Monday, December 12, 2011

In statistics, how do you control for a variable?

If I have 4 variables, 3 of which influence the fourth, how do I control for a variable and determine its effect on the fourth variable?|||Let's say you have four (random) variables x, y, z, and w. suppose there was a relationship beween w and y, for example, say w=ay +b. (y could represent a companies cost of raw meat and w the the price of that meat cooked to order.) That formula then says that if we know the price of the raw meat then we know the price that establishment will charge for the cooked order. So the charge of the cooked order controlled on the cost of the raw meat, w|y, verbalized as "w given y" or "w controled on y" is then w|y=ay+b. In statistics we very often control on certain variables where there is a known relationship and compute (integrate or find the expected value) that relationship across the unknow variables.

No comments:

Post a Comment