The participants in this pilot study included 46 females and 27 males who represented a sequentially-selected, stratified, random sample. The stratification criterion was the presence of children under age 16. The median number of children was 2 and the median years of marriage was 7. Thirty-six individuals had not graduated from high school; only 6 were university graduates. Findings revealed that hassles prior to mediation were associated with post-mediation hassles. In addition, engaging in violence-prevention activities prior to separation and being frightened of the partner were associated with abuse after the separation. Furthermore, evaluations of the mediation process were associated with post-separation abuse and only among those who took part in mandatory legal aid mediation. Results suggested that a screening instrument for abuse should include four variables: hassles prior to separation, being frightened of the partner, engaging in violence prevention activities prior to mediation, and completing an affidavit before mediation. Tables, figure, notes, and 13 references (Author abstract modified)