Introduction to SAS: Project 8

Project 8

Here basic features of proc glm are studied. The focus is on the basics of the analysis of variance.

The iris data, contained in iris.dat, will again be used. In Project 4 a comparison of the means of the sepal lengths measurements was made on a pairwise basis.

A more powerful method of analysis is to test simultaneously for the equality of the 3 means. This is done using a one way analysis of variance.

ASSIGNMENT:

1. In order to do a one way ANOVA using proc glm it is necessary to have a variable to be used in the class statement. In this context the variable should have 3 values depending on the species from which the observation is taken. Create a data set containing the sepal length measurements and an appropriate new variable. The techniques of do loops will be helpful. You don't need to turn in a copy of the data set but do turn in a copy of the program used to create the data set.

2. Use proc glm to test for equality of the means for the 3 species. Also compute Bonferroni type simultaneous confidence intervals for the comparison of the 3 means at both the 90% and 99% confidence level. Turn in a nice copy of this printout.

3. Briefly, are the 3 mean lengths different? Which pairs are different at the 90% level? At the 99% level?

4. Create a SAS data set from the file pig.dat as mentioned in the proc glm discussion. The missover option of the infile statement may be useful. You don't need to turn in a copy of the data set. Then test for the presence of litter effect. Is there litter effect at the 5% level of significance? Submit a nice printout to back up your conclusions.

Contents Index <<Browse<< >>Browse>>