Randomization on Minitab
The following is easier than the steps in Section 3.2:
1. Start as in Step 1 on p. 34: Enter r1 1's. r2 2's,
etc. into a column. You can do this by hand. If your design is balanced,
you can use Set Patterned Data (Calc menu). For example, if you have three
groups of 4 each, specify to start at 1, end at 3, with increment 1, and
then enter 4 in "Repeat each value".
2. To randomize, choose Random Data from the Calc menu, then Sample
from Columns. Enter the total number of observations desired (n = sum of
the ri's)
in "Number of rows samples." Enter the column you just
created in the box "From columns" (Note that columns are labeled
c1, c2, etc.unless you have given them a label. You can see the columns
with entries by clicking in te box. You can then select and click in
the box to enter a particular column.). Enter the column where you want
the randomized numbers in "Store
samples in." Example output (using 2,2, and 3 as treatment sizes):
| obs |
treat |
randtr |
| 1 |
1 |
1 |
| 2 |
1 |
3 |
| 3 |
2 |
1 |
| 4 |
2 |
3 |
| 5 |
3 |
3 |
| 6 |
3 |
2 |
| 7 |
3 |
2 |
(I have added a column "obs" for convenience.) This says assign experimental
units 1 and 3 (first column) to treatment 1 (last column), units 6 and 7
to treatment 2, and units 2, 4, and 5 to treatment 3. (Note that the second
column was just an intermediate step.)