Comparing sample sets of data already aggregated

[#1]

Hello,

I am trying to compare the number of rejected units from 2 sample groups with unequal populations. I have read multiple times that the tool to use in JMP for that purpose is the "Fit Y by X", but it doesn't seem to take into account the population difference in my 2 groups (which I put in the Freq. box). How can I do my analysis? And how can I add more parameters to this analysis (I also want to make sure that there is no significant difference between my products)?

Here is the data I want to test:

Sample Type	Product	Quantity	Fail1 Qty.	Fail2 Qty.	Fail Tot.
Sample1	Prod1	723	7	2	9
Sample1	Prod2	7964	25	8	33
Sample1	Prod3	4651	25	14	39
Sample2	Prod1	4845	36	6	42
Sample2	Prod2	163182	637	123	760
Sample2	Prod3	80716	381	115	496

Best regards,

Samuel Taverne

[#2 · Accepted solution]

Hi STA,

The trick here is formatting your data differently, which I did by adding one formula column, then stacking the data. Then, the Fit Y by X platform should work for you. I did make an assumption though, that what you were looking for was the difference in total failures by each sample type--if you meant to include Fail 1 Qty and Fail 2 Qty, then you'll need to approach differently.

[#3]

Hi Jed,

Thanks for your answer! I also found the  "Two Sample Test for Proportion" and the "Analysis of Means for Proportions" to support my analysis once the fit Y by X was done. The data stacking was not so straight forward, thanks for mentioning how it should be organised.

Best regards,

Samuel