sorta OT, Need some help with statistics class

[infinity5]

I was wondering if any of you guys could help me out. i'm taking a stats class right now and we've been assigned a project. the only problem is that i have to come up with the project topic on my own, and i'm clueless. i'm the kind of person that likes to be told(or know) what to do and get the job done, rather than sitting there thinking about what to do...

so anyway. all i need is two populations i can takes samples of and compare. it must be quantitative data (numeric) and the data must be availible somewhere i can easily get to it (like online). this is NOT an experiment, or a survey, there are no predictor and response varibles, its just two related populations. there can't be more than two populations, and there can't be more than one varible i'm looking at.

the example that was given in class was 'the number of cars parked illegally on campus betwwen 9 and 12 verses the number of cars parked illegally between 1 and 5'

i'm trying to think of a way to apply that to cars.. if there was some good data on, say, oil ananlysis for 20 different cars, 10 runing KN and 10 running stock filters, i could pick a nice descriptive stat for oil-dirtyness and just compare the two groups of cars based on that. the problem is i don't know of anyone who would do that kind of an experiment with cars.

links or just ideas would be unimaginably helpful to me right now, the deadline is comming up.

[Javier]

for example, the two groups may be 535i population in the forum, and 525i population ( or 6 cylinders and 8 cylinders engines). Also the investigated issue may be: Running performance chips, or Total miles per year age, or Manual vs automatic.

Just guessing.

You can open a thread in the forum and get the data results from there. Sure everybody will be willing to help.

Javier

[infinity5]

wow, i never thought of it that way. hmmmm. good idea. i'll try and think of an interesting variable to ask..

[Kalevera]

hmm...I did a project like this a year or two ago, only I think I had two independent vars and one dependent. My data were # of high school graduates per state and # of households with computers per state versus average income (haha....yes, per state).

If it were me, I'd want to do something that involved, for instance, the number of years that someone's owned their e34 and the average amount of money they've spent per year on the car (seems to be a hot topic recently). The problem is that you'd have to mine these kind of data - they're not, obviously, published.

One of the things you COULD do is call Consumer Reports (or equivalent) and see if they'll give you, say, reliability data (say, # of complaints per model year of a car) and then call up a local dealer or indy and see if they'll give you something like average costs for repairs on given model years. Once again, info that's not readily available on the internet.

Maybe see if the teach will let you construct a little web survey...get Ed's help to do it and I bet you'll get a few hundred responses in a day or so of it being on the forums...

best, whit

[infinity5]

those are all good suggestions lowell. i remember there being a thread not too long ago about how long you've owned your car and the repairs you've done.

the only problem is that's not two populations that share a varible i can compare. my professor really made this project a bit of a hassle. if i could break e34 owners down into two groups, then compare something those two groups might share that varies, i would have a topic!

hmmmmmmmmmmmm.
well i have to go to work soon, hopfully i'll get home tonight to a myriad of interesting and useful ideas posted so i can get this all done by wednesday.

if all else fails i'll revert to something like.. umm.. high end truck torque vs high end car torque.. hehehe.. i'll make lots of jokes about towing inordinately large things with very very tiny sports cars. (gearing aside/etc)

[Kalevera]

aaah...my bad - I probably misread that they had to share a variable. My project related to finding correlation and running all sorts of statistical analysis on vars that might have been related, and I guess I got caught up in that tangent.

best, whit