Designers of stated preference studies have placed an emphasis in recent years on ensuring that questionnaires are defensible, and that all ‘hypothetical’ elements are removed. A potential problem with this emphasis is that it can unwittingly increase the hypothetical nature of the survey as well as necessitating the use of ethically questionable statements. An alternative approach was recommended by Morrison (forthcoming) that is ethically better and potentially less susceptible to hypothetical bias. This approach has been used in several studies, with the results indicating that designing questionnaires in an ethically neutral manner does not automatically lead to poorer quality models. In this paper we present the results of a more rigorous split sample test to test the appropriateness of using this approach. Minor evidence of strategic behaviour by a small proportion of the respondents (about 7%) was identified; however the results indicate that welfare estimates were not affected by designing questionnaires in this way.