|Also available as an Acrobat File|
Review of graphical environments on the WWW as a means of widening public participation in social science research
The following discussion is based on information gathered during the research, and preparation of the case studies, and on the authors' knowledge of the subject area. The discussion is broken down into five basic areas: graphical and technical requirements of participatory systems; the potential future role of participatory systems; practical and theoretical difficulties; the need for new visualisation/graphics tools; and social, cultural, political and ethical questions.
5.1 Graphical data and technical requirements of participatory systems
Technically the web and the tools that have been developed specifically to utilise it (e.g. Java and VRML) open up a range of possibilities for graphical interaction with data and information. Everything within the field of graphical exchange of information seems possible, but is on the other hand hampered by practical considerations of restricted band-width, speed of access and hardware/software limitations faced by the client (user). This has significant, if mundane, implications in terms of cost and accessibility to the public user.
Current survey work on the Internet is limited by the software available. Most survey sites use simple forms which are extremely basic graphically, making use of CGI scripts to submit responses in an electronic mail format or as a text file. Advances in this area are likely to make use of the Java programming language mainly due to its platform independence and advanced features allowing greater interaction between the respondent and the survey (see WWW Survey Assistant web pages). Current problems may lie in a lack of understanding with regards to CGI and PERL programming. In a similar way to web authoring only three or four years ago when there were limited numbers of HTML (Hyper Text Mark-up Language) editors around, similar problems with regards to CGI and PERL editors are likely to be overcome. As more custom built and easy to use on-line questionnaire packages become available which automate the CGI and PERL scripting processes, increased use of on-line surveying techniques is likely.
Perhaps the two key aspects regarding the graphical data content and technical requirements of participatory systems are accessibility and usability. There is a great deal of social science relevant data in existence, while the technical and graphical capabilities of the WWW far surpass anything available ten years ago. At the same time the WWW is hugely distributed; providing networked access to vast amounts of information and computing power from almost anywhere in the world (or absolutely anywhere if you have a portable PC linked to a satellite telephone). The possibilities for the WWW in changing the way people use computers and information may seem almost limitless. Web sites need to be written in standard code as not all browser software supports all formats or the latest versions of HTML, VRML or Java.
Web sites that are full of graphics in the form of banners, icons, wallpaper, images and animation's may look pretty and can at first attract users. However, the graphical content of the web can be a victim of its own success. True, a picture is worth a thousands words, but at what cost in terms of download time? Clients are more likely to quit and jump to another web site if the graphics on the current site slow the loading times to minutes rather than seconds. The average human attention span is limited, but when the client is being charged local call rates for the privilege then it is severely curtailed. Web designers therefore need to be careful in establishing trade-offs between graphical design and graphical information and tread a fine line between making a web site attractive and keeping clients on-line.
Hardware advances, improved bandwidth and reduced (or even free) access rates may help reduce or eradicate this particular problem in the future. In North America local calls are free, making the access times less of an issue. In the UK developments in the near future may also help remove this barrier to more widespread use. These include plans for access to the Internet via digital and cable television networks and the possibility of electricity utilities acting as service providers using their own distribution cables.
The potential future role of on-line participatory environments in gathering large amounts of data on both general and specific social science topics is real, but may be misleading within the context of this report. The WWW potentially gives social scientists access to human subjects in every household with an Internet connection, making it a potentially huge source of social science data. Current surveys put the number of households with an Internet connection at between 5% and 20% (NOP, 1997). However, despite these encouraging figures, the public have to want to contribute/participate in web-based social science and cannot be forced to fill in on-line questionnaires or use social science web pages. Access to web pages is, like watching television, entirely voluntary. Just as with television thirty years ago, however, market saturation is just around the corner for the WWW or whatever succeeds it, meaning that the potential (however fraught with difficulties) cannot be ignored.
