0

I want to check whether a substitution effect among two variables exists.

In the models that I have estimated these have been set as independent variables. Each model has 1 "main" dependent variable and a series of control ones.

I want to perform a t-test to compare the coefficient of each model against the other. This way, based on the p-value, I can assume that I'm not talking apples and oranges, and that even if I cant calculate the magnitude of this substitution effect, I can at least prove it exists.

The problem is that I don't know how to perform the t.test in R on the coefficients.

lmo
  • 37,904
  • 9
  • 56
  • 69
minigui
  • 11
  • 1
  • What (I think) you are describing is statistically suspect (but possibly interesting). You should post some actual model-objects using the `dput` function to create an unambiguous coding basis for testing and demonstration. – IRTFM Feb 08 '17 at 00:17
  • @42- So these are my regression models: https://www.dropbox.com/sh/e9tz37ntwymvt4b/AABlux-Q948vkVqTG-QaDHIDa?dl=0; as you can see they have the same dependent variables.. I want to perfomr a t-test to check if the variable named `ifrs` affects both independent variables – minigui Feb 08 '17 at 00:35

0 Answers0