☐ cool new feature
☐ could help many users!
☑ removes a „bug“
☐ nice to have
☐ nobody needs it
What inspired this wish list request?
this community discussion :
@gnu23 wrote:
I think the discussion boils down to the question, if statistics on multiple response columns should generally either use the number of responses (standard now in JMP) or the number of rows as 'n'. In my opinion the second option would make more sense for many use cases and an option to switch between these two values for 'n' would increase the (already very nice) functionality of multiple response columns.
The default statistics in Graph Builder for counts and percentage, is (*):
"one row counts one - Ntotal (denominator for percentages) is the number of rows".
[This statistics can be adjusted via the Freq/Weight drop zone.]
Surprising:
If a column with modeling type "multiple response" is used ANYWHERE in the graph, the statistics changes to
"one response counts one [OK]
Ntotal (denominator for percentages) is number of responses" [maybe: not OK].
So, to generate meaningful bar charts on the porcelain layer, it seems that on the plumbing layer Jmp
- stacks the individual responses, creating additional "rows",
- and then applies the standard logic (*)
For simple bar charts, this might be reasonable, but for Ntotal and the calculated percentages it leads to funny/strange effects.
What is the improvement you would like to see?
Please provide a more meaningful statistics for Multiple Response (e.g. as an alternative to the current version):
"one response counts one
But: denominator for percentages (Ntotal) is the number of rows contributing to the plot - like in standard Graph Builder"
on the plumbing layer one might think of:
Any column with multiple responses is split into "indicator columns", representing each response - the number of rows stays constant.
This approach is much closer to the default Statistics in Graph Builder. As a side effect, this alternative approach will also fix a bug which shows up if multiple columns with modelling type "multiple response" are used in one graph:
https://community.jmp.com/t5/Discussions/Graph-Builder-Unclear-statistics-when-using-multiple-respon...
Why is this idea important?
a) Meaningful statistics.
b) no misleading statistics anymore for users who don't "notice"
1 Comment