Printable Version Email to a Friend |
Online Polls and Surveys
The need for nonprofits to collect information and survey others is critically important, especially in research and message development activities. There are usually, however, two limitations: cost and expertise. A number of online tools and services exist to make it easier and less costly to do surveys and polls.
There are two catches to these services. First, the free access to tools comes courtesy of banner advertising, either through the host server, or through code that is inserted into your website. Second, the actual poll results and data are hosted by an outside service, although you can manage your polls and generate reports at your convenience.
FreePolls
You can set up free instant feedback forms, single issue polls, or multiple item polls of up to ten questions and 15 answers each on your website. You define what items or questions to have visitors vote on or respond to, and the service generates HTML code or JavaScript which can then be inserted into any existing web page. Results can be viewed graphically by other visitors, and the service lets you manage your polling via the web.
Pollit.com
This is a service that lets you insert polls and surveys into your existing website for free. You create a poll or survey from among 12 options, with up to 25 answers. The service then generates the HTML code you need for your web page. You can choose for the survey to be generated through your existing web page, or as a separate "pop-up", and have a large degree of flexibility in determining the look and feel, through an online editor. More significantly, the service can block duplicate voting by either using cookies, tracing the IP address of particpant machines, or a combinations of both. You can have voting results displayed in real time in a variety of graph formats.
Zoomerang This is a free online service which hosts online surveys and databases. Zoomerang differs from Flashbase in that it is specifically a survey research tool as opposed to a general purpose online database hosting service. You can select from over 100 customizable survey templates or create your own from a blank form.
After editing your form, you can then elect to make your survey available as a web-based or e-mail-based survey. While you are limited to 20 survey questions per form, you can choose from 12 different types of questions, including matrix scales, open response, and date/timestamp formats. Three limitations if you are using the free version of the service: (1) you cannot download survey results to your machine, (2) you cannot link back to your organization's web page from the survey; and (3) you can only store your results for 30 days. If you conduct an e-mail based survey, however, you can compile a personal and confidential contact address book for later use.
Since the services are free, there are some caveats. Free accounts can only receive a limited number of responses, and must contend with banner ads at the top of the screen. Also, data and responses are stored on the respective companies' servers, however you access and download data from any machine with a web browser. Paid subscribers can access a range of higher end customization tools and advanced abilities like cross-tabulations.
Will Online Polling and Surveying Replace Traditional Polling?
In late September 1999, the New York Chapter of the American Association for Public Opinion Research conducted a forum on "The Perils and Potential of Online Polling." Among some of the benefits cited were: access to a varied and sizable population that actually wants to be polled, the capacity to survey large numbers of people simultaneously, the ability to use more sophisticated tracking tools, and innovative opportunities to conduct more detailed focus groups.
The chairman of Harris Interactive summed up the unique nature of online polling at the New York forum by noting that it is visual and interactive medium that allows for open-ended (and in many cases) more candid responses with a greater degree of anonymity, more than telephones do.
On the reverse side, the chairman also conceded adequate cross-samples are not guaranteed, because some groups might be underrepresented in a raw data. The argument follows that, since there are more people who have telephones than have Internet access, for example, a sample pool may not be sufficiently representative of a particular population. In addition, there is a tendency to draw more moderate or unsure responses than those reflecting a strong positive or negative opinion. Also, online surveys do not rely on random samples of e-mail addresses or on-screen identities (which can be unreliable), so they invariably involve people who self-elect to participate.
The risk, simply, is that online polling might result in a lot of data that ultimately does not yield accurate or useful results. On the other hand, even telephone surveys have their share of difficulties, especially the murky problem of non-responses that disrupts whatever sampling method you use. Also, no matter how big your sample size, you still do not guarantee that your results are more valid
Harris Interactive, however, has stated that out of some 200 parallel surveys involving polling on the same issues using both the Internet and traditional phone interviews, there was almost no noticeable difference.
Online polling and surveying was also the focus of Robert Schlesinger's January 5, 2000 piece in the newspaper The Hill. In it, he mentions how public response rates to traditional polls and surveys have fallen over the last 15 years, from a range of 55 to 65%, to 25 to 35%.
Schlesinger points to an intertwined set of influences responsible for the dip. It costs more to conduct polls today because it is harder to reach people. Many people are hard to reach because they are so used to receiving calls from telemarketers, they refuse to respond to telephone calls that sound remotely like solicitations. In addition, to meet demands to be protected from telemarketers, consumer technology advanced such that screening and blocking devices (including answering machines and caller ID) now allow potential respondents to screen out telemarketers-- and pollster calls. Pollsters now need to make more calls, and spend more time and money, to get better representative samples.
One big obstacle to widespread Internet polling firms on the landscape is the capitalization costs required to build a strong potential base of respondents for sampling and cross-samples. Harris Interactive cites that it spent some million over two years to build a respondent base of 5 million.
The cost of conducting polls or surveys, however, goes down for each activity. This can be attributed, in part, to the reduced cost of sending an e-mail request to respond to an in-person meeting, e-mail survey, or web-based poll. Compare this to printing, mailings, follow-up phone calls, etc. So instead of the traditional order of representative samples of 500-1000 responses, you might instead yield respondent bases in the tens of thousands. The larger numbers might also ensure better representation of answers in particular response subcategories.
Eli Noam, the director of the Columbia Institute for Tele-Information, points to the ultimate downside of online polling in his speech, "Why Information Technology is Bad for Democracy." Technology, he argues, allows for more voices to participate in public discourse. Yet as more and more of those voices strive to be heard, two things happen: (1) the content from those voices becomes simpler, and (2) the incidence of information overload increases. It is for this reason, that both media and politicians argue that the course of their respective institutions should not be determined by public opinion.
A number of members of Congress at the time, however, expressed disdain for polls that dictated a specific course of action, citing them as (a) the impulsive response of an ill-informed electorate, (b) not representative of their constituents' will, or (c) a political (or partisan) tool which should not be allowed to determine the course of policy.
One of the main problems with online polling can be that it becomes comingled with market research. The latter is used to gauge what news to broadcast or what votes to cast in an elective body on a constant (and now instantaneous) basis. There is also a large difference between using one poll and many surveys, as well as the frequency with which a groups is polled.
Another possible source of distrust for online polling might also be the convener of the polls. A large number of online polls are sponsored by online media entities, which might lead some creedence to the "poll as marketing tool" suspicion that exists.
As online polling and surveying becomes informed by better tools, better methodologies, and better analytical frameworks, is it correct to ask if it will serve as a more legitimate and reliable mechanism to help inform public discourse?
Links Cited
Other Resources
Resource Organizations
American Association for Public Opinion Research
American Statistical Association
Consortium of Social Science Associations
COPAFS: Council of Professional Associations on Federal Statistics
The Council of American Survey Research Organizations
NCPP: The National Council on Public Polls
WAPOR: The World Association for Public Opinion Research
The World Association of Research Professionals