Would like Categorical Means Comparision to use Kish correction.When there is overlap the cell comparisons in Categorical use a correction. We would like the Kish correction to be use when comparing means in the categorical platform when there is overlap.
Re: Would like Categorical Means Comparision to use Kish correction. - Status changed to: DeliveredRe: Would like Categorical Means Comparision to use Kish correction.<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />This has been delivered in JMP 13.
Re: Would like Categorical Means Comparision to use Kish correction.Good afternoon-
I would like to understand a bit more about the Kish correction. We have some data sets we are playing with, but we are unable to get the data to match the book examples, or other software (Desktop Reporter). Can you provide any more guidance or suggestions? I am assuming we are just not using the platform properly; and I can share the actual data set as well. Please let me know the best way to share the data.
Thank you-
Re: Would like Categorical Means Comparision to use Kish correction.Hi Scott,
Were you able to take a look at this?
Thanks,
Stan
Re: Would like Categorical Means Comparision to use Kish correction.This question: Questions about Multiple Response type in Categorical Platform
has a data table that was sent to us by P&G. I have uploaded a version that is a JMP table with a script here: Example Repeated Measures Data .
Run the script to see the calculations.
Re: Would like Categorical Means Comparision to use Kish correction.I'll see what we have that we can share.
Re: Would like Categorical Means Comparision to use Kish correction.Jeff- Thank you. Can you share a data set where we can see how this will look different in the two different cases?
Re: Would like Categorical Means Comparision to use Kish correction.I can confirm that the next EA release will contain the correction described above.
We should be very clear about the terminology here. We've seen the word "overlap" and "multiple response" to describe two different situations.
The correction here is appropriately applied to a case where a participant answers multiple questions but the answers may be related to each other. A more suitable term to describe this, and the one we will be adopting is Repeated Measures.
This correction will only be available the Related tab with the Repeated Measures option.
Re: Would like Categorical Means Comparision to use Kish correction.Posting this for clarity:
We're looking at implementing the formula referenced by varbanov.ar in Re: Questions about Multiple Response type in Categorical Platform :
sec. 12.4, p. 437 of the book Survey Sampling (1965) by Leslie Kish
Re: Would like Categorical Means Comparision to use Kish correction.As indicated in Questions about Multiple Response type in Categorical Platform there is currently no correction applied for the same person comparing two products.
If this is desired, please edit this idea to change it from a question.
Thanks!
