Why are there to different formula’s for finding F-statistic?
-
F = (RSS/k)/(SEE/(n-(K+1))
-
F = (RSS/1)/(SEE/(n-2))
Why are there to different formula’s for finding F-statistic?
F = (RSS/k)/(SEE/(n-(K+1))
F = (RSS/1)/(SEE/(n-2))
I think the latter formula is only applied to single regressions. To be safe, i would always just use the former.
Yup. k = 1; one slope coefficient, one independent variable.
At least that way you have to memorize only one formula, not two. Saves some brain space.
both formulas are same. its basically
F = (RSS/ no. of independent variables)/SSE (total number of observations - total number of parameters)
when we are using one independent variable it takes the form F = (RSS/1)/(SSE/(n-2)
when we are using more than one independet variables it takes the form F = (RSS/k)/(SSE/(n-(k+1))
also their is a difference between SSE (sum of squared error) and SEE (standard error of estimate). In this formula we are using SSE not SEE.