I read very helpful posts about LMC, a coding scheme which I am very new to. Related posts were :
(1) How can logistic regression have a factorial predictor and no intercept?
(2) Linear model in R with multiple dummies but no constant: Choice of included dummy variable levels
I thought LMC can be especially helpful for visualization such as coefficient plot as Reference Level Coding (RLC) does not provide an estimate for the reference group, which is not very visually appealing due to the uncomprehensiveness.
Yet, I am still not sure if LMC can be applied to applied quantitative research in social sciences where dozens, sometimes nearly a hundred, variables are used in a linear and additive model. Say, for instance, I am interested in highlighting factor variable x1 with 3 levels, while I control for additional continuous variables(x2 ~ x5) and other factor variables(x6 ~ x10). The statistical software is R.
In this case, would it be statistically okay to code x1 by LMC, adjust R to not include the intercept in the regression, and include all other covariates(x2 ~ x10) in a linear model, while forcing x6 ~ x10 to be a numeric type dummy variables? (like how they did in the second related post?)