Problems with sample bias
A significant problem facing the use of participatory systems on the WWW to gather social science relevant data is that of sample bias. It can be safely assumed that in most cases those members of the public accessing a social science-based web site are doing so because they want to. In other words, they have not found the web site by chance (though that is possible) nor have they accessed it through a sense of duty to support social science research, but will have actively sought it out because they have a particular interest in the topic covered. This can create critical problems over bias in the sample generated in any on-line surveys used to address topics and issues requiring a stratified sample across the whole population. Surveys of Internet users suggest that those individuals most likely to be represented are young males, further complicating the bias in the list of likely participants. Those individuals most likely to be under-represented include older people, females and those from lower income groups that are less likely to be able to afford a home PC. In other cases, this bias may be a distinct advantage. Returning to the analogy with television, there is some suggestion that "surfing the web" may have replaced channel hopping in some sectors of the population (Internet service providers have noted marked drop off in service access at times when TV programmes such as the X-files are screened). Bias towards this kind of user may be interesting for certain ethnographic/behavioural studies.
The role of schools and higher education
School children and students in higher education represent a significant user group for particular studies. All higher education institutions have access to the Internet and WWW, and it will not be long before the same situation exists in primary and secondary education. Under these circumstances all children of school age and a significant number of young adults (those in higher education) will have direct access to the web via their place of learning. For certain social science applications this may be advantageous, in particular given the view that today's school children are tomorrow's voters, business people, tax payers, etc. Direct, on-line targeting of these groups by participatory systems (with curriculum assistance from teachers and lecturers) is therefore directly relevant to studies of future society. One further point is that if all schools have web access, then many households have indirect access through their children, further broadening the outreach of the WWW.
The web as a source of secondary data?
The WWW in its wider sense may also present opportunities for gathering potentially massive amounts of social science relevant data from secondary sources. The access logs of web pages whose primary purpose is not social science research may yield data of interest to social sciences. For example, social scientists interested in politics may glean useful data from the access of logs of web sites with a political theme. This route to gathering useful social science data is, however, fraught with difficulties. In particular the use of data from web access logs for purposes other than server monitoring may be in contravention of data protection and privacy laws. In many cases it may not be possible to identify individuals without access to commercially valuable and sensitive information about individual clients held on service provider databases. In essence, using web page access logs to generate massive amounts of social science relevant secondary data is a nice idea, but it may never work for all but a limited number of special cases.
Reviving old areas of social science research
The WWW presents some limited opportunities for reviving and reviewing old areas of social science research with much larger sample sizes and/or populations. The problem here is not just ensuring a representative sample, but capturing a large enough sample size. Some mechanism other than personal interest and chance are required to attract more than just the casual surfer. Media science events such as the British Association for the Advancement of Science (BAAS) meetings and the Tomorrow's World MegaLab experiments might prove suitable vehicles in advertising a social science web presence and attracting greater response rates.
Developing new areas of social science
Some areas of social science research that have hitherto been unresearched because of a lack of access to suitable subjects and/or insufficient sample sizes may possibly be opened up through recourse to on-line participatory systems. One obvious example used already in this discussion is that of the behaviour and ethnography of web users. A great deal of potentially interesting research is still be carried out into user attitudes and perceptions, social interaction versus isolation, 'netiquette' and on-line behaviour, virtuality, community development and teleworking, among others.
One area where the web can make considerable inroads is in improving public access to software and data tools developed out of social science research. A number of examples of social science relevant software tools are already on the web. Good examples include the Friends of the Earth Chemical Release Inventory (FoE CRI), the Institute of Fiscal Studies' (IFS) Be Your Own Chancellor site, Stan Openshaw's Geographical Analysis Machine (GAM) cluster hunting software and now many examples of Geographical Information Systems (GIS) demonstrations such as Berkeley University's Research Programme in Environmental Planning and GIS (REGIS). Essentially such web sites provide two valuable public services; making software and data publicly available, and making data useful. This important function serves as a public window on social science research.
Graphics Multimedia Virtual Environments Visualisation Contents