Improvements to Power Explorer Platform

[shampton82]

What inspired this wish list request? While using the Power Explorer platform, I find myself having to jump back and forth between different input boxes as I try and examine different operation conditions due to the fact that few of the boxes can be locked.  This makes for a frustrating and ineffective user experience.  Also, there are some one and two sample test I'd like to do power analysis for an equivalence test but there is currently no option.

What is the improvement you would like to see?

• Add Power to the "Solve for" drop down options
• Add Margin of Error to the "Solve for" drop down options
• Only allow what has been selected in the "Solve for" drop down  to change as the other input boxes are updated
• Add the ability to lock the ratio between G1 and G2 sample sizes
• For Sample Counts per Unit, change the "Assumed Ratio of Group Counts per Unit" inputs to a single input of the ratio of group 1 to group 2 defect counts per unit.
• Add in one and two sample Equivalence test for Proportions and Counts

Why is this idea important? The Power Explorer platform is a key platform but is hard to use and less intuitive than the Minitab equivalent(though it does have more and better capability!).

Thanks for considering this!

Steve

[SarahGilyard]

These are interesting suggestions. Thank you for posting shampton82

[SarahGilyard]

Hi  shampton82  .  In JMP 19 we have added ability to address the following feature request:

• Add the ability to lock the ratio between G1 and G2 sample sizes

We will continue investigating your remaining requests for improving the Sample Size Explorer platforms. Thank you!

[shampton82]

Thanks @SarahGilyard, appreciate all the updates you've been doing on Wish list items.

[jszarka]

In what version were you referring to your original comments, @shampton82? When this platform was introduced in JMP 16 (I think), there were elements that I liked in terms of being able to have k-1 elements fixed, and then 'solve' for the kth element. I consider this as a more 'static' case, which can be very useful when you have a specific situation. The current platform in JMP 19 is much more 'dynamic' where you can see the Profiler view with many more features turned on. If I have a specific situation, it is likely much more than I need to consume.

Below is a comparison of the cases. For instance, if I already know my stdev=2, I don't need the dynamic element of how if the stdev was actually 3, that the power would be significantly lower.

Perhaps my adjacent 'wish list' item is to have 2 modes - static and dynamic. In static mode, it could be easy for instance to have 'equal sample sizes for 2 groups' enabled, whereas the dynamic mode could look at a much broader range of options.