Thirty percent of the participants were classified as not impaired with ADLs and only 3% were impaired with all ADLs. However, it seems that if the therapist being present during task performance improved ADL function, the correlations that resulted would have been lower or none existent.