IMPROVING VALIDITY OF SELF REPORTS OF CANCER SCREENING

这项建议的具体目的是进行一项复杂的析因实验，以评估四种收集自传信息的方法，我们假设这些方法可能有助于减少对促进健康和预防疾病行为的过度报道。要评估的方法包括：a)从受访者那里收集数据的方式；b)在询问行为是否发生之前，询问获得检查的意图(一种社会合意的行为)；c)在询问他们是否接受了检查之前，询问可能阻止他们进行检查的障碍或问题(不执行该行为的理由)；以及d)询问作为年度体检的一部分，他们可能报告进行检查的规律性的例外情况，以此作为改进他们估计的一种方式。我们将把这项研究的重点放在三个特定的癌症筛查程序上：巴氏试验、乳房X光检查和乳房检查。数据将从在伊利诺伊州香槟-厄巴纳生活和接受医疗保健的800名55岁及以上的妇女那里收集，她们在该地区生活至少五年。在使用计算机辅助电话采访(CATI)或音频计算机辅助自我采访(ACASI)完成采访后，我们将要求受访者允许摘录他们的医疗记录，以验证他们的自我报告或癌症筛查检查。分析将侧重于评估每一种实验操作的实际效用，以提高自我报告的准确性。

The specific aim of this proposal is to conduct a complex factorial experiment to evaluate four approaches to collecting autobiographical information that we hypothesize may help reduce over reporting of health promoting and disease prevention behavior. The methods to be evaluated include: a) the mode in which the data are collected from the respondent, b) asking about intention to obtain the procedure (a socially desirable behavior) before asking whether the behavior occurred, c) asking about barriers or problems that might keep them from getting the procedure (justification for not performing the behavior) before asking about whether they had received it, and d) asking about exceptions to the regularity with which they might report getting the examination as part of their annual physical examination, as a way of improving their estimates. We will focus this research on three specific cancer screening procedures: Pap tests, mammograms and breast examinations. Data will be collected from 800 women aged 55 and older who live and receive their health care in Champaign-Urbana, Illinois and who have lived in the area for at least five years. After completing interviews using either computer-assisted telephone interviewing (CATI) or audio computer- assisted self interview (ACASI), we will ask respondents for permission to abstract their medical records in order to validate their self reports or cancer screening examinations. Analyses will focus on assessments of the practical utility of each experimental manipulation for improve the accuracy of self-reports.