I am trying to track down a formula that will accept degrees of freedom
and the computed value of a Pearson r and output the p value for the
statistic.

If you're testing the hypothesis that the population correlation is zero, you can convert your r to a t-statistic using an expression such as this in Excel:


t = (r * sqrt(n - 2)) / (sqrt(1 - r^2))

...and then use Excel's TDIST function with n - 2 df to return the p-value.

If you're testing another hypothesis (such as r being equal to some nonzero value in the population), the sampling distribution of r can get pretty skewed, so you're safer going with a Fisher's Z transformation of your r and using that to test the hypothesis.

Hope this helps...



Charles Perdue
Psychology Department
WV State College
[EMAIL PROTECTED]
304.766.3271


--- You are currently subscribed to tips as: [EMAIL PROTECTED] To unsubscribe send a blank email to [EMAIL PROTECTED]

Reply via email to