![[personal profile]](https://www.dreamwidth.org/img/silk/identity/user.png){kind=small}
![[identity profile]](https://www.dreamwidth.org/img/silk/identity/openid.png){kind=small}
So I may have a Phd in "maths", but it's pure maths, and I am woefully ignorant of statistics, not having studied it beyond a first year level (I've tutored second year stats. That was a challenge :D)
Thus my mind is actually quite easily blown by fairly basic statistics facts presented in an engaging way. See for example I'm not normal, which makes an obvious point I'd never really thought about before: The normal distribution became central to statistics back before computers, and once people get taught it, unless they go on to study more stats it's their only hammer so everything starts to look like a nail. But lots of large samples are NOT normally distributed, and given computing power we generally don't need to smoosh them into a nice distribution at all but can actually look at what the real data is doing.
My mind was also blown by Full House: The Spread of Excellence from Plato to Darwin which is pretty much just a basic description of evolution being a bounded random walk (Namely, the mean complexity goes up over time while still leaving 99.9% of life at the same basic level, with anything but the most simple species being as likely to get more simple as less)
Thus my mind is actually quite easily blown by fairly basic statistics facts presented in an engaging way. See for example I'm not normal, which makes an obvious point I'd never really thought about before: The normal distribution became central to statistics back before computers, and once people get taught it, unless they go on to study more stats it's their only hammer so everything starts to look like a nail. But lots of large samples are NOT normally distributed, and given computing power we generally don't need to smoosh them into a nice distribution at all but can actually look at what the real data is doing.
My mind was also blown by Full House: The Spread of Excellence from Plato to Darwin which is pretty much just a basic description of evolution being a bounded random walk (Namely, the mean complexity goes up over time while still leaving 99.9% of life at the same basic level, with anything but the most simple species being as likely to get more simple as less)
no subject
The only maths I've done at a tertiary level is statistics (unless you count finance, which I guess is applied maths too), which I found unexpectedly fascinating and fulfilling. (Interesting to see the random walk concept applied in that context; it's something I apply every day to value financial instruments.)
Have you read Freakonomics? It's a pretty light read and I think you might find it enjoyable, both from a statistical point of view and to pick holes in. :)
no subject
I remember the random walk from thermodynamics in physics, it's funny how these things come up.
no subject
Absolutely agree with this; it's one of those peculiar books where I read it and then figured that even if it was All Wrong (and I'm certain much of it was), then the thinking I got out of it was valuable. I think I might be a little peculiar.
If you do decide you're feeling critically robust enough to go there at some point, I'd take the writers' conclusions with a pinch of salt, but the themes of challenging common wisdom and looking for alternative theories to support statistical findings (not to mention confusing correlation and causation - a trap the authors themselves fall into at times) are worthwhile.
no subject
no subject
When I was working in finance, I was exposed to a few of the ideas mentioned in the articles you linked to, and also discovered the whole controversy over the philosophy of statistics. Are you a Bayesian or a frequentist?
no subject
Are you a Bayesian or a frequentist?"
*googles*
*reads*
Um..
o.O
x_x
I dunno :)
(My brain is a bit dead today after just mopping the whole house after the cat was sick everywhere. I may come to a conclusion later after some pondering)
no subject
One of the things I especially like is that Salsburg roots his narrative in a social world, so you get stuff about how all the early computing power was provided by "computers" -- women who were employed to compute calculations all day, much in form like a secretarial pool -- and how the statisticians running the firm thought women were especially well-suited to that work, since they were so "docile and patient" and could be trusted to not let bursts of creativity disrupt them from the tedium. (It gets better, later, fortunately. In fact, women in stats is in way better shape than women in pure maths.) In terms of social world, you also get to see some of the academic fights, like how math departments missed the boat in not embracing stats as a legitimate, substantive academic field, and then how statisticians did exactly the same thing with computer science. ;-)
And now that I've got the book review out of the way... *puts on stodgy stats-geek hat*
Most stats education is pretty bad -- the vast bulk of it is for people who don't want to know anything more than how to calculate a number to put into a paper or report, and so what most people get is a simple recipe book. What is generally left out -- or rather, mentioned as a set of assumptions at the beginning of every recipe, and which is thereafter ignored by the students who just want their number -- is when and how these techniques break, what they can and can't do, how to decide if the question being answered is the question you wanted answered, let alone decide if that question is even answerable.
