JMP 18 introduced an enhanced Python integration, and we will have R integration support via that Python Integration in JMP 19.
Regarding your second question, JMP does offer sample size calculations under the DOE menu, specifically in the Sample Size Explorers section. Please note that the names of these tests may differ from those in SPSS. If you’d like assistance navigating the Sample Size Explorers or have specific questions, I recommend reaching out to our Technical Support team—they’ll be happy to help!
Hello thiennguyen. Adding to this conversation... Under the JMP sample size explorers you will find several options for computing power for various scenarios. It is hard for us to compare calculations to what SPSS is doing in their related samples proportion calculator since we do not have access to their computations. One thing you could try in JMP is using the one sample proportion calculator for the paired sample data. This should give a reasonable estimate assuming you want the null hypothesized difference in sample proportions to be zero and the alternative hypothesized proportion would whatever difference you wish to be able to detect. It would be very interesting to see how these computations compared to the SPSS related samples computations under the same conditions.
Another option if your data is actually in the form of numerical proportions would be to use the one sample mean calculator for the difference in sample means you want to detect for the proportion data (ie: perform a typical paired sample means test, which is really a one sample test on the differences). This option would likely be the best option, especially if you used something like the delta method approximation for the group standard errors. There is currently no option to directly apply the delta method in JMP, so this would require a little extra work.
We are always looking to improve our sample size explorers to account for common scenarios that our users encounter. We will be continuing to evolve the platforms in JMP 20 and beyond.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
