Hints and/or comments about Problem 14

• For the 10 point estimates, I basically just want the estimates in the boxes (rounded to the nearest tenth) --- you don't have to turn in any work to show how you got your answers (but you can turn in some supplementary work if you wish).
• You should show some work/partial results for the confidence interval parts. Don't just give the answers for those parts. But do put the answers in the boxes on the same sheet that you put your point estimates on, and attach that sheet as the 1st sheet of your solutions. (You should show how you used a Satterthwaite approximation to get the degrees of freedom. For the jackknife interval, just give some partial results, like the midpoint of the interval, and the standard error estimate that goes into the plus/minus piece.) As with the point estimates, round final answers given in answers boxes to nearest tenth.
• For the 2nd confidence interval that is requested (the one using the Stein estimators), you are still supposed to use the approximate pivot given at the bottom of p. E31 of the class notes, and you are still supposed to use the df formula given on p. E31. The only possible difference between the two Satterthwaite confidence intervals would due to the fact that for some data sets one may get different values for the df when the two sets of estimates of the variances are plugged in. (I fear that some of you may replace (MSA - MSE)/n in the pivot by the Stein estimator, but if that is done, one would also have to work through a new Satterthwaite approximation to determine the proper df, and that would be very tough, due to the complicated nature of the Stein estimator. So for both Satterthwaite intervals, use the pivot based on MSA - MSE, but for the 2nd of the two requested, use the Stein estimators to get the variance estimates required for the df calculation.)