cancel
Showing results for 
Show  only  | Search instead for 
Did you mean: 
JMP is taking Discovery online, April 16 and 18. Register today and join us for interactive sessions featuring popular presentation topics, networking, and discussions with the experts.
Choose Language Hide Translation Bar
syy010
Level I

Standard deviation and p-value for difference scores?

Does anyone know how to create standard deviations and p-values if you are looking at the difference scores of two averages? I think I have to do a one tailed t-test but I am really unsure how I would do that in JMP.

1 ACCEPTED SOLUTION

Accepted Solutions
julian
Community Manager Community Manager

Re: Standard deviation and p-value for difference scores?

Hi syy010,

It sounds like you're looking to perform a matched-pairs/dependent-measures t-test. When you have data in dependent pairs (e.g. repeated measurements from the same individual, or across matched individuals) the analysis boils down to a one-sample t-test on the difference scores (*this is a different issue than a "one-tailed test" -- see below). These can be performed in JMP using the Analyze > Matched Pairs platform, or by using Analyze > Distribution after making a difference score column. In the latter case, you are testing whether the mean of the difference score column is different from 0 (which can be done by selecting the Red Triangle in the distribution platform > Test Mean). You will also get the standard deviation of the difference scores by using this second approach. 

Here is a page with details on these kinds of tests, including a video and a PDF with step-by-step instructions:

https://community.jmp.com/docs/DOC-6775

You can find a complete listing of these JMP guides here: Collection: All One-Page Guides

I hope this helps!

julian

*A "one-tailed" test is a separate issue, and that has to do with how you specify your null and alternative hypotheses. A one-tailed hypothesis test, or a directional test, in when you are testing an alternative hypothesis in one direction (e.g. whether there is specifically an increase, or specifically a decease) rather than whether there is evidence of any difference (direction ignored). Directional hypotheses (one-tailed tests) are carried out the same way in JMP, just look at the p > t or p < t sections (depending on the direction you specify) rather than the p > |t| sections. If you'd like to know more about directional hypothesis (and some considerations when using them) here is a video of mine covering them in more detail:        Factors Affecting Power - Directional Hypotheses (Module 1 8 9)    

View solution in original post

1 REPLY 1
julian
Community Manager Community Manager

Re: Standard deviation and p-value for difference scores?

Hi syy010,

It sounds like you're looking to perform a matched-pairs/dependent-measures t-test. When you have data in dependent pairs (e.g. repeated measurements from the same individual, or across matched individuals) the analysis boils down to a one-sample t-test on the difference scores (*this is a different issue than a "one-tailed test" -- see below). These can be performed in JMP using the Analyze > Matched Pairs platform, or by using Analyze > Distribution after making a difference score column. In the latter case, you are testing whether the mean of the difference score column is different from 0 (which can be done by selecting the Red Triangle in the distribution platform > Test Mean). You will also get the standard deviation of the difference scores by using this second approach. 

Here is a page with details on these kinds of tests, including a video and a PDF with step-by-step instructions:

https://community.jmp.com/docs/DOC-6775

You can find a complete listing of these JMP guides here: Collection: All One-Page Guides

I hope this helps!

julian

*A "one-tailed" test is a separate issue, and that has to do with how you specify your null and alternative hypotheses. A one-tailed hypothesis test, or a directional test, in when you are testing an alternative hypothesis in one direction (e.g. whether there is specifically an increase, or specifically a decease) rather than whether there is evidence of any difference (direction ignored). Directional hypotheses (one-tailed tests) are carried out the same way in JMP, just look at the p > t or p < t sections (depending on the direction you specify) rather than the p > |t| sections. If you'd like to know more about directional hypothesis (and some considerations when using them) here is a video of mine covering them in more detail:        Factors Affecting Power - Directional Hypotheses (Module 1 8 9)