Duroc-Jersey Pigs Data
|
No. Males |
Frequency |
|
0
|
3
|
|
1
|
16
|
|
2
|
53
|
|
3
|
78
|
|
4
|
53
|
|
5
|
18
|
|
6
|
0
|
If you have data values from a continuous distribution, you can just put the data values in a column of
StatXact's CaseData editor, and then click that variable in as the Response when running a one-sample K-S test.
But with data like this you don't want to enter in three 0s, sixteen 1s, 53 2s, etc. It's easier to put the numbers
in the table into two columns of the CaseData editor, and then click one in as the Response and click the other in
as the Frequency when running a one-sample K-S test.
To do the desired K-S test, use Nonparametrics > One-Sample Goodness-of-Fit > Kolmogorov ..., click the proper variables into the
Response and Frequency boxes, under Distribution select Binomial for Type, and enter 0.5 and 6 into the Probability
and Trials boxes. Then under Compute, select Exact Using Monte Carlo (having previously set the Monte Carlo options to do
1000000 trails with a seed of 23456), and finally click OK.
To do the desired Person's chi-square test, recall that you have to put the data into the Table Data Editor (instead of the Case Data Editor),
choosing a 1 by 7 table with column scores. (You can enter the binomial distribution probabilities for the outcomes 0, 1, 2, ..., 6 as the scores.) Then use
Nonparametrics > One-Sample Goodness-of-Fit > Chi-Square ..., being sure to select Exact under Compute.