How would I constrain a regression model with two covariates so that the coefficient B2 is equal to 1/2B1?

41 Views Asked by At

If the original model is y = B0 + B1X1 + B2X2

How could I manipulate the equation or the data so that the coefficient B2 is equal to 1/2 B1?

y = B0 + B1X1 + 1/2B1X2

In theory, how could I do this with a real dataset?

1

There are 1 best solutions below

2
On BEST ANSWER

Rewrite your model as $y = B_0 + B_1 (X_1 + \frac{1}{2} X_2)$, introduce a new variable $X_3 = X_1 + \frac{1}{2} X_2$, and perform the regression on $y = B_0 + B_1 X_3$.