Post Hoc Tests - Agricultural Statistical Data Analysis Using Stata - page 164

squared results in the error mean square or variance. This value along
with the number of replications can be used to calculate the standard
error of the mean difference. The first loop uses the forvalues loop,
which loops from 2 to the number of treatments and is used to calculate
the Studentized Range value with the qsturng command. This then
is used to calculate the comparison values based on the rank difference.
With the duncan.do and watertrial2007frtchar2.dta files loaded,
make sure the duncan.do file has been executed, which can be done
from the Do-File Editor window by clicking the Run or Do icons in
the upper right corner on a Macintosh or the same icons in the icon
bar on a Windows computer (see Chapter 7, Figure 7.1). Once this is
done, enter the following command:
duncan length trt rep
This will result in the following output:
Number of obs = 51 R-squared = 0.9276
Root MSE = .743141 Adj R-squared = 0.8708
Source | Partial SS df MS F Prob > F
----------+----------------------------------------------------
Model | 198.193996 22 9.008818 16.31 0.0000
|
trt | 195.604525 19 10.294975 18.64 0.0000
rep | 2.25708993 3 .752363311 1.36 0.2746
|
Residual | 15.4632584 28 .552259229
----------+----------------------------------------------------
Total | 213.657254 50 4.27314509
Duncan's Multiple Range Test (P≤0.05)
17.37 a
12.56 b
12.12 bc
12.05 bc
11.88 bcd
11.62 bcde
11.38 bcde
11.28 cde
11.19 cdef
11.19 cdef
11.15 cdef
11.10 cdef
11.00 cdefg
10.75 defg
10.73 defg