Statistics for behavioural sciences class

wormie

Asshat wormie

2023 Asshat Award Winner
<Gold Donor>
16,820
30,964
Soriak_sl said:
That's actually a good question -- I only know of one course for sure at Carnegie Mellon, which is for undergrads in statistics:36-402, Undergraduate Advanced Data Analysis (2013)

Check out thismidterm, which is posted on the site. That's what a good assignment in the social sciences should look like: here's a problem, here's a data set, now go use what you learned and write a paper with your analyses and findings.

I actually heard aboutthis homework assignment, which came out around the time when a mistake made by two Harvard economists made it into the popular press. Basically, they failed to drag their Excel formula all the way down; once you fix that, their findings change significantly. (Oh yeah, their research was cited by policymakers in Europe for their austerity measures -- whoops.) Turns out that there was much more wrong with the analysis, and that assignment showed undergrads how to do it properly. Spoiler alert: the entire effect goes away and, in some cases, more debt actually leads to more growth. And that's why you don't skip statistics classes... or use Excel for your analysis.

Helps that all the professor's lecture notes are written like a book and available as a PDF free of charge, so a lot of people outside of Carnegie Mellon use it to teach themselves. The perk of being a grad student or professor is that you can just stop by your stats department and there's someone who will take the time to help you out if you get stuck. And the thing with stats is that if you use it, you should continuously refresh your understanding and learn new tools. Some of the directed acyclic stuff in the last chapters of his book is based on publications from 1-2 years ago and lets you do really cool stuff with respect to establishing causality (a huge problem in economics). The computing power to do that just wouldn't have been available even 5 years ago, whereas now you can just create your farm on Amazon and the sky's the limit.

So I figured I'd be save with up to 5 universities, because I really can't think of more than one and have not heard about a similar class offered elsewhere.
Click to expand...
You are comparing a course requiring probability theory and math stats as pre reqs to hatorade's class that probably barely requires high school algebra.
 
S

Soriak_sl

shitlord
783
0
wormie_sl said:
You are comparing a course requiring probability theory and math stats as pre reqs to hatorade's class that probably barely requires high school algebra.
Click to expand...
I guess I'm calling for more probability & stats requirements for people in the social sciences. It's pretty atrocious right now.
frown.png
 
wormie

Asshat wormie

2023 Asshat Award Winner
<Gold Donor>
16,820
30,964
Soriak_sl said:
I guess I'm calling for more probability & stats requirements for people in the social sciences. It's pretty atrocious right now.
frown.png
Click to expand...
Agreed. For those looking to pick up a thing or two, following is in the order from easiest to hardest:

Coursera.org

Statistical Learning | Stanford Online

Introduction to Statistics Course | edX,Introduction to Statistics: Probability | edX,Introduction to Statistics: Inference | edX

And a bit more advanced:

http://statweb.stanford.edu/~tibs/ElemStatLearn/
 
C

Creslin

Trakanon Raider
2,375
1,077
Knowing all the underlying math isn't really important unless you are doing actual research imho. Most people would be far better off just knowing how to read the conclusions of a paper and what the stats mean, but I don't see all that underlying matrix algebra and optimization stuff being that important unless you are doing your own work and not just trying to read someone elses.
 
Big Phoenix

Big Phoenix

Pronouns: zie/zhem/zer
<Gold Donor>
44,782
93,637
Lol stats. Business math is a prerequisite for business stats here. Show up to business stats and proceed to use zero of the math I learned and did everything by excel. Very jimmy rustling.
 
Palum

Palum

what Suineg set it to
23,595
34,111
Creslin_sl said:
Knowing all the underlying math isn't really important unless you are doing actual research imho. Most people would be far better off just knowing how to read the conclusions of a paper and what the stats mean, but I don't see all that underlying matrix algebra and optimization stuff being that important unless you are doing your own work and not just trying to read someone elses.
Click to expand...
Well we all know from Mist you only have to be able to read and interpret the work of others to form a complete understanding of life. Wait, how is that not a religion?
 
