Efficient use of telephone survey response facilities: a decision aid

Abstract

This thesis develops an object-oriented simulation model of the Computer Aided Telephone Inquiry (CATI) system currently employed by the Defense Health Resources Study Center, which allows recipients of mailed survey questionnaires to respond to the mailed questionnaires via telephone. The simulation models system performance and the response arrival process as a transitory queuing system. The primary focus of this study is to develop a predictive decision aid for effective and efficient employment of the CATI system, while minimizing response attrition due to system overload. Sensitivity analysis is conducted to determine arrival rates which overload the system, mean service time effect on system capacity, and effects of various retry decision processes (i.e., the arrival process for respondents who fail to access the system because of system overload). Additionally, possible network optimizations designed to aid in the development of appropriate mailing strategies are discussed. As a predictive tool, the model appears to be quite accurate. Network optimization solutions for mailing strategies may achieve a significantly lower caller attrition rates than strategies which call for evenly distributed batch survey mailings.