There is no doubt that web surveys have many advantages for researchers: lower cost, less manual handling and increased speed. However, for the respondents there are significant drawbacks. These drawbacks often lead to design recommendations to limit the number of questions to 25 or 2-3 screens. Obeying these recommendations would make computer-based self-administered surveys practically impossible to use for most research. For example, in Epidemiology, the number of questions in a questionnaire can be over one thousand and can easily take more than two hours to answer.

So, the question is: how should we make people spending their leisure time sitting in front of a computer for several hours in a row? The answer is: use the same design techniques the computer game industry already has explored for years. They can tie gamers to their computer for days, and so should we.

As a first attempt to include computer game design in a survey we have implemented personalized feedback in a study on physical activity and nutrition (Bälter et al. 2005). The respondents were given feedback on their energy expenditure, intake of vitamin-C, Calcium, Iron and Fibre, body mass index and meal composition. This increased the response rate compared to a paper version of the same questionnaire. The survey was directed to the general population aged 20-60. On the first part of the questionnaire the response frequency was higher for the paper version, but no more than expected due to the loss caused by lack of Internet access among the respondents. On the second part of the questionnaire, there was practically no difference in response rate between paper and web. That is, with the design used, there is no loss in response rate compared to paper if you use the web for long questionnaires.