We describe a new approach to fault collapsing that extends fault collapsing based on fault equivalence and fault dominance. The new approach is based on a metric called level of similarity between faults. Informally, a fault f_j is said to be similar to a fault f_i with a level of similarity SL_{i,j ?} ≤ 1 if a fraction SL_i,j of the tests for f_i also detect f_j. If SL_i,j is high enough, one may exclude f_j from the set of target faults and rely on the test for f_i (and tests for other faults) to detect f_j. We describe a procedure for fault collapsing based on the level of similarity, and study its effectiveness experimentally.