|Also available as an Acrobat File|
Review of graphical environments on the WWW as a means of widening public participation in social science research
4.2 Case Study 2: A Review of Web-based Tools for Creating Questionnaires and Inputting and Analysing Survey Results.
Due to a lack of web-based social science research which make extensive use of graphics to gather survey data it is felt appropriate to review software tools which are starting to become available on the Internet which automate the survey process. Two software tools are discussed which provide solutions to allow web-based surveys to be undertaken without having to gain an understanding of the computing languages which many of the current web-based surveys require.
If the future use of the web for carrying out social science surveying is to expand the development of software for the automation of Internet based research surveys needs to be made available. the majority of web-based surveying systems requires the use of CGI and PERL programming (see Section 5.1) which more often than not requires some knowledge of computing languages and scripting. If a researcher wishing to undertake a web-based survey has no or little knowledge of this, the likelihood of not using the web to conduct a survey will be increased.
The WWW Survey Assistant package which has been developed by Schmidt came about due to some of the problems associated with CGI programming which have been outlined above. As Schmidt realised the problems many researchers would encounter when attempting to develop on-line questionnaires.
"WWW Survey Assistant came about as a by-product of my efforts to implement WWW surveys for one of my students. I realized that the process could be automated in order to provide professionals and amateurs alike with wider access to survey research via the web. Nobody would have to go through what I went through: WWW Survey Assistant is my answer to increasing accessibility." http://or.psychology.dal.ca/~wcs/SAdocs/about.html
The software creates a HTML web form to display the survey to be administered and then automatically generates a CGI program to run the survey therefore eliminating any problematic programming which may put people off from carrying out web-based surveys. This can either be installed on Schmidt's server or your own. The software has options to dynamically reply to user's responses and can handle dynamic forms. Once the questionnaire is finalised and contains the necessary questions the software package first writes an HTML document which will present your questionnaire then writes a PERL CGI program which is then executed each time a user submits a response to your survey through their WWW browser. The software contains many features which allow the researcher to tailor the questionnaire for specific tasks which overcome many of the problems which are often associated with web-based surveying. The software can filter out respondent's re submissions thus preventing biases in responses. Specific domain names can be incorporated so that access to respondents from the specified domains only can participate in the survey. Essential questions for which the researcher requires a definite response can be set up in a way that the questionnaire cannot be submitted if these are left blank. The software also has capabilities to present sound and images within the survey providing the server which is running the questionnaire contains the appropriate software.
The WWW Survey Assistant software is available from the company's web page at: <http://or.psychology.dal.ca/~wcs/>.
A similar software package which uses the Java programming language is also available from Senecio Software. This benefits from platform independence and therefore problems associated with different client operating systems is over come so long as they are operating Java enabled browsers. Due to its use of Java this software appears to be a little more advanced than the previous package. This allows the researcher to create questionnaires faster and provides more control over the respondent's questionnaire. The software can prevent the respondent from looking ahead in the questionnaire or leaving questions unanswered which often occurs in web-based surveys. The questionnaire can be made randomly to present questions and response categories to each respondent which can be an important factor when trying to reduce the bias from order-effect. The software can deal with adaptive surveys, where previous answers are dynamically inserted into later questions and response categories, or where behind the scene calculations are incorporated into questions and skip logic.
Other features of this web-based surveying software package include the ability to create a questionnaires with contingency skips, adaptive fills, calculations, analogue scales, pictures, rankings and many other necessary survey features which are all integral to the programs environment. Rapid changes to the questionnaire are straightforward with easy to use page layout environments and the simplification of surveying on the web by eliminating the need for extensive CGI and PERL scripts.
Senecio's On-line Survey Product, the MaCATI - ePoll software is available from the company's web page at: <http://www.senecio.com/>.
An on-line tutorial for students at Plymouth University provides an example of how psychology students can put their own surveys on the web with relative ease thanks to the CGI scripting being continued within a programme called Polyform which executes the scripts that deal with data sent back by the forms that users fill in. Students receive questionnaire responses in the form of an e-mail. The URL for the tutorial is: <http://salmon.psy.plym.ac.uk/mscprm/forms.htm>.
Graphics Multimedia Virtual Environments Visualisation Contents