GVU Center's 2nd WWW User Survey
GVU Center's Second WWW User Survey

GVU's 2nd WWW User Survey

GVU's WWW User Survey Home Page

Welcome, this is the home page for the Graphic, Visualization, & Usability Center's 2nd WWW User Survey. The second survey, which was run from October 10th 1994, through November 16 1994, and endorsed by both CERN and NCSA, is now over. Four sets of surveys were run: General Demographics, WWW Browser Usage, Authoring Information, & Consumer Surveys (pre-test). This page contains: Feel free to join our WWW Surveying Mailing List to keep up to date with the latest survey information (expected to be low volume). Additionally, we'd like to extend our special thanks to those of you who participated in the surveys and made it all possible.
Survey Background - The Web is one of the fastest growing Internet resources, both in terms of the number of users and the number of servers (ref: Merit NSFNET statistics). Yet because of its distributed, global nature, very little is known about its users, their characteristics, and why they are using the Web. A better understanding of these users can lead to improved development of Web-related tools and technologies. Towards this end, we decided to conduct the WWW User Surveys. Rather than using Newsgroups or cluttering up email, we felt that the most natural approach would be to use a point-and-click, graphical user interface. Quite logically, we used the Web. This way, users could complete the survey at their own convenience, and answer questions in a low-overhead fashion. The First User Survey was conducted in January of 1994, and received over 4,500 responses. Since these results proved to be of great interest to many people, in many different sectors, we decided to conduct the Second WWW User Survey. Back to the top
Experimental Confounds - Highly distributed, heterogeneous, electronic surveying is a new field, especially with respect to the Web. Our adaptive WWW based surveying techniques are pioneering and as such, require conservative interpretation of collected data due to the absence of time-tested validation and correction metrics. Basically, our survey suffers two confounds: sampling and self-selection. Random selection as a means of sampling is intended to ensure equal representation amongst populations. While our survey did not randomly select users for participation, we note that the survey was advertised in the following diverse mediums: One could argue that this diversified exposure minimizes any systematic effect introduced via the sampling method. While this constitutes a valid argument, we remain unconvinced that our sampling methodology is optimal and welcome suggestions and further comments on this subject. To futher study this effect, we are planning on adding a question in the next round of surveys quering the respondent as to how they found out about the survey, followed by stratification analyses. The second confound is of self-selection. Essentially, when people decide to participate in a survey, they select themselves. This decision may reflect some systematic selecting principle or judgement that effects the collected data. Initial cross-validation is being performed and seems to indicate a negligible self-selecting bias. We are devising ways to minimize this effect in future surveys and openly welcome helpful suggestions. Finally, we note that these confounds do not invalidate our survey, but place contraints upon the generalizability of the results to the entire WWW population. Back to the top
Technical Details & Results - The second surveys utilized WWW technologies to enable 'adaptive' questionnaires. That is, answers provided to certain questions determined the next series of questions. Below are pointers to the technical schematics of the survey implementation as well as preliminary analysis of the data for the General Demographics, WWW Browser Usage, and Authoring Information Surveys. The below information is also available as GVU Technical Report GIT-GVU-94-40 Using the Web as a Survey Tool: Results from the Second WWW User Survey in Postscript.

Consumer Pre-Test Results - As per our philosophy of cooperating with other researchers on survey content and topics, the Consumer Survey Pre-Tests were incorporated into the WWW User Surveys. Developed by Dr. Sunil Gupta of the Michigan Business School (sunil_gupta@ccmail.bus.umich.edu), pre-tests of Web users' attitudes, usage, perceptions and preferences were designed to better understand current and potential customers of Web vendors. The Consumer Pre-Test Results are available in Postscript and HTML versions. Back to the top
Graphs & Tables of Results - Index of graphs and tables of results for all questions in the General Demographic, WWW Browser Usage, and Authoring Information surveys. Results for each question are presented in multiple forms - pie graphs, column plots, and tables - to facilitate data interpretation. All plots were created using Delta Graph Professional for the Macintosh, with input from summary statistics processed via S-plus for UNIX. A tar'd file of the graphs is available via ftp.cc.gatech.edu in /pub/gvu/www/survey/survey-09-1994/graphs.tar. Back to the top.
Collected Datasets - As part of our commitment towards the WWW community and it's success, we are making available the entire datasets for this set of surveys. This also enables specialized analysis for those whose needs exceed our analysis. The datasets are however subject to certain copyright restrictions. A tar'd file of the datasets is available via ftp.cc.gatech.edu in /pub/gvu/www/survey/survey-09-1994/datasets.tar.Z. Back to the top.
Original Questionnaires - These pages simulate the surveys, i.e. no responses are being logged. Some of the questions in the below surveys are 'adaptive.' That is, responses to these questions trigger subsequent questions to be asked. An example of adaptive surveying can be found in the first question of the demographics survey (Which browser do your primarily use?). This adaptive technology enables fine grain data capture while minimizing cognitive load on the respondent and decreasing the screen demands for all possible choices - all in a point and click GUI interface! The surveys also enforce question completion. The idea of adaptive questioning is not new - it occurs all the time in face-to-face interview - Educational Testing Services (ETS) currently uses it for GRE testing - & John Mallery's Communications Linker System also employs similiar functionalities. In fact, it was John who initially explained hierarchial questioning to me at the First International Conference on the World-Wide Web. Back to the top.
Other Information Back to the top.
For more information or to submit comments:
send e-mail to www-survey@cc.gatech.edu.

GVU's WWW Surveying Team
Graphics, Visualization, & Usability Center
College of Computing
Georgia Institute of Technology
Atlanta, GA 30332-0280