Using Telephone Survey Data to Predict Participation in an Internet Panel Survey
Roe, D., Stockdale, J. D., Farrelly, M. C., & Heinrich, T. D. (2007, May). Using Telephone Survey Data to Predict Participation in an Internet Panel Survey. Presented at American Association for Public Opinion Research Conference, Anaheim, CA.
As declines in telephone response rates continue, the need for flexible data collection strategies adaptable to a rapidly changing environment is clear. Advances in technology during the past decade have allowed Internet surveys to become a useful data collection strategy. While concerns surrounding the representativeness of internet surveys among the general population exist, the potential benefits; absence of interviewer bias, convenience, lower costs, and the use of multimedia present an attractive option for collecting data from longitudinal panels and list samples.A pilot project was conducted to gain insight into collecting data from youth online. The study examined barriers, advantages and the feasibility of converting a longitudinal panel of youth from telephone surveys to online surveys. After completing two longitudinal telephone surveys, 84% of respondents agreed to participate in an Internet Panel. Despite this apparent interest, we were only able to complete on-line interviews with approximately 54% of respondents.The research presented here will examine whether responses to items in the telephone survey are significant predictors of participation in the web survey. Of key interest are variables which capture information about a respondent’s willingness to use the internet and existing habits, such as daily use of a PC, time spent on the internet, connection speed and type and where they access the internet. We would expect to see strong relationships between a respondent’s perceived ability to use the internet and their participation in the web survey. However, demographic variable such as age, race, income (if any) and parental influence and oversight may also play a role in limiting participation. By identifying predictors, future data collection efforts can be tailored to assist and encourage those who may face barriers or limitations to participating.