Survey Design Web Surveys

Web Surveys The public did not have widespread access to the Internet and e-mail until the end of the 1990s. For example, in 1994, only 3% of the U.S. population had e-mail at home or work; by 2007, 62% of households had both e-mail and Internet connections. By 2012, some projections suggest that 77% of households will be connected. Internet connection rates are higher in other nations, for example, 97% in South Korea, 82% in the Netherlands, 81% in Hong King, 79% in Canada, and 77% in Japan.

Advantages Web-based or e-mail surveys are very fast and inexpensive; they allow flexible design and can use visual images and even audio or video. The two types of Web surveys are static and interactive A static Web or e-mail survey is like the presentation of a page of paper but on the computer screen. An interactive Web or e-mail survey has contingency questions and may present different questions to different respondents based on prior answers.

Disadvantages An unusual disadvantage of Web surveys is that they are cheap and easy. As Weisberg (2005:38) remarked: “Putting a poll up on the Internet can be inexpensive, so many groups put up polls without paying attention to quality.” Web surveys have three disadvantages or areas of concern: coverage, privacy and verification, and design issues.

Coverage The first concern involves sampling and unequal access to and use of the Internet. Older, less educated, low-income, and rural people are less likely to have access, and a majority without access now say that they do not plan to acquire it in the future. In addition, many people have multiple e-mail addresses.

Privacy and Verification A second concern involves protecting respondent privacy. Secure Web sites with passwords or PINs and high confidentiality protection can help. Respondent verification is needed to ensure that the sampled respondent alone participates and only once.

Design Issues For example, it appears A third concern involves design complexity and flexibility. The compatibility of various Web software and hardware combinations must be verified. We are just beginning to learn the most effective way to design Web surveys. For example, it appears best to provide one or a few questions per screen best to making the entire question visible on the screen at one time in a consistent format with dropdown boxes for answer choices

Design Issues best to include a progress indicator (as motivation) such as a clock or symbol indicating progress (how far respondents have gone and how much questionnaire remains). Keeping visual appearance simple (limited colors and fonts) and maintaining consistency is best. Very clear instructions are needed for any computer action (e.g., use of drop-down screens) and they should include “click here” instructions. Also, making it easy for respondents to move back and forth across questions

Design Issues Providing detailed questions and a large space for answers in open-ended questions on Internet surveys helps elicit longer and more complete answers Avoiding technical glitches and “bugs” at the implementation stage with dedicated servers and sufficient broadband to handle demand is important

Thank you!