BookmarkSubscribe
Choose Language Hide Translation Bar
5tudent
Community Trekker

significance labels

Hi.

I'm new to JMP and Google could not help find an answer, probably because different search terms should have been used.  We performed some experiments in the lab and came up with a bar chart of 14 mean values from 14 experiments.  A simple t-test with hypothesized mean in JMP showed the test statistic for each of the 14 experiments, so now we can accept or reject the null hypothesis for each of those experiments.  Now we hope to publish in a manuscript, so we created a bar chart showing the 14 mean values.

We know that JMP can show confidence intervals on the bar chart, but the professor wants only a symbol over the bars for which the null hypothesis was rejected.  How can this be done in JMP?  Other software might do this, but JMP graphics look nicer for a publication.  Thank you for your advice!

0 Kudos
1 ACCEPTED SOLUTION

Accepted Solutions
julian
Staff

Re: significance labels

Hi 5tudent,

One way you can accomplish this in JMP is by using row labeling. I've attached a table with the setup you would need (which has a column formula to take your p-values and make them into stars based on a < 0.01, and < 0.05 cutoffs). I'm also including a quick screen clip of the actions you need to take to turn the labeling on in graph builder. Here are the steps written out:

1. Right click the "Label" column, and select "Label/Unlabel" to toggle labeling for that column

2. Launch Graph Builder - Graph > Graph Builder:

     - Y: Response

     - X: Experiment (as nominal)

     - Toggle Bar Chart
     - Right Click Bars > Bars > Label > Row Label

I hope this helps!

Julian

5 REPLIES 5
Steven_Moore
Super User

Re: significance labels

Just a point regarding hypothesis testing:  You never accept the null hypothesis.  You either reject the null hypothesis or you fail to reject it.  This may seem like nit-picking, but it is important, especially if you are going to publish the results.

Steve
5tudent
Community Trekker

Re: significance labels

You're correct, of course.  It's a very important point in case I have to include it in the manuscript.  When the conclusion is stated your way, the manuscript reviewer/referee will not yell at me!

0 Kudos
julian
Staff

Re: significance labels

Hi 5tudent,

One way you can accomplish this in JMP is by using row labeling. I've attached a table with the setup you would need (which has a column formula to take your p-values and make them into stars based on a < 0.01, and < 0.05 cutoffs). I'm also including a quick screen clip of the actions you need to take to turn the labeling on in graph builder. Here are the steps written out:

1. Right click the "Label" column, and select "Label/Unlabel" to toggle labeling for that column

2. Launch Graph Builder - Graph > Graph Builder:

     - Y: Response

     - X: Experiment (as nominal)

     - Toggle Bar Chart
     - Right Click Bars > Bars > Label > Row Label

I hope this helps!

Julian

5tudent
Community Trekker

Re: significance labels

Thank you so much!  It's pure genius.  And thank you for the video showing how it works!

0 Kudos
julian
Staff

Re: significance labels

You're welcome -- I'm glad I was able to help!

Julian

0 Kudos