400 Chapter 12 Multiple Comparisons Among Treatment Means
Data Siegel;
Infile ‘Alexander:SAS610:Data Files:Siegel.dat’;
Input Condtion Latency;
run;
Proc GLM Data = Siegel;
Class Condtion;
Model Latency = Condtion/SS3;
Means Condtion /SNK Tukey REGWQ Scheffe;
Run;
-------------------------------------------------------------------------------------------------------------------------
The SAS System 16:51 Thursday
December 20, 2007
The GLM Procedure
Dependent Variable: LATENCY
Sum of Mean
Source DF Squares Square F Value Pr > F
Model 4 3497.600000 874.400000 27.33 <0.0001
Error 35 1120.000000 32.000000
Corrected Total 39 4617.600000
R-Square Coeff Var Root MSE LATENCY Mean
^2
0.757450 36.26189 5.656854 15.60000
Dependent Variable: LATENCY
Source DF Type III SS Mean Square F Value Pr > F
CONDTION 4 3497.600000 874.400000 27.33 0.0001
Ffor
-------------------------------------------------------------------------------------------------------------------------
Condition
Student-Newman-KeulsTest for LATENCY
NOTE: This test controls the Type I experimentwise error rate under the
complete null hypothesis but not under partial null hypotheses.
Alpha = 0.05 df = 35 MSE = 32
Number of Means 2 3 4 5
Critical Range 5.7420598 6.921941 7.6279952 8.1319061 w 5 =q 5
Means with the same letter are not significantly different.
SNK Grouping Mean N CONDTION
A 29.000 8 Mc-M
A 24.000 8 S-M
B 11.000 8 S-S
B 10.000 8 M-M
C 4.000 8 M-S
-------------------------------------------------------------------------------------------------------------------------
Ryan-Einot-Gabriel-WelschMultiple Range Test LATENCY
NOTE: This test controls the Type I experimentwise error rate.
Alpha = 0.05 df = 35 MSE = 32
Number of Means 2 3 4 5
Larger Critical Range 6.8765473 7.5391917 7.6279952 8.1319061
than for SNK Same as SNKExhibit 12.1 (continues)