Palum

Palum

what Suineg set it to
23,595
34,111
Big Phoenix_sl said:
Lol stats. Business math is a prerequisite for business stats here. Show up to business stats and proceed to use zero of the math I learned and did everything by excel. Very jimmy rustling.
Click to expand...
Yea OK, I really want to differentiate trigonometric equations on paper again... In actuality there's a theoretical component that gives way to the practical component to pretty much every curriculum besides hard math. Even sometimes math, I guess. I remember learning some things that the professor would tell is we could not understand the proof, so here's how you solve it in 3 easy steps.
 
wormie

Asshat wormie

2023 Asshat Award Winner
<Gold Donor>
16,820
30,964
The way math is taught is a fucking travesty. College math should start with a logic/proofs course. Followed by discrete math and then by one or two courses in slightly above basic graph theory/combinatronics/group theory/number theory to solidify logic abilities. Then they should start teaching calculus. But not the retarded "here is some formulas, memorize this to solve problems on tests and nowhere else" method. It should be what most colleges call Advanced or Honors Calc. Teach that shit up to derivatives then follow up with another semester up to integration/measure theory. Then dive into more interesting shit which would be for math/stats/CS majors. Oh and linear algebra, linear algebra everywhere.

The way it is now just makes people not understand math and hate it and you cant blame them because it seems useless when you only learn shit that lets you pass a test in a course that you are taking only because its required and one that you dont understand in the first place.
 
khalid

khalid

Unelected Mod
14,071
6,775
wormie_sl said:
The way math is taught is a fucking travesty. College math should start with a logic/proofs course.
Click to expand...
I agree with you in principle, but if anything the pushback is harder in the other direction. It is the rare student who is interested in proofs, everyone just wants to learn the equations, not how to prove them or why they work. If you have a math class that is required for any business degree, they are constantly pushing for all proofs or even talk of proofs to be taken out. Even for courses required by engineering degrees, you have the same sort of thing. My facepalm as an undergrad was having a choice between "statistics" and "mathematical statistics". Wtf, there is a statistics class without the math?? Pretty much.

It seems that professors in any degree outside of mathematics just want their student to memorize the equations and could give a fuck about anything else. "Just apply it, you don't need to understand it". It seems weird to me, but changing it isn't going to happen and maybe I am in the wrong. Maybe all they really need to do is apply it shrug. Hell, look at this thread. Hatorade, who is probably far above the average student, still was tempted to completely skip the one math course in his entire curriculum.
 
wormie

Asshat wormie

2023 Asshat Award Winner
<Gold Donor>
16,820
30,964
khalid_sl said:
I agree with you in principle, but if anything the pushback is harder in the other direction. It is the rare student who is interested in proofs, everyone just wants to learn the equations, not how to prove them or why they work. If you have a math class that is required for any business degree, they are constantly pushing for all proofs or even talk of proofs to be taken out. Even for courses required by engineering degrees, you have the same sort of thing. My facepalm as an undergrad was having a choice between "statistics" and "mathematical statistics". Wtf, there is a statistics class without the math?? Pretty much.

It seems that professors in any degree outside of mathematics just want their student to memorize the equations and could give a fuck about anything else. "Just apply it, you don't need to understand it". It seems weird to me, but changing it isn't going to happen and maybe I am in the wrong. Maybe all they really need to do is apply it shrug. Hell, look at this thread. Hatorade, who is probably far above the average student, still was tempted to completely skip the one math course in his entire curriculum.
Click to expand...
The problem probably goes back to k-12. People are so terrified of math that when they are faced with something other than rote memorization, they shit themselves and complain to departments who in turn bitch at teachers. If logic was taught earlier than college all this shit might be avoided. Then again we have Professors who think shit like this:

Is Algebra Necessary? - NYTimes.com

so who the fuck knows.
 
D

Dabamf_sl

shitlord
1,472
0
Creslin_sl said:
Knowing all the underlying math isn't really important unless you are doing actual research imho. Most people would be far better off just knowing how to read the conclusions of a paper and what the stats mean, but I don't see all that underlying matrix algebra and optimization stuff being that important unless you are doing your own work and not just trying to read someone elses.
Click to expand...
Knowing the underlying math is not useful for people doing research in the behavioral sciences either. You need to know the conceptual very well, and to be able to test assumptions. The math is not needed.

I don't get the critique. No one with a bachelors degree in beh science is doing any research to my knowledge. The stats that we teach, at least, is half math half conceptual, and that + methods is fine for a bachelors. The info classes (ie everything else) are way more important. Only once you go to grad level is a thorough understanding of stats important, but still not the math.
 
