Turn on suggestions

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

Showing results for

- JMP User Community
- :
- Discussions
- :
- Discussions
- :
- Residuals save as zero, should be .00005 (is JMP rounding?)

Topic Options

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

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

Residuals save as zero, should be .00005 (is JMP rounding?)

Mar 7, 2015 9:51 AM
(1129 views)

Is there some setting that would round the residuals down to zero? I'm having trouble because I do a full factor analysis, save the residuals, and they're all zero. They should be in the neighborhood of .00005 to .00010

1 REPLY

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

Re: Residuals save as zero, should be .00005 (is JMP rounding?)

Mar 7, 2015 11:27 PM
(1031 views)
| Posted in reply to message from lizstatistics 03/07/2015 12:51 PM

Hi lizstatistics,

Without knowing more about your data it's hard to guess what is happening here, but I can tell you that JMP will not round the saved residuals from your model. Can you describe your data a bit more, and why you expect residuals around 0.00005?

Here are some guesses: It is possible that the saved column is not displaying that many decimal places (though it should) -- if for some reason that's the case, you can change the number of decimal places displayed by right-clicking the column > Column Info, and then select "Fixed Dec" under formatting, and enter a value more than 6 for decimal places. By default you should have well more than that displayed, so I don't think that's the issue

Another guess: When you fit this model, is your R^2 = 1 (in the Summary of Fit table)? It could be that you have have a model in as many parameters as you have data points, resulting in a model that fits the data perfectly, resulting in residuals that are all 0. If that's the case you will need to remove some some of the higher-order interactions since you don't have the data to estimate those independent of error.

I hope this helps!