Head-Fi.org › Forums › Misc.-Category Forums › Members' Lounge (General Discussion) › Statistics Project Due Tomorrow!!
New Posts  All Forums:Forum Nav:

Statistics Project Due Tomorrow!!

post #1 of 15
11/17/09 at 12:32am
Thread Starter 
Quick question - I'm supposed to gather, plot, and analyze bivariate data for this statistics assignment due tomorrow. Is stock price vs trading volume of a particular company (such as Google) a valid bivariate correlation? Would the addition of time (i.e. 1st day of each month in 2009) make it a trivariate data collection?..
Reply
post #2 of 15
11/17/09 at 1:39am
Thread Starter 
Anyone taking stats? Or basic stats?
Reply
post #3 of 15
11/17/09 at 1:57am
Stock prices have no or insignificant correlation to trading volume. Choose something else that has a stronger positive or negative correlation. And always keep in mind that just because two things have a correlation, doesn't mean there is a logical connection between them. Applying statistics to predict stock prices, especially in a in basic stats course, is a bad idea. Its almost like dealing with randomness.
Reply
post #4 of 15
11/17/09 at 2:03am
Thread Starter 
But is it bivariate or trivariate? Do I take time into account?

Any suggestions of stronger correlation that have to deal with finance/economics, and has easily attainable data points?
Reply
post #5 of 15
11/17/09 at 2:44am
Time will not make it any better in this case.

Economics has a lot more interesting correlation relationships than finance. Demand of product A vs product B, pricing, GDP, etc. There is plenty of info out there.
Reply
post #6 of 15
11/17/09 at 3:24am
Thread Starter 
OK. Decided on GDP. Thanks.
Reply
post #7 of 15
11/17/09 at 10:54am
You have not yet learned how to statisticulate. An excellent place to start is the possibly hard to find book How to Lie with Statistics. It taught me everything my formal courses did not.

Follow up: I did a search on-line, and the book does not seem to be hard to find. Get a copy, and consider that an assignment.
Reply
post #8 of 15
11/17/09 at 12:48pm
It appears you've made your decision, but the FTC can be a decent resource if you need it: FTC Bureau of Economics
Reply
post #9 of 15
11/17/09 at 4:12pm
Quote:
Originally Posted by Lazarus Short View Post
You have not yet learned how to statisticulate. An excellent place to start is the possibly hard to find book How to Lie with Statistics. It taught me everything my formal courses did not.

Follow up: I did a search on-line, and the book does not seem to be hard to find. Get a copy, and consider that an assignment.
QFT. This book is awesome. It's a sad truth that even though I took stats for years through school and uni (as part of a bio degree), I probably learned more useful stuff from reading this (pretty slim) volume, than all of the rest put together. It's a mini-masterpiece.
Reply
post #10 of 15
11/17/09 at 4:49pm
I'd suggest looking at GDP and oil consumption.
Reply
post #11 of 15
11/17/09 at 8:46pm
Thread Starter 
Already turned in, but thanks. 1st yr of stats (AP stats) for me. It's considered to be one of the easy AP courses at our school....
Reply
post #12 of 15
11/17/09 at 10:03pm
Two classes I wish I had taken: stats and econ. Learning what little I have on the fly has not been fun.
Reply
post #13 of 15
11/17/09 at 10:09pm
Thread Starter 
I've already taken introductory college econ (micro and macro). What really interests me ATM is strategy and game theory...
Reply
post #14 of 15
11/17/09 at 10:15pm
Quote:
Originally Posted by DeusEx View Post
I've already taken introductory college econ (micro and macro). What really interests me ATM is strategy and game theory...
But the number of real-world applications of those is small. A lot of stuff in the textbooks, but in real business seldom used.

Statistics is way more useful.

(I have papers in Econometrica and Management Science on pricing optimization, plus papers in The Annals of Statistics on multivariate two-sample tests, and have been doing business applications for 40 years, so trust me on this).
Reply
post #15 of 15
11/17/09 at 10:38pm
Quote:
Originally Posted by Lazarus Short View Post
...How to Lie with Statistics ...
From 1954. And pretty elementary. Still a quick read, and fun.

But you have to go beyond that. Try "Mostly Harmless Econometrics" ... this is what you need to read!

Amazon.com: Mostly Harmless Econometrics: An Empiricist's Companion (9780691120355): Joshua D. Angrist, Jorn-Steffen Pischke: Books

Not as fun or on-point, but still useful to have on your shelf: "Common Errors in Statistics"

Amazon.com: Common Errors in Statistics (and How to Avoid Them) (9780470457986): Phillip I. Good, James W. Hardin: Books

Finally, here's the best intro stat textbook ever written, IMHO:

Amazon.com: All of Statistics: A Concise Course in Statistical Inference (Springer Texts in Statistics) (9780387402727): Larry Wasserman: Books

I don't know any of these authors (I may have met Phil Good once years ago) and have no vested interest in any of these. But reading them will pay back, I promise.
Reply
New Posts  All Forums:Forum Nav:
  Return Home
Head-Fi.org › Forums › Misc.-Category Forums › Members' Lounge (General Discussion) › Statistics Project Due Tomorrow!!