Here's a statistical question that I think you may have some interesting opinions on. Let's say you've created a measure of anxiety with 30 items on...

Here's a statistical question that I think you may have some interesting opinions on. 

Let's say you've created a measure of anxiety with 30 items on it.  You've collected some data and your ready to analyze how well your items performed.  One analysis you might do is to see how strongly your items correlate with each other.  What would you do if you had three items that had extremely high correlations with one another (e.g., r=.98)?  Would you want to keep all these items because of their strong relationship or would you want to get rid of some or all of them?  Why?