Hi everyone,
I am analyzing data from a large-scale survey. I used the command mi (sequential imputation using chained equations) to impute the missing values in the dataset.
I am trying to figure out how individual-level factors (e.g., parental education, household income) and neighborhood factors (e.g., grand mean of parental education, urbanicity) affect students’ school drop-out (binary response: 0 for completion, and 1 for drop-out). I also want to include sampling weight (weight) into the model.
I tried to use mi estimate: meqrlogit function to run a two-level model, with individual-level factors and sampling weight included into level 1, and neighborhood-related factors included in level 2.
The code is like below:
mi estimate: meqrlogit Dropout Parent_edu income [pweights = weight] || neighborhoodid: urban Paredu_grand
However, it comes out an error that “weights not allowed”. My question is, how can I include sampling weight into a two-level logistic model of imputed data?
Thank you so much in advance!
I am analyzing data from a large-scale survey. I used the command mi (sequential imputation using chained equations) to impute the missing values in the dataset.
I am trying to figure out how individual-level factors (e.g., parental education, household income) and neighborhood factors (e.g., grand mean of parental education, urbanicity) affect students’ school drop-out (binary response: 0 for completion, and 1 for drop-out). I also want to include sampling weight (weight) into the model.
I tried to use mi estimate: meqrlogit function to run a two-level model, with individual-level factors and sampling weight included into level 1, and neighborhood-related factors included in level 2.
The code is like below:
mi estimate: meqrlogit Dropout Parent_edu income [pweights = weight] || neighborhoodid: urban Paredu_grand
However, it comes out an error that “weights not allowed”. My question is, how can I include sampling weight into a two-level logistic model of imputed data?
Thank you so much in advance!
