Evaluation 2009 Banner

Return to search form  

Contact emails are provided for one-to-one contact only and may not be used for mass emailing or group solicitations.

Session Title: Hot Topics in Surveying/Polling: Mixed Phone, Web, and Mail Surveys, and Barriers to Randomization in Mobile Polling
Multipaper Session 811 to be held in Suwannee 15 on Saturday, Nov 14, 11:50 AM to 12:35 PM
Sponsored by the AEA Conference Committee
Chair(s):
Britteny Howell,  Cincinnati Children's Hospital Medical Center, britteny.howell@cchmc.org
Mixing Modes to Maximize Results: Combining Telephone, Web and Mail Surveys
Presenter(s):
Julie Rainey, Professional Data Analysts Inc, jrainey@pdastats.com
Lija Greenseid, Professional Data Analysts Inc, lija@pdastats.com
Abstract: "Mixed Mode" has become a buzzword in survey research circles. What is a mixed mode survey? Why would you want to use one? This presentation will highlight the benefits and challenges of mixed-mode designs, and provide criteria to consider when selecting survey modes that are best suited to a particular population or situation. We'll compare the advantages, costs and response rates of telephone, web and mail surveys, and describe common pitfalls to avoid during survey implementation. Attendees will understand issues to consider when selecting an appropriate survey mode or modes, choosing which mode should come first, developing strategies to do critical data tracking, and planning to avoid mode bias.
Practical Barriers to Effective Randomization During a Pilot Project on Mobile Polling
Presenter(s):
Jonathan Rubright, University of Delaware, rubright@udel.edu
Kristin Harkins, University of Pennsylvania, kristin.harkins@uphs.upenn.edu
Charles Sabatino, American Bar Association, sabatinoc@staff.abanet.org
Deborah Markowitz, State of Vermont, dmarko@sec.state.vt.us
Robert Boruch, University of Pennsylvania, robertb@gse.upenn.edu
Jason Karlawish, University of Pennsylvania, jason.karlawish@uphs.upenn.edu
Abstract: In this paper, barriers and solutions to effective randomization will be discussed in the context of an evaluation that assigned nursing homes to a mobile polling program. The evaluation took place in Vermont and was conducted to examine whether mobile polling reduces between-site variance in the proportion of residents who vote. Methods involved matching sites according to number and cognitive status of residents. Twenty eight nursing homes met selection criteria and were placed in matched pairs and randomized within pairs. Randomization was threatened by problems with implementation at the site-level and by the election clerks, some of whom were unwilling to partake in this demonstration project. Of those nursing homes that were involved, a few refused to provide full data on their residents. Small sample size, along with validity threats of comparing voting rates between sites, required a greater emphasis on survey and qualitative data than initially planned.

 Return to Evaluation 2009

Add to Custom Program