For hypothesis, we are told to put whatever we need test as an alternative hypothesis. The statement that needs to be rejected as a Null. However in case of correlation, we are trying to find out correlation =0 and inspite of that put it in NULL. Should nt it be in alternate?

We want to reject that the correlation coefficient is equal to zero. Only by doing so can we embrace the alternate that the correlation is statistically significant.

You are trying to reject the idea that r = 0. You can’t assume that two variables are related until you prove they are.

chebychev Wrote: ------------------------------------------------------- > We want to reject that the correlation coefficient > is equal to zero. Only by doing so can we embrace > the alternate that the correlation is > statistically significant. Wow. First we have the “absence of zero” and now “embrace the alternate”. I think chebychev is a pastor or something on the side.

Highly impressed thus far with chebychev’s posts! Lucid and accurate - even poetic. Makes a change from mr grumpy ^^^^^ [/kidding] Am I missing a joke with the misspelt name though?

Have I ever been grumpy with you? I don’t think so…

Not particularly. I love you really.