中文版精神病人刑事责任能力评定量表的信度和效度研究

A study on reliability and validity of Chinese version of criminal responsibility assessment scale in mental patients

目的：探讨中文版《精神病人刑事责任能力评定量表》（中文版R-CRAS）的信度和效度。方法：国内12名不同省市的资深司法精神病学鉴定专家运用R-CRAS量表采用盲法对3例犯罪嫌疑人的辨认能力或者控制能力以及刑事责任能力的评定结果进行比较。结果：量表内部一致性系数（Cronbach＇sα）r=0.846。12名评定者对全量表评定的一致性系数r=0.493,一致性最差的4个条目（第10、11、14、17题）占全量表的22.22%（r=0.099~0.294,P均〉0.05）,提示这些条目与刑事责任能力评定或许没有必然联系。12名评定者36例次评定结果：部分、小部分、大部分、完全和无刑事责任能力的分别为7（19%）、2（6%）、8（22%）、15（42%）和4（11%）例次,提示评定者根据量表对刑事责任能力评定的结果明显不同。对于12名评定者的36例次的评分转换成等级资料,进行Friedman检验（χ2=208.824,df=17）和Kendall＇W协同系数检验（Kendall＇W系数=0.341,χ2=208.824,df=17）,显著性均P=0.000,提示量表评分标准存在随意性。多维尺度分析发现,该量表条目按其相似性和差异性可以被划分为若干不同的范畴,仅以＂辨认能力＂和＂控制能力＂不能概括他们的特征。结论：中文版R-CRAS的设计欠合理,需进一步研究改进。

Objective： To study the reliability and validity of Chinese version of criminal responsibility assessment scale（ R-CRAS） in mental patients. Method： Twelve forensic psychiatrists from different provinces were presented three cases,and conducted blindly assessment of identification capacity,control ability and criminal responsibility of the three suspects. the assessment results were analyzed. Results： The internal consistency coefficient（ Cronbach＇s alpha） of the scale was r = 0. 846. The consistency coefficient of 12 forensic psychiatrists was r = 0. 493. The obvious inconsistency items were 10 th,11th,14 th and 17 th item,accounted for22. 22% of full items（ r = 0. 099 ~ 0. 294,all P 〉0. 05）,suggesting that there was no necessarily link between these items and criminal responsibility. Thirty-six rating results for criminal responsibility according to 12 forensic psychiatrists assessments as follows： partial 7（ 19%）,small partial 2（ 6%）,most 8（ 22%）,full 15（ 42%）and no 4（ 11%） responsibility,suggesting that assessment results varied much. Based on 36 rating results for criminal responsibility of 12 forensic psychiatrists converted to rank data,Friedman test（ χ2= 208. 824,df = 17）and Kendall＇W test（ Kendall Wa = 0. 341,χ2= 208. 824,df = 17,all P = 0. 000） results showed diversified assessments of the scale. Multi dimensional scaling（ MDS） found that the scale items could be divided into several different categories according to their similarities and differences.＂identification capacity＂and＂control ability＂could not generalize their characteristics. Conclusion： The Chinese version of R-CRAS is still imperfect and needs further improvement.