Like wind power, they just seem to make good sense. Web surveys are green, inexpensive, and fast—and anyone can create one. However, when you look deeper into the technology, citizen surveys on the web remain a heavenly idea with a devil of a road to practicality.
Posted link for opt-in response. A survey link displayed on your jurisdiction’s home page will attract some respondents, but those people should not be mistaken for your typical citizen. Opt-in web respondents have computer access (true, most people do, but not everyone, and especially not groups you may be more interested in hearing from) and opt-in web respondents tend to be more opinionated and critical of local (and all other) government. With the posted link, respondents have found you and the edge on which to sharpen their disaffection. Results of this type of survey won’t represent the whole community, so it will be hard to tell council what the results mean. Input, yes. Representative public opinion, no.
Invited web survey. As with surveying by mail or phone, invited web surveys require that you think about issues of sampling and representativeness. The goal is to ensure that all members of the community have a chance to be chosen for the survey, and because there is no complete list of all email addresses within a given geographic area, you have to recruit residents for your web survey by mail or phone. Since relaying a web site URL over the phone is complicated, expensive, and leads to error, mail surveys are the best vehicle for delivering the invitation to the web survey. Mail delivery has great geographic precision, includes all household residents in the community, and is relatively inexpensive. When the invitation is a “stand-alone,” mailed contacts typically are postcard and letter notifications with reminders.
Although this method of web surveying has the greatest potential to include a representative sample of residents, response rates are low. Still, if the choice is between a phone survey and a stand-alone web survey, web survey response rates are not much worse, these days, than phone—stand-alone web response rates from a mailed invitation have been increasing, receiving about 13%-18% response in National Research Center, Inc.’s (NRC’s) recent testing, while phone response rates have dropped to 12%-25%.
Web survey option with citizen survey delivery. More common than stand-alone web surveying is to permit web responses as an alternative to returning print surveys delivered by mail. For this option, residents receive all materials in the mail, but are able to choose whether to type in the web address included on the survey cover letter or to fill in the print survey and return it in the postage prepaid envelope. NRC’s research has shown that providing this option doesn’t improve a jurisdiction’s response rate but it doesn’t hurt, either. The mailed citizen survey with the web response option gives potential respondents who like the web their preferred method to participate and it helps to characterize the jurisdiction as one with forward-thinking leaders. (The web continues to carry the cachet of the future.) Among the three main citizen survey data collection methods (mail, phone, or web with mailed invitation), mailed surveys still get the highest response rates (typically 25–40%).
NRC pioneered the use of the Internet for citizen surveying (Citizen Surveys on the Web: General Population Surveys. Miller, T. Kobayashi M. Caldwell E. Thurston S. and Collet B. Social Science Computer Review. SAGE publications. 2002, vol 20 (2): 124–136) and continues in-house research and testing in order to provide the best guidance regarding web surveys. NRC continues to test and develop web surveying options to turn the dream of web surveying into reality for local government surveys.
NRC currently is working with a professor at Rutgers University to compare the response rates and responses to a national opt-in web version of The National Citizen Survey™ to a national mailed survey with the same content. For now, recent improvements in web-only responses rates suggest a possible shift in residents’ willingness to respond online. In the meantime, including a web address along with a mailed survey is the best option.
This article was written by Thomas I Miller, president and founder of National Research Center, Inc. and Shannon Elissa Hayden, and first appeared in Perspectives, Volume 9, No. 3., a newsletter about survey research for local government managers and elected officials. Tom and Shannon co-authored Citizen Surveys for Local Government: A Comprehensive Guide to Making them Matter, published in 2009 by ICMA.
To learn more about The National Citizen Survey, ™ visit icma.org/ncs.