RSS feed, go do so now." ;) LL is awesome. (I added it after ... leora? darch? linked to one (of many!) anti-Strunk&White posts.)

(no subject)

Date: 2009-04-07 03:55 am (UTC)
From: [identity profile] leora.livejournal.com
Darch. While I like LL (although I don't always read it) the anti-S&W thing is very darch. Although I had to take back my criticism and then take back my taking back of it. See, my edition of S&W is actually really good and sensible. It doesn't state rules as absolutes, but states general guidelines that are actually true.

Later I read that later editions after the death of (I forget, one or both of) the authors modified it in ways that they would not have approved of.

(no subject)

Date: 2009-04-06 11:54 pm (UTC)
From: [identity profile] drspiff.livejournal.com
Yeah. As a scientist and a teacher I read that and consider the interesting effect that nurture vs nature appears to have on working memory.

But the other thing a lay person might miss is that your working memory can be developed or atrophy. It isn't static.

But I also see how this gets twisted in the MSM. Good link. Thanks!

(no subject)

Date: 2009-04-07 03:57 am (UTC)
From: [identity profile] leora.livejournal.com
It can? Do you have any details on that? As I wrote, mine is extremely small. I have practiced chunking techniques so that I can effectively carry more data, but I cannot find a way to increase the number of chunks I can hold. I'd love to read more about it. Although, realistically, if the info isn't on the web then I probably won't end up accessing it.

(no subject)

Date: 2009-04-08 03:41 pm (UTC)
From: [identity profile] drspiff.livejournal.com
I'll admit I have no citations, only personal experience and experience with students. "Chunking" works but is it expanding slots or making better use of what you have? I'd argue that it is hard to tell.
There is also the a nature versus nurture thing. I have trouble believing with everything we know about the adaptability of the brain that one is stuck with a fixed number of slots. Chunking seems to show that it is more of a software issue than a hardware issue. I'm not saying its easy to change it but considering all factors it seems reasonable to assume that a brain can be conditioned to expand the number of slots by adapting to prolonged environmental stimuli. A static limit at a small number would seem to be highly inflexible considering all the evidence that brains will adapt to environmental challenges, even retasking parts of the wetware to do so.

(no subject)

Date: 2009-04-08 06:14 pm (UTC)
From: [identity profile] leora.livejournal.com
Well, chunking has been studied. They got someone up to the ability to memorize some insane number, like 100 random digits, through chunking. But it doesn't affect the base limitation of number of units. This is part of why I try to work on chunking.

However, my question really is not so much, can the limit be changed, but can the limit be changed in an adult, since I'm more interested in my personal case. And if so, how? I don't know what I'd need to do to change my limit. When I try to hold more items, I just drop some. Which is why I use mnemonics and chunking and other memory tricks. I'm getting a bit old to have brain flexibility do me much good :/ Although maybe someday my limit will get larger.

(no subject)

Date: 2009-04-07 01:03 am (UTC)

(no subject)

Date: 2009-04-07 01:20 am (UTC)
ext_25834: (Default)
From: [identity profile] miang.myopenid.com (from livejournal.com)
I never learned it in school even though I have a PhD.

I find that really surprising, as the stats courses you and I took overlapped quite a bit.

That said, it's nice to have a plain-language link to throw at people when I'm on a bitchfest about media treatment of science research, so thanks. :D

(no subject)

Date: 2009-04-07 04:00 am (UTC)
From: [identity profile] eirias.livejournal.com
Oh, sure, people talked about effect sizes. But I don't recall anyone casting it in this stark light to make clear that what it really meant was that with most psychology effects, a randomly chosen person from High Group is not going to have very high odds of scoring higher than a randomly chosen person from Low Group. Better than even -- but not very high. That means ever so much more to me than just saying "it's a small effect," which is vague and mushy.

I got another layer of understanding from teaching for RJ, when explaining the relationship between standard deviations and the standard error of the mean; but that happened after ML's blog gave me the initial clue. I do think RJ understands the meaning of stats more clearly than most of the other teachers I've had. Not all, but most. They've all been decent technically, but this intuitive stuff, the part that is really important epistemologically? I didn't get it from them, in any case.

(no subject)

Date: 2009-04-08 06:18 pm (UTC)
From: [identity profile] leora.livejournal.com
I may have mentioned this before, but my first college stats teacher was one of the most amazing teachers I have ever had. He had almost the entire class bored. Everyone felt the material was ridiculously easy and going way too slowly... so, I looked back over how much we'd actually covered, and I realized it really wasn't. Somehow he made what he taught all seem so obvious that the whole class was not just learning it, but feeling it was way too easy. I was so glad to have him for stats 1 and 2. My stats 3 prof was good and he did things to make the class interesting, but he wasn't wonderous like my first stats prof. We need to clone him and send him around America to teach.

(no subject)

Date: 2009-04-24 03:34 am (UTC)
From: [identity profile] botia.livejournal.com
So, you list lobster porn as in interest.

Does that mean you are a scientist that has to watch lobsters mate as part of your work (one of my friends has been watching vole sex for a while), or does it mean you are a fan of Salvador Dali?

(no subject)

Date: 2009-04-24 12:07 pm (UTC)
From: [identity profile] eirias.livejournal.com
It means I saw it on someone else's interest list and thought it was funny. :) (A few years ago my husband decided he wanted his totem animal to be a lobster, so when I saw that interest I found it kind of hilarious.)

(no subject)

Date: 2009-04-24 01:10 pm (UTC)
From: [identity profile] botia.livejournal.com
Haha, fair enough. You should read "Dali & I" ... apparently Dali was a huge freak about lobsters and included live ones in his orgies.

Profile

eirias: (Default)
eirias

December 2016

S M T W T F S
    123
45678910
11121314151617
18192021222324
25262728293031

Most Popular Tags

Style Credit

Expand Cut Tags

No cut tags