In: Statistics and Probability
Perform's Tukey's multiple comparison procedure in SAS (using alpha = 0.05) and summarize which insecticides are significantly different from each other, in terms of mean number of deaths.
data insectdeaths; input INSECTICIDE $ DEAD; cards; A 85 B 90 C 93 D 98 A 82 B 92 C 94 D 98 A 83 B 90 C 96 D 100 A 88 B 91 C 95 D 97 A 89 B 93 C 96 D 97 A 92 B 81 C 94 D 99 ; run;
Treat insecticides as a grouping variable using class statement in proc glm
use model statement
Ho:
he mean number of insect deaths are same across insectside types.
Ha:
The mean number of insect deaths differs across insectside types.
SAS Code:
data insectdeaths;
input INSECTICIDE $ DEAD;
cards;
A 85
B 90
C 93
D 98
A 82
B 92
C 94
D 98
A 83
B 90
C 96
D 100
A 88
B 91
C 95
D 97
A 89
B 93
C 96
D 97
A 92
B 81
C 94
D 99
;
run;
proc glm data = insectdeaths;
class INSECTICIDE ;
model DEAD = INSECTICIDE ;
means INSECTICIDE/cldiff tukey ;
run;
Output:
The GLM Procedure
| Class Level Information | ||
|---|---|---|
| Class | Levels | Values | 
| INSECTICIDE | 4 | A B C D | 
| Number of Observations Read | 24 | 
|---|---|
| Number of Observations Used | 24 | 
The GLM Procedure
Dependent Variable: DEAD
| Source | DF | Sum of Squares | Mean Square | F Value | Pr > F | 
|---|---|---|---|---|---|
| Model | 3 | 488.7916667 | 162.9305556 | 17.99 | <.0001 | 
| Error | 20 | 181.1666667 | 9.0583333 | ||
| Corrected Total | 23 | 669.9583333 | 
| R-Square | Coeff Var | Root MSE | DEAD Mean | 
|---|---|---|---|
| 0.729585 | 3.264029 | 3.009707 | 92.20833 | 
| Source | DF | Type I SS | Mean Square | F Value | Pr > F | 
|---|---|---|---|---|---|
| INSECTICIDE | 3 | 488.7916667 | 162.9305556 | 17.99 | <.0001 | 
| Source | DF | Type III SS | Mean Square | F Value | Pr > F | 
|---|---|---|---|---|---|
| INSECTICIDE | 3 | 488.7916667 | 162.9305556 | 17.99 | <.0001 | 

The GLM Procedure

The GLM Procedure
Tukey's Studentized Range (HSD) Test for DEAD
Note:This test controls the Type I experimentwise error rate.
| Alpha | 0.05 | 
|---|---|
| Error Degrees of Freedom | 20 | 
| Error Mean Square | 9.058333 | 
| Critical Value of Studentized Range | 3.95825 | 
| Minimum Significant Difference | 4.8635 | 
| Comparisons significant at the 0.05 level are indicated by ***. | ||||
|---|---|---|---|---|
| INSECTICIDE Comparison  | 
Difference Between Means  | 
Simultaneous 95% Confidence Limits | ||
| D - C | 3.500 | -1.364 | 8.364 | |
| D - B | 8.667 | 3.803 | 13.530 | *** | 
| D - A | 11.667 | 6.803 | 16.530 | *** | 
| C - D | -3.500 | -8.364 | 1.364 | |
| C - B | 5.167 | 0.303 | 10.030 | *** | 
| C - A | 8.167 | 3.303 | 13.030 | *** | 
| B - D | -8.667 | -13.530 | -3.803 | *** | 
| B - C | -5.167 | -10.030 | -0.303 | *** | 
| B - A | 3.000 | -1.864 | 7.864 | |
| A - D | -11.667 | -16.530 | -6.803 | *** | 
| A - C | -8.167 | -13.030 | -3.303 | *** | 
| A - B | -3.000 | -7.864 | 1.864 | |
since F=17.99,p<.0001,
p<0.05
Reject Ho that all the 4 type insecticides means are equal
and conclude that there is sufficient evidence at 5% level of significance that
he mean number of insect deaths differs across insectide types.
From tukey,the following means are different
D-B
D-A
C-B
C-A
B-D
B-C
A-D
A-C