This paper extends symmetrically trimmed least squares estimators to a
fixed effects bivariate truncated/censored regression model. The esti
mator proposed in this paper is consistent and does not require specif
ying the joint density of the error terms.