Big Phoenix

Big Phoenix

Pronouns: zie/zhem/zer
<Gold Donor>
44,782
93,637
wormie_sl said:
The way math is taught is a fucking travesty. College math should start with a logic/proofs course. Followed by discrete math and then by one or two courses in slightly above basic graph theory/combinatronics/group theory/number theory to solidify logic abilities. Then they should start teaching calculus. But not the retarded "here is some formulas, memorize this to solve problems on tests and nowhere else" method. It should be what most colleges call Advanced or Honors Calc. Teach that shit up to derivatives then follow up with another semester up to integration/measure theory. Then dive into more interesting shit which would be for math/stats/CS majors. Oh and linear algebra, linear algebra everywhere.

The way it is now just makes people not understand math and hate it and you cant blame them because it seems useless when you only learn shit that lets you pass a test in a course that you are taking only because its required and one that you dont understand in the first place.
Click to expand...
I think people hate math because beyond simple multiplication and division, its almost all irrelevant in basic day to day living. You dont need to know why derivatives exist or what they are to get through your day.

Also you need not know the principles or mechanics behind something to make use of it.
 
Palum

Palum

what Suineg set it to
23,595
34,111
Big Phoenix_sl said:
I think people hate math because beyond simple multiplication and division, its almost all irrelevant in basic day to day living. You dont need to know why derivatives exist or what they are to get through your day.

Also you need not know the principles or mechanics behind something to make use of it.
Click to expand...
It's a valid point, scientific advancement at some point has to rely upon accepting the foundation and work of others once accepted as fact instead of reinventing everything. I mean, I went through differential equations, multidimensional calculus and linear algebra. I know for a fact many of the proofs for some differential equations are just beyond my learning to comprehend at this point. I use some of those math concepts at work, but I am content with the fact that I have the capacity to prove most of these things (or at least understand the proofs) IF I needed to put the effort into them (years in some cases) but I don't need to.

Practical vs. theoretical I suppose.
 
Mist

Mist

Eeyore Enthusiast
<Trapped in Randomonia>
30,474
22,325
Dabamf_sl said:
Knowing the underlying math is not useful for people doing research in the behavioral sciences either. You need to know the conceptual very well, and to be able to test assumptions. The math is not needed.

I don't get the critique. No one with a bachelors degree in beh science is doing any research to my knowledge. The stats that we teach, at least, is half math half conceptual, and that + methods is fine for a bachelors. The info classes (ie everything else) are way more important. Only once you go to grad level is a thorough understanding of stats important, but still not the math.
Click to expand...
I was arguing with my professor in one of these classes and accidentally derived the Poisson Distribution almost perfectly over the course trying to make my point. I mean, the professor obviously knew where I was headed and kinda tricked me into going through with the whole thing.

That's when I started taking more serious math classes. I still feel like I kinda fucking suck at it most of the time though.
 
Mist

Mist

Eeyore Enthusiast
<Trapped in Randomonia>
30,474
22,325
Palum_sl said:
It's a valid point, scientific advancement at some point has to rely upon accepting the foundation and work of others once accepted as fact instead of reinventing everything. I mean, I went through differential equations, multidimensional calculus and linear algebra. I know for a fact many of the proofs for some differential equations are just beyond my learning to comprehend at this point. I use some of those math concepts at work, but I am content with the fact that I have the capacity to prove most of these things (or at least understand the proofs) IF I needed to put the effort into them (years in some cases) but I don't need to.

Practical vs. theoretical I suppose.
Click to expand...
The vast majority of people who use computer programs have no idea how a finite-state automaton works, but most of the time get along fine.
 
Hatorade

Hatorade

A nice asshole.
8,198
6,627
Fucking shit, 45 mins on one test problem I couldn't get the answer right, turned out the answer in the book was fucking wrong. Almost a four-hundred dollar fucking book and it lies to me, fuck american schools and their insanely shitty way to get more money from students.
 
T

The Master

Bronze Squire
2,084
2
Hatorade_sl said:
Fucking shit, 45 mins on one test problem I couldn't get the answer right, turned out the answer in the book was fucking wrong. Almost a four-hundred dollar fucking book and it lies to me, fuck american schools and their insanely shitty way to get more money from students.
Click to expand...
Some textbook companies have a bounty for finding errors in the current printing.
 
You must log in or register to reply here.