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

- 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

Interpretation of leverage plot axes

Sep 22, 2020 11:22 AM
(237 views)

Hello,

I am trying to understand how to interpret the leverage plot axes-- specifically why negative x and y values are plotted when my data have none of these. I understand, I supposed, that negative y values could be the result of model uncertainty, but why the x values?

For sake of discussion, if it matters, I am running a multiple regression of rainfall amount (cm), development level of watershed (% developed) and looking at the creek response (flow). I have no negative values, yet the leverage plot shows X values well below my actual data range.

Thanks

3 REPLIES 3

Highlighted
##

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

Re: Interpretation of leverage plot axes

JMP adds the mean of Y to the Y-residuals and the mean of X to the X-residuals. So I expect you may have some very negative residuals ?

detail

https://www.jmp.com/support/help/en/15.2/index.shtml#page/jmp/leverage-plots.shtml

JMP Systems Engineer, Pharm and BioPharm Sciences

Highlighted
##

Thank you. Yes there are some fairly negative residuals but the model appears very good and the Y residuals are nicely scattered.

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

Re: Interpretation of leverage plot axes

Highlighted
##

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

Re: Interpretation of leverage plot axes

As long as the magnitude of the residuals aren't concerning... (and without looking at anything) you're probably OK.

When I have extreme values (high or low X) that are far from the center of the data (Y) I'll take a little closer look at that point because it exerts a higher degree of "leverage" on the regression.

JMP Systems Engineer, Pharm and BioPharm Sciences