计算机辅助电话调查系统（Computer Assisted Telephone Interviewing System），简称CATI系统，是利用计算机辅助电话调查而开发的调查访问操作系统，是近年来高速发展的通讯技术及计算机信息处理技术与传统的电话调查相结合的产物。CATI系统自20世纪70年代在美国诞生以来，以其无可比拟的优势在美国、西欧等地得到迅速推广应用，在许多调查中已取代传统的面对面访问方式，成为一种重要的调查方法。CATI系统在我国的应用虽然较晚，但已有不少信息咨询公司和官方统计机构运用CATI系统开展调查，正得到越来越广泛的应用。本报告从内容上看有两个部分。第一个部分是在介绍计算机辅助电话调查实施的过程中，计算机辅助电话调查系统的革新和发展的基础上，以国内著名的计算机辅助电话调查系统，雅典娜CATI系统为例，阐述了社会调查的关键环节如何计算机辅助电话调查系统中得以实现。第二部分则是在计算机辅助电话调查实施中一些关键的问题如何实现。首先，本工作报告在给出计算机辅助电话调查中随机拨号法抽样的原理基础上，给出如何估计的参数，并创新性的给出了序贯抽样方法，即在一些社会调查问题中，给定调查费用和估计的误差项后，无须计算样本容量，而是，在电话调查的同时，判断调查是否可以结束，即估计的误差项是否被满足，相比而言，比面访中样本量的缺点更为合理。其次，电话调查中一个弊端是调查对象，即访问者没有按理想的方式被等概率访问，本工作报告给出了事后分层技术，借助调查对象的背景信息对电话访问样本做后期校正，从而弥补访问者没有按理想的方式被等概率访问这一弊端。最后，则是从数据挖掘的角度给出了利用关联规则方法做背景分析的探讨。

Computer Assisted Telephone Interviewing System, abbreviated as CATI, is a tool to carry a survey research by making use of computers and telephone. The first CATI was born in United States in 1970s, and then rapidly applied and developed in many investigations from United States to Western Europe and other places as a substitute of the traditional way of interview face to face. The applications of CATI was late, in a few consulting company and official statistical agencies at firstly, is being more and more widely used.This report has two parts. The first part is to introduce the history of innovation and development of CATI, and then Athena CATI system, an domestic well-known computer-assisted telephone interviewing system, the implementation process of computer-assisted telephone survey lastly.The second part is discussion about some key issues of computer-assisted telephone survey. First of all, the theory of random digit dial in the sample of computer-assisted telephone survey and estimate method of parameters about population are showed. Then a sequential sampling method is suggested, by which we need not to calculate the sample size restricted to cost and estimated error. Sequential sampling method is more economic than sample method in interview face to face. Second, an emendation method of sample of telephone interview is suggested to remedy the fact that visitors of telephone interview are not phoned with equal probability. Finally, we discuss cross-analysis of telephone interview sample using association rules from the perspective of data mining.