turn on suggestions

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

Showing results for

Find a Community

- Home
- /
- Analytics
- /
- Stat Procs
- /
- What to do when two models have exactly same AIC/B...

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
- Highlight
- Email to a Friend
- Report Inappropriate Content

07-19-2014 11:06 PM

I have two models in NLMIXED (one linear-broken line and the other a quadratic broken-line) and they have the exactly same AIC, BIC and variance components. Their AIC, BIC are smaller than other models I am comparing them to. I am using this models to estimate nutritional requirements of animals. Now, I am normally reporting only the estimated requirement of the best fitting model. In this case, should I report both of them? Or there is any further analysis/investigation can be done to find which one is better?

Thank you,

Marcio

Accepted Solutions

Solution

07-22-2014
11:42 AM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

07-22-2014 11:42 AM

I would expect the quadratic to be more easily explainable in biological terms (at least if it is quadratic to a plateau). Otherwise, linear is simpler to explain--there is an additive response above a certain cut-off.

Steve Denham

All Replies

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

07-20-2014 06:20 AM

Hi,

One approach is focusing on the "Story Telling" side of your model.

Best,

Mohammad

Solution

07-22-2014
11:42 AM

- Mark as New
- Bookmark
- Subscribe
- Subscribe to RSS Feed
- Highlight
- Email to a Friend
- Report Inappropriate Content

07-22-2014 11:42 AM

I would expect the quadratic to be more easily explainable in biological terms (at least if it is quadratic to a plateau). Otherwise, linear is simpler to explain--there is an additive response above a certain cut-off.

Steve Denham