A lot of webmasters invite their visitors to respond to a website exit survey in order to evaluate visitor experience of their site. Of course, some visitors do not accept the invitation. As a result, the collected data are susceptible to nonresponse bias. In this paper I try to provide answers to the following questions: Is the sample collected with an exit survey representative of the population of all website visitors? Are there any differences between the sample and the population regarding visitor’s sex, age and education level? Can we use website exit surveys to estimate the users’ satisfaction level? Are we going to get answers only by those users who were satisfied by the website and as a result, will we overestimate the satisfaction level? The findings presented in this paper are based on the analysis of data collected by a Greek voting advice application and the analysis of data collected by the corresponding exit survey. The comparison between the two datasets provides some interesting answers to the aforementioned questions.