Hi all,
I'm currently working on micro data for graduates and their first-market wage. The idea is to check to what extent the class of origin affects the wage of the respondents, controlling for a set of educational and labour market controls.
However, a further step is to test the hypothesis that this relation is changing depending on the province of residence. For this purpose I would like to fit a multilevel model with random slopes and intercepts.
The problem I have is that the father class of origin is a factor variable and I do not exactly know how to specify a model where at each province I have a random intercept and a random slope for each class of origin. For example, if I would have a continuous variable I could fit something like:
However, as I have a factor main predictor, I guess I need a cross-effect multilevel model, in order to have different effects within each province depending on the different levels of the factor. But here is my problem, how shall I specify the multilevel model to have random slopes and intercepts for each level of the factor variable within and between provinces?
Fitting something like:
gives me only a random slope, without any random(_cons). Is this correct?
Many thanks for your support
I'm currently working on micro data for graduates and their first-market wage. The idea is to check to what extent the class of origin affects the wage of the respondents, controlling for a set of educational and labour market controls.
However, a further step is to test the hypothesis that this relation is changing depending on the province of residence. For this purpose I would like to fit a multilevel model with random slopes and intercepts.
The problem I have is that the father class of origin is a factor variable and I do not exactly know how to specify a model where at each province I have a random intercept and a random slope for each class of origin. For example, if I would have a continuous variable I could fit something like:
Code:
mixed log_wage x_continuous controls || province: x_continuous, var(unstructured)
However, as I have a factor main predictor, I guess I need a cross-effect multilevel model, in order to have different effects within each province depending on the different levels of the factor. But here is my problem, how shall I specify the multilevel model to have random slopes and intercepts for each level of the factor variable within and between provinces?
Fitting something like:
Code:
mixed log_wage ib4.father_class controls || province: R.father_class
Many thanks for your support
