0 thoughts on “”

  1. Avatar

    Interesting article Mr. Quinn. One of the problems of Chi-square and Chi-square based measures is that they are effected by n and by table size. As a result, when using Chi-square would you not apply some sort of control or alternate test to try and deal with that? In this case, I may look at Phi or Cramer’s. In other cases one could use a correction such as Bonferroni (realizing that can be conservative in cases where there are large numbers of tests, or positive correlation of the test statistics). I suppose the other way to look at this is to take an ecological significance approach; you are getting 78 more responses with an extended subscription and is that good enough for your business objectives? I have enjoyed reading your columns; keep it up!

  2. Avatar

    Thank you for posting a good article on correlation!

    So, from the above article – considering the case where correlation between the variables is less(0.43) due to some external factors, how do you improve the correlation between two variables to predict the response variable with good R-square value i mean with good accuracy?

Leave a Comment

Your email address will not be published. Required fields are marked *

Scroll to Top