In this section, the biases and variances of LL estimation and DSLL estimation in partial linear models are analyzed under the following assumptions. Let K1(v)=∫K(v−u)K(u)du, K2(v)=∫(v−u)uK(v−u)K(u)du, and V(K)=∫{K1(v)−K2(v)/μ2}2dv. Set μk=∫ukK(u)du, and νk=∫ukK2(u)du. Obviously, μ1=1