Traditionally, hypothetical evaluation methods such as Stated Preference analyses (SP analyses) have been used to identify the preferences of the various passenger groups for non market goods such as travel time and comfort. Evaluations from SP analyses are used in the planning and decision making process in the transport and communications sector, for example in cost benefit analyses. Correct evaluations are therefore important in order to make the right planning decisions.

During the last years Institute of Transport Economics (TOI) has carried out several Internet based SP-surveys to find passengers’ preferences. This paper will summarise some of the experiences we have had with Internet based SP analyses in order to look at the differences between data collected through self-administered Internet surveys and data from paper forms or home interviews.

The analyses are based on five different SP surveys. The recruitment strategy varied from survey to survey. We have recruited respondents by post, telephone, e-mail or by handing out cards. The response rate varied from 16% to 72% and depends on the target group and how they react to various different recruitment strategies.

In all the surveys there are two ways of collecting the data material, self-administrated Internet interviews and pen-and-paper or Internet based home interview. We have analysed if there are any differences between respondents who choose one method rather than another. We find that women choose pen-and-paper or home interview, average age is higher amongst those who choose pen-and-paper or home interview and those in employment choose self-administered Internet. Use of public transport, holding a driving licence and access to a car cannot explain the choice of method beyond that which is explained by employment.

Comprehensive SP surveys to find passengers’ preferences are relatively new and it is interesting to see that Internet based home interviews replace pen-and-paper. By carrying out Internet based home interviews we catch respondents who don’t prefer self administrated Internet survey and at the same time making good use of the advantages of Internet based studies.

In order to see whether those who choose self-administered Internet have other preferences to those who choose pen-and-paper or Internet based home interview, we have run several analyses of the stated choice-sequences which calculate separate estimates for each method. These estimates are then compared to see whether the different methods give different estimates which, in turn, lead to different valuations. We have also carried out calculation where we have taken into account the proven sample bias in gender, age and employment by weighting the data material with regard to these factors. This is done to see whether we can find an isolated method effect.

We find fewer differences when two Internet solutions are combined compared to combination of Internet and pen-and-paper. And weighting has greater effects on surveys where the choice is between two Internet methods.

When analysing our five surveys we concluded that Internet is an effective method in complicated studies which require advanced or tailor-made designs, studies which require a large sample, when large elements of the target group have the chance to carry out a self-administered Internet interview, when the e-mail addresses of the target group are easily accessible or when instant access to data is needed.