De Irala J, Bigelow C, McCusker J, Hindin R, Zheng L
University of Massachusetts at Amherst, School of Public Health, AIDS Research Unit, MA 01003, USA.
Am J Epidemiol. 1996 Apr 1;143(7):725-32. doi: 10.1093/oxfordjournals.aje.a008806.
This study examined test-retest reliabilities of self-reported human immunodeficiency virus (HIV) sexual and drug injection behaviors among 246 prior drug users admitted to either of two residential drug treatment programs in Westborough, Massachusetts, and Providence, Rhode Island, between June 1990 and September 1992. Participants, selected by their date of admission, were administered admission and reliability questionnaires pertaining to HIV risk behaviors, the latter at approximately 2 weeks after admission. Estimated reliabilities (kappa coefficients) of the sexual behaviors ranged from 0.72 to 0.91; those for the drug injection variables ranged from 0.63 to 0.98. These results were consistent across groups defined by sex and injection of drugs. The consistently good reliabilities are significant to the design of independent studies of drug treatment populations utilizing self-report measures of sexual and drug behaviors.
本研究调查了1990年6月至1992年9月期间，在马萨诸塞州韦斯托伯勒和罗德岛州普罗维登斯的两个住院戒毒项目中收治的246名既往吸毒者自我报告的人类免疫缺陷病毒（HIV）性传播行为和药物注射行为的重测信度。根据入院日期选择参与者，让他们填写与HIV风险行为相关的入院问卷和信度问卷，后者在入院后约2周填写。性行为的估计信度（kappa系数）在0.72至0.91之间；药物注射变量的信度在0.63至0.98之间。这些结果在按性别和是否注射毒品定义的各组中是一致的。持续良好的信度对于利用性传播行为和药物行为自我报告测量方法对戒毒人群进行独立研究的设计具有重要意义。
