confidence interval/f-tests

#1
Basically Fe (iron) is the x variable/predictor, and MDD loss is the y variable/response in minitab.





1. I generated some data (regressoin analysis). How do i find a standard error for the estimated regression coefficient? (with or without minitab, either is fine)



2. I generated the anova table (see below), but how do i find the estimate of sigma^2?



3. How do i do an F-test and t-test in minitab or outside minitab (by hand)?



4. I calculated the 95% confidence interval using the formula given in my notes (see below), but how do i do it for the 4th value of the explanatory variable?



Many thanks.



 

Mean Joe

TS Contributor
#2
1. The standard error for the estimated regression coefficient is in the "SE Coef" column for Fe.
How did you calculate the 95% confidence interval for beta? The directions in that box tell you "betahat is the value in the Coef column for Fe and s.e.(betahat) is the value in the SE coef colun for Fe."

2. I can't remember, oops.

3. The ANOVA table gives you the F-test results: F=352.27 (calculated by dividing MS values Regression/Regression Error: 3293.8/9.4=352.27) The p-value corresponding to this F is right next to it: p=0.000 (report as p<0.001).
The t-test for beta can be read from the first table (beta is the general name for your specific Fe): t=-18.77.
Note that t^2 = F: (-18.77)^2 = 352.27

4. You need to use the formula for prediction intervals--they are wider than confidence intervals. Throw in the value x4=0.95