When we are testing the eqaulity of two coefficient in the model .
let's say B2=B3
H_A:B2 not equal B3
Assume using F-test
Why should we force B2 and B3 to be equal by summing up variable X2 and X3 .
Then do regression .
Why don't we multiply X2 and X3 instead???And run a regression ,then use it as our restricted model??
Advertise on Talk Stats