Friday, May 7, 2010

F-statistic test (F-test)

  • A test for the overall significance of the regression, which is formulated with the null hypothesis that all slopes (independent variables' coefficients) simultaneously equal zero.
  • This null hypothesis is identical to a test that the R-square = 0.

F = (explained variance) / (unexplained variance)
= mean regression sum of squares / mean squared error
= (regression sum of squares/number of independent variables) / (error sum of squares/degrees of freedom;error)

= (RSS/k) / (ESS/(n-k-1))

0 comments:

Post a Comment