How to determine assumptions

xtreme2001 · 11-06-2009, 10:06 AM

Hi everyone.

Newbie stats question.

Say I'm completing a parametric test which takes into account several assumptions (interval data and normal distribution). How do I determine these assumptions? Is it by actually making calculations (mean/median/mode) to determine the normal distrubition? If that's right, how do I determine the assumption of interval data?

The question seems odd - I know.

Thanks

vinux · 11-06-2009, 10:41 AM

If you are using any statistical package,in that there will be option for "Normality test" or PP plot. Based on the p value you can conclude distribution part.

WeeG · 11-06-2009, 12:52 PM

you should check for normality !
you can see with your eyes if it's interval or not

as for normality, there are several ways to check that.

http://en.wikipedia.org/wiki/Normality_test

I recommend that you try the Shapiro-Wilks test, it is more reliable than the others, or that you look at the Q-Q graph, it is even better. Like vinux said, every statistical package can easily do that, you just need to find how. If you name the software you work with, we'll try to tell you how to run the tests...

11-06-2009, 10:06 AM	#1
xtreme2001 Junior Member Join Date: Nov 2009 Posts: 4 Thanks: 2 Thanked 0 Times in 0 Posts	How to determine assumptions Hi everyone. Newbie stats question. Say I'm completing a parametric test which takes into account several assumptions (interval data and normal distribution). How do I determine these assumptions? Is it by actually making calculations (mean/median/mode) to determine the normal distrubition? If that's right, how do I determine the assumption of interval data? The question seems odd - I know. Thanks

Thread Tools	Search this Thread
Show Printable Version Email this Page	Search this Thread: Advanced Search
Display Modes
Linear Mode Switch to Hybrid Mode Switch to Threaded Mode

11-06-2009, 12:52 PM	#3
WeeG Senior Member Join Date: Mar 2006 Location: Israel Posts: 140 Thanks: 18 Thanked 5 Times in 5 Posts	you should check for normality ! you can see with your eyes if it's interval or not as for normality, there are several ways to check that. http://en.wikipedia.org/wiki/Normality_test I recommend that you try the Shapiro-Wilks test, it is more reliable than the others, or that you look at the Q-Q graph, it is even better. Like vinux said, every statistical package can easily do that, you just need to find how. If you name the software you work with, we'll try to tell you how to run the tests...