Academic journal article The Public Manager

Conducting Web-Based Surveys: Keys to Success

Academic journal article The Public Manager

Conducting Web-Based Surveys: Keys to Success

Article excerpt

Agencies can successfully conduct surveys over the Internet, but careful execution is required for sound results.

Use of surveys in government to gauge performance and improve operations is widespread. Managers must either understand survey complexities enough to oversee work done in-house or, more frequently, to monitor the work of a contractor. This report of our experiences will help managers obtain sound results from their survey efforts.

Three Phases of Survey Projects

In general, survey projects have three basic phases. The first is the planning stage, which involves designing and pre-testing a questionnaire and determining the best way to conduct the survey. The second phase--conducting the survey--consists of several steps. They include acquiring a list of the survey population, sampling, sending survey communications, and monitoring the responses. In the third phase, the data are recorded from the questionnaires and the results are analyzed and presented.

Increasingly, the Internet is being used as the medium for conducting surveys. While the debate continues on whether the Internet can be used for general populations, it has been widely accepted as a medium appropriate for special or known populations. Web-based surveys can be used with a variety of populations of concern to government:

* external customers (vendors, for example);

* internal customers (program managers, for example);

* managers and supervisors; and

* employees.

The Logistics Management Institute (LMI) has successfully used the Internet to survey over 30,000 government managers, supervisors, employees, and internal customers during the past two years (Schwartz, 1999).

Planning/Design Phase

Computer and Internet Access

Government employees are using computers and the Internet on a regular basis. In our recent survey work with government agencies, we have found that virtually all employees are using computers as part of their daily work and have desktop access to the Internet. Thus, the basic issue of comfort with a computer is generally no longer a deterrent with government employees and their internal customers. However, as discussed below, some users still may be unfamiliar with special conventions such as drop-down boxes.

It is important to note that for some agencies, using mixed survey media may be more appropriate. In one project that included industrial-type operations in which some employees lacked access to computers, we used paper surveys for those employees; others with e-mail and Internet access participated on the Web.

Identification of Participants

While the Web provides many benefits, several considerations are involved in selecting it to be the survey medium. The first consideration is identifying the participants. Just as with postal mail surveys, a mailing list needs to be generated in order to reach potential respondents. With Internet-based surveys, we have primarily used e-mail to urge people to take the survey and to prod nonrespondents. Although one would think that a copy of any organization's email address list would be readily available, we have often found that not to be the case. As a result, one of the earliest tasks is to collect the e-mail addresses and determine Internet access. The consolidation of the organization's email list should be done as early as possible--during the survey design phase--to avoid delays.

Unlike postal mail, the Web environment is unforgiving of the slightest deviation in specifying addresses. A misplaced letter, period/dot, or @ symbol in an e-mail address can keep a potential respondent from receiving correspondence. A clean mailing list is essential for selecting and reaching all potential participants.

A sufficiently bad postal address usually results in the item being returned to the sender. E-mail communications generally work the same way. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed


An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.