topic Re: How to do significant difference analysis when we only know the average and stdev data? in Discussions

How to do significant difference analysis when we only know the average and stdev data?

Freya — Wed, 18 Apr 2018 02:47:04 GMT

Is it possible to do significant difference when we only have the average data and stdev data on JMP? I know when we have the data from several groups of repeated experiments, it can be analysised through oneway analysis<< compare means<< each pair, student's t.

Re: How to do significant difference analysis when we only know the average and stdev data?

cwillden — Wed, 18 Apr 2018 04:10:32 GMT

With scripting, all things are possible in JMP. Using a JMP platform is probably not going to happen for you with just summary statistics on your groups. How comfortable are you with the idea of using a script and doing the computations yourself? How many groups are you comparing?

If you are comparing several groups simultaneously, I'd recommend using ANOVA and Tukey's HSD. Doing t-tests on all pairs is the classic multiple comparisons situation. If you are unfamiliar with that, check out this Wikipedia page: https://en.wikipedia.org/wiki/Multiple_comparisons_problem. However, these days people seem to care less about family-wise error rates. Up to you.

If you want to do the ANOVA/Tukey approach with your summary statistics, you can use this handy little calculator: http://statpages.info/anova1sm.html. I'm sure there's a million t-test calculators out there that just take the summary statistics if you Google it.

Re: How to do significant difference analysis when we only know the average and stdev data?

Dan_Obermiller — Wed, 18 Apr 2018 14:09:00 GMT

This will depend on the analytical technique you are calling "significant difference analysis". If you are referring to ANOVA, then there is an add-in to do this:

https://community.jmp.com/t5/JMP-Scripts/ANOVA-from-Data-Summary/ta-p/21487