Summary: For linear models in a parametric mold, independence, homoscedasticity and normality of errors constitute the basic regularity conditions, and these have been relaxed to varying extents in alternative approaches based on robust and nonparametric methods. In mixed effects models, the random effects components may generally introduce some complications in such robust and nonparametric procedures. Some of these difficulties are eliminated here through a conditional functional estimation approach, and in the light of that, improved estimation of the fixed effects parameters is presented in a unified manner.