The fact that survey respondents do not attend to all the attributes presented in choice experiment surveys is fast becoming a key issue in CE studies. This study proposes a new method aimed at eliciting consumers’ stated attribute non-attendance (ANA) behavior at the levels of each attribute, and compares it with the commonly used stated ANA approach, where non-attendance behavior is captured at the attribute level (i.e., not levels of the attributes). Results generally indicate that respondents do indeed ignore some of the levels of an attribute, suggesting that capturing non-attendance behavior at the attribute level would be insufficient at accurately and totally capturing stated ANA behavior. This finding implies that future choice experiment studies should take ANA into account not only at the attribute level but also at the levels of an attribute, both when asking respondents their stated ANA behavior during the survey and also in CE model specifications.