Auto-suggest helps you quickly narrow down your search results by suggesting possible matches as you type.

- JMP User Community
- :
- Discussions
- :
- I'm completely lost as to how to set up my data and what test to run. Could some...

- Subscribe to RSS Feed
- Mark Topic as New
- Mark Topic as Read
- Float this Topic for Current User
- Bookmark
- Subscribe
- Printer Friendly Page

Highlighted

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Jul 24, 2020 4:53 AM
(2328 views)

I'm new to JMP and statistics. I just collected the first round of data for my masters research and am at a loss on how to analyze it. I have eight plant species, from which I made four concentrations of leachate (1.87, 3.75, 7.5, and 15g/25ml). The leachate was was used to test germination inhibition (each concentration was tested on four replicates for each plant for a total of 128 data points). Just by looking at the data, it seems clear that the amount of inhibition varies by both concentration and species. A simple Y by X showed significance with concentration. However, I'm looking to test whether or not each species at each concentration is significantly inhibiting germination. I'm at a loss on what to do.

2 ACCEPTED SOLUTIONS

Accepted Solutions

Highlighted

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

I hope that you realize that you put 'the cart before the horse.' The design of your study should have been for the analysis to follow. Collecting data by any means and then wondering about how to analyze it is risky, at best.

Set up data columns for each factor (Species, Leachate Concentration) and the response. Each row is one observation, so enter the replicate observations on separate rows. Then select Analyze > Fit Model. Select Species and Leachate Concentration and click Macros > Full Factorial. Select the response column and click Y. Click Run.

Come back when you get this far to ask about interpretation.

Learn it once, use it forever!

Highlighted

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

You described your study but you did not show your design in the beginning. Your description led me to believe you could model/test the interaction effect but you cannot. The Singularity Report in one of your pictures highlights that you model is over-specified. So run the analysis again without the cross term and see what happens.

Learn it once, use it forever!

5 REPLIES 5

Highlighted

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

I hope that you realize that you put 'the cart before the horse.' The design of your study should have been for the analysis to follow. Collecting data by any means and then wondering about how to analyze it is risky, at best.

Set up data columns for each factor (Species, Leachate Concentration) and the response. Each row is one observation, so enter the replicate observations on separate rows. Then select Analyze > Fit Model. Select Species and Leachate Concentration and click Macros > Full Factorial. Select the response column and click Y. Click Run.

Come back when you get this far to ask about interpretation.

Learn it once, use it forever!

Highlighted
##

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: I'm completely lost as to how to set up my data and what test to run. Could someone help?

Created:
Jul 24, 2020 6:13 AM
| Last Modified: Jul 24, 2020 6:23 AM
(2311 views)
| Posted in reply to message from markbailey 07-24-2020

Thank you for your response! I now see the type of output for which I was looking! Both Species and Concentration are significant. However, the LS Means Student's T Test (with the connecting letters) as well as Tukey's test show that there is no statistical difference between species or concentrations. I'm not sure how this can be if both are deemed significant predictors of germination.

Highlighted
##

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: I'm completely lost as to how to set up my data and what test to run. Could someone help?

Can you post a picture of the results to which you are referring?

Learn it once, use it forever!

Highlighted
##

See attached. I believe the problem is stemming from the LSM being taken from the species/concentrations as a whole group and not species by concentration (of which there seems to be an obvious significance). However, when I use Fit Model with Germination as the Y and Species, Concentration, and Species*Concentration as the model effects, no discernible information is in the output (I've attached an image of this as well).

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

Re: I'm completely lost as to how to set up my data and what test to run. Could someone help?

Highlighted

- Mark as New
- Bookmark
- Subscribe
- Mute
- Subscribe to RSS Feed
- Get Direct Link
- Email to a Friend
- Report Inappropriate Content

You described your study but you did not show your design in the beginning. Your description led me to believe you could model/test the interaction effect but you cannot. The Singularity Report in one of your pictures highlights that you model is over-specified. So run the analysis again without the cross term and see what happens.

Learn it once, use it forever!