1) Using one of the four sets of means from Homework 2, part 1, use 1000 iterations to simulate the values of the mean-squared error (MSE) from the ANOVA analysis. Do this two ways: one using the OUTSTAT command in PROC GLM, and the second way using ODS statements. Run PROC UNIVARIATE with the PLOT option for each of the two simulations, and email me your two sets of code and the PROC UNVARIATE results ONLY. Hand in the code and UNIVARIATE printout.
2) Each of the three variables in this data set is highly skewed and needs a log transformation. However, 0 values occur in the data. One often-used approach in this case is to add a small constant to each data point before taking the log value. Use ARRAY statements along with other DATA step statements and/or PROC's to create new variables which are equal to the log of (original value + (minimum of variable)/4 ) for each of these three variables.
3) Below is a SAS program to do a regression simulation and part of its output. Something is wrong with the output - report what is wrong with the output, and what mistake in the program causes the problem. Although you can enter the code yourself to find or verify the answer, try to do it first just by looking at this information.
data regsim ;
do i = 1 to 100 ;
x = 15 + 2*rannor(0) ;
y = 7 + 3*x +x2 + 8*rannor(0) ;
x2 = x**2 ;
x3 = x**3 ;
x4 = x**4 ;
output ;
end ;
proc reg ;
model y = x ;
model y = x x2 ;
model y = x x2 x3 ;
model y = x x2 x3 x4 ;
run ;
The REG Procedure
Model: MODEL2
Dependent Variable: y
Analysis of Variance
Sum of Mean
Source DF Squares Square F Value Pr > F
Model 2 9852.15067 4926.07533 1.33 0.2681
Error 96 354354 3691.19013
Corrected Total 98 364206
Root MSE 60.75517 R-Square 0.0271
Dependent Mean 274.38425 Adj R-Sq 0.0068
Coeff Var 22.14237
Parameter Estimates
Parameter Standard
Variable DF Estimate Error t Value Pr > |t|
Intercept 1 136.61230 221.94699 0.62 0.5397
x 1 13.78691 29.87413 0.46 0.6455
x2 1 -0.29757 0.99843 -0.30 0.7663