Using Correlation

[bradyle]

Hi!

I am doing tests to see if the density of a piece correlates to the % pick up of resin the piece gets as it goes through an impregnation process. I am planning to use Pearson's correlation coefficient for this test.
I also want to do a correlation test on the density of the piece and whether or not it passes an air pressure leak test. The result of the leak test will just be yes or no. I don't think I can use Pearsons correlation in this case, Can I?
Is there any other tests you could suggest?

Thanks!

 

[GregLocock]

I don't think you can, There again i suspect you are tesing the hypothesis 'if density > x then leak test passes'

which strikes me as a probabilty test.

alternatively google for statistics of attribute data



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[IRstuff]

?? You claim you're doing a go-no-go test, so what are you correlating? Do you have multiple pieces?

TTFN
faq731-376

 

[PFF]

There are several ways to calculate the two category correlation. You could break the density into several categories then do a chi square to measure strength of relationship. Or you could use a K-category correlation. or you could cook up your own measure if you know the distributions of the two variables.