Although measuring intrinsic quality has been a key factor in the advancement\nof Machine Translation (MT), successfully deploying MT requires considering not\njust intrinsic quality but also the user experience, including aspects such as\ntrust. This work introduces a method of studying how users modulate their trust\nin an MT system after seeing errorful (disfluent or inadequate) output amidst\ngood (fluent and adequate) output. We conduct a survey to determine how users\nrespond to good translations compared to translations that are either adequate\nbut not fluent, or fluent but not adequate. In this pilot study, users\nresponded strongly to disfluent translations, but were, surprisingly, much less\nconcerned with adequacy.\n