A data analyst wants to fit a complete second order model using two quantitative variables. The full model was fit to a data set of 90 observations. The results showed that TSS = 130 and SSE = 50. When a first order model was fit SSE was found to be 90. Calculate the sample partial F statistic for the comparison of the first order and complete second order models.
Partial F test:
(SSE Reduced - SSE Full) / g , where g is the # of parameters set to
________________________ zero in the null hypothesis
SSE Full / (n-k-1)
My questions are:
How do you know that k is equal to 5?
How do you know that g is equal to 3?
How do you distinguish which is SSE Reduced and which is SSE Full?