In this paper we propose a robust version of Reduced and Factorial k- means, based on a trimming strategy. Reduced and Factorial k-means are data re- duction techniques for simultaneous dimension reduction and clustering. The occur- rence of data inadequacies can invalidate standard analyses. An appealing approach to develop robust counterparts of Reduced and Factorial k-means is given by impar- tial trimming. The idea is to discard a fraction of observations that are selected as the most distant from the centroids.