Multi-Mode Political Surveys an Exploratory Study
Jacquelin Redman
Franklin & Marshall College Center for Opinion Research
Scottie Thompson
Franklin & Marshall College Center for Opinion Research
Berwood Yost
Franklin & Marshall College Center for Opinion Research
Despite the fact that recent polling in the US has performed no worse than in the past on average, significant misses in some races have pollsters and poll watchers concerned about the future of political polling. Of primary concern is the potential biases created by increasing cellphone-only households, declining response rates, and the relationship between non-coverage, nonresponse and voting. The increasing costs required to conduct methodologically sound pre-election polling is another concern. These concerns are forcing pollsters to adopt new methodologies, one of which is the use of multi-mode surveys. Multi-mode surveys are used to combine the strengths of multiple methodologies in order to achieve higher response rates, more representative samples and better response quality. There is also evidence that using a multi-mode design reduces the cost and increases the efficiency of a survey. This paper explores the feasibility of using a multi-mode design (web and telephone) to conduct voter surveys by presenting a comparison of survey efficiency, response rates, sample representativeness, and survey estimates produced by using telephone and web-based data collection strategies.