On Tue, 10 Feb 2004 20:55:17 +0100, "sali" <[EMAIL PROTECTED]> wrote:
> Hi. > I have 8 lines created from x,y scatters, I would like to test whether the > slope is significantly different from line to line. How do I do this? (I use > STATA, exel and sigma) If you have 8 lines, then you have 8 slopes. I assume that you refer to the slopes of prediction y from x. Do you want to use the internal measure of consistency, to test whether there are different intercepts, as well? Or, can you make a statement beforehand about where the intercept is supposed to lie? If x predicts y, then you will get a certain R-squared for predicting y from x. Now, you will get an increased R-squared if you predict y from the set that has x and also has a 7 dummy variables: one for each extra degree of freedom of slope. Then you test the increase in R-squared. There could also be dummy variables for additional intercepts, if those are possible, too. Hope this wasn't too terse. -- Rich Ulrich, [EMAIL PROTECTED] http://www.pitt.edu/~wpilib/index.html "Taxes are the price we pay for civilization." . . ================================================================= Instructions for joining and leaving this list, remarks about the problem of INAPPROPRIATE MESSAGES, and archives are available at: . http://jse.stat.ncsu.edu/ . =================================================================
