A researcher asked an experienced statistician what procedure should be used to obtain the correlation between two normally distributed variables that were artificially dichotomized. Why did the researcher suddenly rush from the statistician's office and run straight to the pharmacy to buy a bottle of carbon tet cleaning fluid? The statistician told him a TETRACHORIC SOLUTION was appropriate for his problem!!!
*If you don't get this joke don't despair. The tetrachoric coefficient is legitimate but is rarely used in modern practice. I happen to own a crumpled original monograph by Thurstone that presents a table for computing this index. Any bidders out there before I auction the item on Ebay?
News 
Market Data 
Discover 
