症状自评量表测验答题时间的规律

Law of reaction time in symptom checklist-90

目的:分析大规模群体心理测量中答题时间的规律,以鉴别被试个体在测量中是否存在不认真以及装好、装坏的现象,提高测试结果的可靠性。方法:于2005-01选择武装警察某部2004年底来自山东、安徽、福建等12个省和直辖市的男性新战士472人作为被试,应用解放军第四军医大学航空航天医学系医学工程研究室研制的DXC-Ⅳ型心理测评仪进行症状自评量表测验,该量表包括躯体化、强迫症状、人际关系敏感、忧郁、焦虑、敌对、恐怖、偏执、精神病性、附加项目10个因子,答题按1～5级记分:1无,2轻度,3中度,4重度,5严重。同时采用自编字对匹配测验检测被试对汉语语词的理解以及准确表述的能力,共24题,要求被试按“是(有意义),否(无意义)”键回答。DXC-Ⅳ型多项心理测评仪自动记录每人每题的首次按键时间、最后松键时间(完成时间)、答题中间有无修改、最后答案4个参数。数据处理由DXC-Ⅳ型多项群体心理测评仪配套软件完成。结果:新战士472人均完成测试,全部进入结果分析。①症状自评量表测试中答题时间随题目和因子的不同而有所差异,每题的答题时间均值与题目字数正相关(相关系数=0.671,P<0.01),题目字数和答3(中度)比例进入所建立的回归方程Y=2.324+0.179X1+8.438X2,即被试“中度”的选项越多则答题时间越长。②症状自评量表在10个因子中附加项目得分与各因子答题时间均有显著正相关,恐怖因子得分与该因子答题时间存在正相关但与其他因子答题时间相关不显著。10个因子中除强迫症状以外其他因子的答题时间均与该因子得分存在正相关。③症状自评量表10个因子的答题时间和总时间在城乡之间、不同学历之间的差异无显著性(P>0.05)。结论:在群体测量中计算出症状自评量表中各因子答题时间,通过其大小及分布情况的分析将有助于判断被试个体测量结果的真实性。

AIM： To analyze the law of the reaction time of answering a question in large-scale group psychological test, so as to identify whether there manifestations of unseriousness and pretending to be good or bad in testecs, and improve the reliability of the tested results. ale the METHODS： Totally 472 male recruits, who enrolled at the end of 2004 from 12 provinces of Shandong, Anhui, Fujian, etc., were examined with symptom checklist-90 （SCL-90） an army of armed police in January of 2005. The DXC-Ⅳ machine of psychometrics made in Research Room of Medical Engineering, Department of Aerospace Medicine, Fourth Military Medical University of Chinese PLA was used in the SCL-90 test. The SCL-90 consisted of 10 symptom dimensions of somatization, obsession-compulsion, interpersonal sensitivity, depression, anxiety, hostility, phobic anxiety, paranoid ideation, psychoticism, and additional items. The answers were scored by 1-5 classes： 1 as none, 2 was mild, 3 as moderate, 4 as severe, 5 as very severe. At the same time, the abilities of comprehension and expression to Chinese language of the subjects were examined with the character-match test, which included 24 topics. The subjects were requested to press ＂yes （meaningful） or no （nonsense）＂ key to answer. We collected 4 15arameters： the first-key-tinie, the last-key-time, the modification and the answer, and processed the date by the DXC-Ⅳ machine of psychometrics. The data were dealt with by the matched software of DXC-IV machine of psychometrics RESULTS： All the 472 recruits completed the test and were analyzed. ①In the SCL-90 test, the reaction time of answering a question differed for different question and factor, and the average reaction time was positive correlated with the word count of the question （r=0.671, P 〈 0.01）, the word count of the question and the proportion that the subject chose 3 in five answers entered the regression equation Y =2.324 ＋0.179X1 ＋8.438X2, that was to say, the more contents in the choice of ＂moderate＂, the longer the reaction time to answer the question. ② In the 10 items of SCL-90, the score of additional item had significant positive correlation with the reaction time to answer the all questions, the score of phobic anxiety was positively correlated with the reaction time to answer this question, but had insignificant correlation with the reaction times to answer the other questions. Except obsession-compulsion, the reaction times to answer the other questions all had positive correlation with the scores of corresponding factors. ③ The reaction times to answer the questions and the total time in SCL-90 had insignificant differences among the recruits from urban and rural areas and among those with different educational background （P 〉 0.05）. CONCLUSION： The study suggests that we can analyze the reaction time to find out that if the subject is faithful to the exam and what is his motivation in it and to increase the validity of the SCL-90 ultimately.