# Breush-Pagan Test

For those with schweser… on page 199, why have they used one degree of freedom on the chi-square distrbution. I understand the squared residuals have been regressed on the original coefficients, but doesnt that give us the original degrees of freedom? (apologies if my post is not clear)

BP test has k degrees of freedom - where k=number of independent variables… see page 198!!!

ahhh maketh sense. U know sometimes u stare at the notes for so long even simple stuff starts to get confusing lol… if i am getting u right, the only figure not from the original data in BP test is R^2, which is derived from the residuals regressed on the coefficients… n & k are from the original regression, which in this case was a simple linear regression???

One Wrote: ------------------------------------------------------- > if i am getting u right, the only figure not from > the original data in BP test is R^2, which is > derived from the residuals regressed on the > coefficients… > > n & k are from the original regression, which in > this case was a simple linear regression??? Regression for BP test and the Original Regression are 2 different regressions, being regressed for different items. In the original regression, we are regressing Dependent Variable against Independent Variable. Whereas in the BP test, we are regressing Variance of Error terms (got from original Regression) against the same Independent Variable as used in the original Regression. This is because in the new Regression, we are testing to see the extent to which variation in Error term can be explained by variation in the original Independent Variable. If it can be significantly explained, we have a case for Conditional Heteroskedasticity.