Dunnett Method

Mice Example

A research psychologist wishes to investigate the difference in maze test scores for a strain of laboratory mice trained under different laboratory conditions. The experiment is conducted using 18 randomly selected mice of this strain, with six receiving no training at all (control group), six trained under condition 1, and six trained under condition2. Then each of the mice is given a test score between 0 and 100, depending on its performance in a test maze.

Program Listing

OPTIONS LS=80 NODATE NONUMBER NOCENTER;

DATA mice (DROP=I);
 DO group = 'C', '1', '2';
  DO I = 1 TO 6;
   INPUT score @@;
   OUTPUT;
  END;
 END;
 DATALINES;
 58 32 59 64 55 49    73 70 68 71 60 62    53 74 72 62 58 61
;

TITLE1 'One-Way ANOVA Model';
PROC GLM DATA=mice;
 CLASS group;
 MODEL score = group;
RUN;
TITLE2 'Dunnett Method';
 MEANS group/HOVTEST=BF CLDIFF DUNNETT('C');
 LSMEANS group/ADJUST=DUNNETT PDIFF=CONTROL('C');
RUN;
QUIT;

Output Listing

One-Way ANOVA Model

The GLM Procedure
Dependent Variable: score
                                                                    
                             Sum of
Source             DF       Squares   Mean Square   F Value   Pr > F
Model               2    673.000000    336.500000      4.55   0.0286
Error              15   1109.500000     73.966667                   
Corrected Total    17   1782.500000                                 


R-Square     Coeff Var      Root MSE    score Mean
0.377560      14.06058      8.600388      61.16667

Source       DF      Type I SS    Mean Square   F Value   Pr > F
group         2    673.0000000    336.5000000      4.55   0.0286

Source       DF    Type III SS    Mean Square   F Value   Pr > F
group         2    673.0000000    336.5000000      4.55   0.0286

One-Way ANOVA Model
Dunnett Method

The GLM Procedure
 Brown and Forsythe's Test for Homogeneity of score Variance
       ANOVA of Absolute Deviations from Group Medians
                                                             
                      Sum of        Mean
Source        DF     Squares      Square    F Value    Pr > F
group          2     37.0000     18.5000       0.48    0.6277
Error         15       577.5     38.5000                     

One-Way ANOVA Model
Dunnett Method

The GLM Procedure
Dunnett's t Tests for score

NOTE: This test controls the Type I experimentwise error for comparisons of all
treatments against a control.

Alpha                              0.05
Error Degrees of Freedom             15
Error Mean Square              73.96667
Critical Value of Dunnett's t   2.43926 ⇒ d.05;2,15
Minimum Significant Difference   12.112 ⇒ d.05;2,15 × std.err

Comparisons significant at the 0.05 level are indicated by ***.
                                             
              Difference
  group          Between     Simultaneous 95%
Comparison         Means    Confidence Limits
  1 - C           14.500       2.388   26.612  ***
  2 - C           10.500      -1.612   22.612

One-Way ANOVA Model
Dunnett Method

The GLM Procedure
Least Squares Means
Adjustment for Multiple Comparisons: Dunnett
                                   
                         H0:LSMean=
                           Control
group    score LSMEAN      Pr > |t|
1          67.3333333        0.0195
2          63.3333333        0.0916
C          52.8333333