Statistics – how to test in Custom Tables?
I got a question from a user: How do I use the testings within the Custom Tables command? Which test should I use? (Custom Tables is an add on module, for user that makes a lot of reports/tables).
I will do a crosstable between 2 category variables and my question is if the gender distribution differs between the different job categories? And if yes, where?
(Click on the Picture below to see the bigger view).
There is 2 steps that a researcher has to follow in the right order to test a frequency cross table:
- Do a chi2-test (as it is a frequency cross table)
- If the result is significant, then do proportion-test (=z-test/p-test)
So I start with the chi2-test in the Custom Tables command, and you find the test when you click on the tab: “Test Statistics”. Fill in as below:
The result is significant as you see below (significance value is far under 0.05, see “Sig.”)
So then as step 2, I Wold like to test where the differences are, and you can test the column distributions among each other. In my table I have 7 columns to compare to each other 1 against 2, 1 against 3, 1 against 4 and so on.
Then you choose the proportions-test from the same tab you used Before:
In this example I think the first choice (in separate table) is easier to understand:
I have concentrated on the first row (Male), and you can see which job categories differs from each other below:
So column A, B, C, E and F all differs from column G (“Econom”). While column D differs from all other categories but not column A.
Thanks for watching!