本研究期望以URICA量表為工具探討家暴加害人在接受12至24次之團體諮商輔導效果，各改變階段間之差異，及討論URICA量表作為輔導效果評估工具之適用性。研究設計以中部某縣市22位參加團體諮商輔導之家暴加害人為研究對象，這些家暴加害人須參加6、12、18或24次的團體諮商輔導。使用家庭暴力評估量表(University of Rhode Island Change Assessment-Domestic Violence, URICA-DV)為研究工具，在團體諮商輔導前後進行量表測量，以評估參加團體前後差異及團體次數對URICA各階段分數之效果。研究發現：以URICA四個階段分數作為指標，各組均處於思考期與行動期。將前後測的分數進一步分組進行比較，參加12次團體組的成員，URICA四個階段的分數，並未有明顯變化；但參加24次團體者，URICA四個階段的分數則有所差異，「前思考期」分數降低，另外三個階段（思考、行動、維持）的分數則上升，且參與者間變異降低，表示參與者的表現越來越好且越來越接近，符合參加團體所預期的效果。其次，亦計算前測與後測四階段分數之相關，以了解是否可以前測分數預測後測之表現。在12次團體組，前測的思考期與前測行動期、後測思考期、後測行動期均有正相關；在24次團體組，前測的思考期則僅與前測的行動期有正相關。再其次，參加團體的次數與後測維持期的分數有正相關，表示參加團體次數越多，維持期的表現越好。最後，根據研究結果提出對實務與未來研究之建議。

The purpose of this study was to explore the applicability of using the URICA scale as a tool for evaluating the effectiveness of group counseling for domestic violence perpetrators and the differences between the stages of change. There were 22 batterers who were mandated to participate 6, 12, 18, or 24 group counseling sessions. The University of Rhode Island Change Assessment-Domestic Violence (URICA-DV) was used as a research tool to assess the effectiveness of the group counseling program and to measure the numbers of participating sessions on the scores of URICA at each stage. The study result showed that using URICA four-stage scores as indicators, each group was either in the thinking period or in the action period. Further comparing the pre- and post-test scores of the groups, the scores of the four stages of URICA of the 12-sessions group did not change significantly. On the other hand, for those who participated in 24 sessions, the scores of the four stages of URICA were different. Their "pre-thinking period" scores decreased, and scores in the other three stages (thinking, action, maintenance) increased and their variation decreased, indicating that the participants' performance was getting better and closer which was in line with the expected effect of participating groups. In addition, the number of sessions participating is positively correlated with the posttest maintenance period score, indicating that the more the number of sessions participating, the better the performance of the maintenance period. Furthermore, the correlation between the pre-test and post-test four-stage scores is also calculated to understand whether the pre-test score can predict the performance of post-test. In the 12-session group, the score of pre-test thinking period is positively correlated with the pre-test action period, post-test thinking period, and post-test action period. In the 24-session group, the score of pre-test thinking period was only positively related to the pre-test action period. The conclusion was that according to the changes and correlation in the scores pre- and post-test, it might be reasonably speculated that the effect of participating in 24 sessions is better than the effect of 12 sessions and the scores of “thinking period” of pre-test could predict the post-test score better. Finally, based on the results of the study, suggestions for practice and future research are put forward.