|
|
 |
| Significace of non-linear slopes |
 
|
| Message/Author |
|
|
| Anonymous posted on Friday, June 24, 2005 - 4:20 am
|
|
|
How do you test if you should include a non-linear slope in the model?
Do you compare model fit with and without the higher order slope term? If so, if there is a significant difference, do you conclude that the higher order slope accounts for significant variance in the model?
Thanks. |
|
|
| BMuthen posted on Sunday, June 26, 2005 - 2:11 am
|
|
|
| Yes, you compare those two models. But a significant difference may only indicate that the mean of the non-linear slope is significantly different from zero not necessarily the variance. |
|
|
| Back to top |
|
|