It's a discipline with tricksy philosophical issues threaded deeply throughout it, and most of that gets completely ignored by most people who handle or calculate statistics. Which makes me sad, because it's a pretty mind-blowingly awesome field.
no subject
It's a discipline with tricksy philosophical issues threaded deeply throughout it, and most of that gets completely ignored by most people who handle or calculate statistics. Which makes me sad, because it's a pretty mind-blowingly awesome field.
So my statsy friends keep telling me :)
no subject
I suppose you get a similar effect (long, low-population "tail" of higher-complexity organisms) regardless of the exact probabilities of species evolving to more / less complex forms. I wouldn't have thought those probs were necessarily close, or consistent from organism to organism? Cool stuff anyway.
no subject
no subject
The only maths I've done at a tertiary level is statistics (unless you count finance, which I guess is applied maths too), which I found unexpectedly fascinating and fulfilling. (Interesting to see the random walk concept applied in that context; it's something I apply every day to value financial instruments.)
Have you read Freakonomics? It's a pretty light read and I think you might find it enjoyable, both from a statistical point of view and to pick holes in. :)
no subject
I remember the random walk from thermodynamics in physics, it's funny how these things come up.
no subject
Absolutely agree with this; it's one of those peculiar books where I read it and then figured that even if it was All Wrong (and I'm certain much of it was), then the thinking I got out of it was valuable. I think I might be a little peculiar.
If you do decide you're feeling critically robust enough to go there at some point, I'd take the writers' conclusions with a pinch of salt, but the themes of challenging common wisdom and looking for alternative theories to support statistical findings (not to mention confusing correlation and causation - a trap the authors themselves fall into at times) are worthwhile.
no subject
no subject
When I was working in finance, I was exposed to a few of the ideas mentioned in the articles you linked to, and also discovered the whole controversy over the philosophy of statistics. Are you a Bayesian or a frequentist?
no subject
Are you a Bayesian or a frequentist?"
*googles*
*reads*
Um..
o.O
x_x
I dunno :)
(My brain is a bit dead today after just mopping the whole house after the cat was sick everywhere. I may come to a conclusion later after some pondering)
no subject
One of the things I especially like is that Salsburg roots his narrative in a social world, so you get stuff about how all the early computing power was provided by "computers" -- women who were employed to compute calculations all day, much in form like a secretarial pool -- and how the statisticians running the firm thought women were especially well-suited to that work, since they were so "docile and patient" and could be trusted to not let bursts of creativity disrupt them from the tedium. (It gets better, later, fortunately. In fact, women in stats is in way better shape than women in pure maths.) In terms of social world, you also get to see some of the academic fights, like how math departments missed the boat in not embracing stats as a legitimate, substantive academic field, and then how statisticians did exactly the same thing with computer science. ;-)
And now that I've got the book review out of the way... *puts on stodgy stats-geek hat*
Most stats education is pretty bad -- the vast bulk of it is for people who don't want to know anything more than how to calculate a number to put into a paper or report, and so what most people get is a simple recipe book. What is generally left out -- or rather, mentioned as a set of assumptions at the beginning of every recipe, and which is thereafter ignored by the students who just want their number -- is when and how these techniques break, what they can and can't do, how to decide if the question being answered is the question you wanted answered, let alone decide if that question is even answerable.
It's a discipline with tricksy philosophical issues threaded deeply throughout it, and most of that gets completely ignored by most people who handle or calculate statistics. Which makes me sad, because it's a pretty mind-blowingly awesome field.
no subject
It's a discipline with tricksy philosophical issues threaded deeply throughout it, and most of that gets completely ignored by most people who handle or calculate statistics. Which makes me sad, because it's a pretty mind-blowingly awesome field.
So my statsy friends keep telling me :)
no subject
I suppose you get a similar effect (long, low-population "tail" of higher-complexity organisms) regardless of the exact probabilities of species evolving to more / less complex forms. I wouldn't have thought those probs were necessarily close, or consistent from organism to organism? Cool stuff anyway.
no subject