Tell me more ×
Cross Validated is a question and answer site for statisticians, data analysts, data miners and data visualization experts. It's 100% free, no registration required.
up vote 0 down vote favorite

i got a problem with cross-validation!

I got two studies, t(N=1200) and t+1(N=200) respectively. The 200 corporations form t+1 are also part of the t0-Data. In t+1, there are two significant relationships using kendall's tau, but they are unexpected relationships. In order to check for errors, i want to do a cross validation using the whole data from t. But there's a problem: I can remove only 165 (t+1)-corporations from t, because 35 corporations cannot be identified.

Thus, i cannot postulate independence between results and cross-validation.

Is there a way around this?

Thanks in advance!

share|improve this question
For clarification - are the two significant relationships also significant in study t? If so, given that t appears to be a superset of t+1, why would you pay attention to the t+1 result? – jbowman Mar 31 '12 at 16:35
because t+1 contains a scale measure, which has been constructed based on the experience drawn from t, and t does not. – Jack Shade Mar 31 '12 at 16:40
Using only 165 corporations the correlation gives non-significant results, but using 200 yields significant ones! – Jack Shade Mar 31 '12 at 16:43

Know someone who can answer? Share a link to this question via email, Google+, Twitter, or Facebook.

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Browse other questions tagged or ask your own question.