Assignment:
For the fat data used in this chapter, a smaller model using only age, weight, height and abdom was proposed on the grounds that these predictors are either known by the individual or easily measured.
(a) Compare this model to the full thirteen-predictor model. Is it justifiable to use the smaller model?
(b) Compute a 95% prediction interval for median predictor values and compare to the results to the interval for the full model. Do the intervals differ by a practically important amount?
(c) For the smaller model, examine all the observations from case numbers 25 to 50. Which two observations seem particularly anomalous?
(d) Re-compute the 95% prediction interval for median predictor values after these two anomalous cases have been excluded from the data. Did this make much difference to the outcome?
Your answer must be typed, double-spaced, Times New Roman font (size 12), one-inch margins on all sides, APA format and also include references.