Problem with level 1 . variance

[georgecorbett]

Hi, I am running a three-level iterative measurement model with binary results (level 1, person level 2, area level 3). In my model I can't select the first level box when I add the cons because it doesn't appear, only the fixed parameters, level 2 and level 3 boxes appear.
How can I choose the level 1 variance. Thanks!

[ChrisCharlton]

If you are running a binary logit model then the variance is fixed at (π^2)/3 (≈3.29) rather than being an estimated parameter, and this is why you cannot add this to your model.

See the following FAQ for more information about variances: Partitioning variation across levels

[Michael42]

If you are running a binary logit model then the variance is fixed at (π^2)/3 (≈3.29) rather than being an estimated parameter, and this is why you cannot add this to your model.

See the following FAQ for more information about variances: Partitioning variation across levels

Hey, thanks! I am running a binary logit model and it really worked.

[singthink]

Hi, I am running a three-level iterative measurement model with binary results (level 1, person level 2, area level 3). In my model I can't select the first level box when I add the cons because it doesn't appear, only the fixed parameters, level 2 and level 3 boxes appear.
How can I choose the level 1 variance. Thanks!

In a three-level model with a binary outcome, level 1 does not have independent variance as in continuous linear models because the binary variable is fitted to a logistic or binomial distribution. Therefore, you cannot choose the variance at level 1 like you would at other levels. Instead, the variance at level 1 is dictated by the nature of the binary distribution. The variances can be modeled at level 2 (person) and level 3 (area), where you can control for random effects.

[Saedovis]

Hi, I am running a three-level iterative measurement model with binary results (level 1, person level 2, area level 3). In my model I can't select the first level box when I add the cons because it doesn't appear, only the fixed parameters, level 2 and level 3 boxes appear.
How can I choose the level 1 variance. Thanks!

In a three-level model with a binary outcome, level 1 does not have independent variance as in continuous linear models because the binary variable is fitted to a logistic or binomial distribution. Therefore, you cannot choose the variance at level 1 like you would at other levels. Instead, the variance at level 1 is dictated by the nature of the binary distribution. The variances can be modeled at level 2 (person) and level 3 (area), where you can control for random effects.

Great and detailed answer. Thank you.