Ok guys I have a question I would like every one's help with. I am taking statistics this summer and our teacher has given us a chance to earn some extra credit (which I need lol). I have read some post from some pretty intelligent people even some with Engineering degrees regardless I need help. Here is the problem I am having: I just need n>=30 variables. I want to do something over cars (hp, time slips etc) Can you guys help me out? Thanks

Extra credit project: this is the one we talked about in class. Think of a variable, a dependent variable (y), that may be changing in response to changes in another variable, the independent variable (x). Email, or speak to me about your idea for clearance. Then, try and collect data (shoot for 30 obs. for each variable). Show me what you have in an excel file. Then, write a short paragraph describing what you plan to do….just tell me the story. After we cover Regression next week, you’ll use SAS to run the model for your data. You’ll then submit your paragraph and final SAS output to me. Worth +10 points on the Final Exam (due on the last day of classes, Tuesday, July 29th by 5pm).

Its like a cause and effect weather temps vs e.ts. I just need a sample data of 30 or close to it...Give me some data! lol Thanks everyone so much!

 

OOOOOFFFF.........TTTTOOOOOPPPPIIICCCC!!!!

 

It's GERNAL DISSCUSSION STEVE! Thanks for the help. I just came seeking help from friends I have met on the board and people who might be interested in helping me out giving it could retain to the stats of certain events dealing with our cars. I am sorry if you are not interested. I know there is a few who might be and might be tired of answering "how to remove the air silencer" for the 1,000,000x.So for any who are willing to help I would really appreciate it. Thanks

 

actually i just didn't understand anything you just said, i have a history degree...i leave the math to my fiance the accountant... let me know if you wanna know how to spell something...

 

hahahah, good luck with doing one on cars, your going to have to seriously find a single car, basically a race car, that will run 30 different runs in like 10 minutes, because, otherwise, you are going to have too many variables. Even then, there are going to be a bunch of different variables.

You could do a dyno test on a car, and go through 30 different degree's total timing, and eventually you'll get to a point where each degree of spark plug advance will increase HP an average ammount. It'll never increase at perfect incriments, but you can get a general number.

Just make sure whatever you do, you repeat the test enough, and hopefully, I would think that hopefully, you know how to do a histogram, at which point you should have data that meets a perfect bell curve, that is, when you do an average, 67% of your data resides within 0 standard deviations, 95% should lie within 1 standard deviation, and 2 standard deviations 99.7% of your data should be there.

if you need me to explain it more, just shoot me a PM, I took alot of these courses in school, so I have a little bit of a backround in it (BS in Engineering Mathematics - concentration in physics)

Edit: you could probably fake something as well, if you know how to fake something well enough, it's better than doing the experiment anyway, because that way you know the in's and out's of whatever your faking, and that's the whole point right (well not the faking thing).

If you do a histogram, kind of like a bar chart, where the ammount of times something occurs on the Y axis, and the group it falls into on the x axis. Your data is seperated into "bins" of equal size, and for each bin, you chart how many times it occurs.

say for instance you are doing it on a drag car (you actually could do the whole project on an open track car, and do lap times versus changes in weather) your going to chart how many times there is a change in performance versus a change in your baseline temperature. You will chart how many times a + or - 1 degree temperature made an impact (what you really want is a car that is very consistant, and you have to establish a set limit to how much of an impact your going to record, say you will only call something dramatic if there is a .1sec (just an example) in the ET) on the lap time or whatever your doing. then do the same thing for + 2 degrees and -2 degrees, and so on and so forth.

as long as the data set follows a bell curve (just look it up, I don't feel like typing it up right now) you can pretty much make anything you want to up. You can make up a data set that follows the curve, say you did the experiment, and it took you alot of time. Instead of using 30 variances in the variable, do like 10 occurances for each variable

 

E=mc2

did that help? btw the 2 is a squared idk how to do it on my keyboard.