Analyzing regression based on different levels of a variable

[rbbtchnnl]

Hello all,

I need to compare effects on a dependent variable based on different levels of an independent one, x2 in the attached equation. I want to look at the effect on predicted y when x2 changes from 4 to 7



I'm using eviews 7 and what I did was I sampled x2 based on the if condition x2=4 and ran the regression without the x2 term (as eviews told me there was a near singular matrix error and I assume that it is set to 4 anyway). So I get coefficients BUT what does the equation look like now? I assume that x2 is still part of the model, only constant, but what is its coefficient of x2? Or do I just omit x2 and say y = B1*x1 + B3*x3 when x2=4?

Now I do the same thing for x2=7. Do I interpret the change in the coefficients of x1 and x3 as the effect of x2 (as it moves from 4 to 7) on y?

Thanks!

[eyas]

You can substitute x=2 and find y and subst x=4 and find also y and find the diff bet the first y and the second on as apercent or number , i think this will solve the change in y after the change in x
Thanx

[jason_ll]

Hello all,

I need to compare effects on a dependent variable based on different levels of an independent one, x2 in the attached equation. I want to look at the effect on predicted y when x2 changes from 4 to 7



I'm using eviews 7 and what I did was I sampled x2 based on the if condition x2=4 and ran the regression without the x2 term (as eviews told me there was a near singular matrix error and I assume that it is set to 4 anyway). So I get coefficients BUT what does the equation look like now? I assume that x2 is still part of the model, only constant, but what is its coefficient of x2? Or do I just omit x2 and say y = B1*x1 + B3*x3 when x2=4?

Now I do the same thing for x2=7. Do I interpret the change in the coefficients of x1 and x3 as the effect of x2 (as it moves from 4 to 7) on y?

Thanks!

The difference is simply 7*B2 - 4*B2.

In other words, 3*B2.