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 1 down vote favorite

I have a problem with my mixed effects model where I found that both subjective evaluations of my stimuli (valence - positive or negative and self-relevance - highly self-relevant or not self-relevant) are highly correlated with one another. Theoretically this is not too surprising but poses a problem in my model. Thus can I take out one of these two variables in my model? When I run the model on solely one of the two variables, I obtain more or less the same t-values of my significant interactions. I also have age as a categorical variable and two dimensional continuous variables in my model. My dependent variable is response times.

share|improve this question
Could you clarify what problem this is posing? – onestop Jan 25 '11 at 9:52
Do you suspect multicollinearity or something? Are your parameter estimates not stable when you include/exclude additional observations? Could you suggest latent variable for the combination of two valence and self-relevance? – Dmitrij Celov Jan 25 '11 at 12:44
Actually, there is multicollinearity between the two variables. I'm not sure how I would be able to construct a latent variable crossing Valence and Self-relevance especially when I did a correlation, it was significant at r=.755. Thanks for any input! – Joanna Jan 26 '11 at 8:44
When you say "two-dimensional," "interactions," and even "mixed effects," I'm having trouble telling whether you're using the jargon in standard ways or more impressionistically. This makes it hard to give you a helpful response. – rolando2 Jan 29 '11 at 20:57
Why do you think you need to take out one of the two variables? Do you perhaps fear that if they are highly correlated, you won't be able to study their independent effect as they maybe would average each other out? I amactually wondering the same thing and posted my question here: stats.stackexchange.com/questions/26402/… Do you know the answer by now? – user9203 Apr 20 '12 at 6:35

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.