Residential satisfaction is often seen as an indicator for quality of life, which makes research into people’s neighborhood preferences an important topic. Traditionally a stated preference study (SP) contains written attributes (text), which leave the respondent to make mental images for him- or herself. With current progress in virtual reality (VR) applications more SP surveys are performed with the use of multimedia. This study adds to the neighborhood preference literature, as well to the discussion whether multimedia application in SP surveys has added value. In this study six neighborhood attributes have been used and presented to two different respondent groups. One group filled in the conventional text-only experiment, while the other group carried out the video experiment. Both experiment designs used a ten point rating scale to express neighborhood preference and were analyzed using an ordinal regression.
In total 215 respondents completed either one of the survey types. The overall model fit of the text-only experiment was better compared to the video experiment. Additionally, respondents slightly enjoyed the text-only experiment more. In both experiment types the attribute ‘green type’ scored almost 50% on relative importance of the whole attribute set. The other five attributes showed a different importance order between the two experiment types. The results indicate that respondents that filled in the text-only survey thought more about how important they considered the individual attributes, while in the video experiment the larger visual attributes gained importance. This raises the question if the application of multimedia may lead to incorrect preference scores. Researchers need to be very careful when using a multimedia survey and evaluate if using multimedia has added value for the specific research they want to perform. For overall preferences the use of a text-only SP survey should be sufficient. The pitfall in using multimedia is that attributes and levels can be shown in multiple ways, which can deviate from how the neighborhood is actually going to be. Therefore, using multimedia is advised only in situations when the representation of the attributes is the same as how they are going to be made in the real neighborhood, or when the attribute cannot be explained in a few words (text).