Timeline for lme4_fixed-effect model matrix is rank deficient so dropping 1 column / coefficient
Current License: CC BY-SA 3.0
5 events
| when toggle format | what | by | license | comment | |
|---|---|---|---|---|---|
| Dec 15, 2020 at 18:18 | comment | added | ccw |
Also, just for clarity and brevity of code, you don't need to include both Language * Canonicity and Language + Canonicity in your model. Language * Canonicity is equivalent to Language + Canonicity + Language:Canonicity, where Language:Canonicity is the interaction effect between language and canonicity. :)
|
|
| Apr 2, 2017 at 2:09 | comment | added | Jacob Socolar |
I'm not sure i follow what a subject is in this model, or what model exactly is producing the error. If you run the exact same model as above but replacing data$useOfInt with its logarithm, it still has the same problem as before--every language has exactly one value of log(data$useOfInt) associated with it, so you can't put both that and language in a single model.
|
|
| Apr 1, 2017 at 23:33 | comment | added | Aloush87 | Totally make sense. I instead used the frequency of intrinsic use per subject. The model works fine with that. However, I get the same error if I scaled this factor 'log(data$useOfInt)', I am wondering why is that? Thank you | |
| Mar 30, 2017 at 3:50 | history | edited | Jacob Socolar | CC BY-SA 3.0 |
added 824 characters in body
|
| Mar 30, 2017 at 3:45 | history | answered | Jacob Socolar | CC BY-SA 3.0 |