Abstract Multiattribute theory is an important conceptual framework for assessing recreation choice. Methods of assessing attribute importance, varying in the amount of cueing they provide, can affect study results through context effects, attribute omission‐inclusion effects, and group rating effects. Attribute omission‐inclusion effects occur when important attributes are omitted or unimportant attributes are included on a list. Context effects may result when attributes are evaluated in the context of different sets. Group rating effects occur when respondents rate attributes that may or may not be salient to their decision. Three commonly used strategies of generating attributes—labeled the researcher‐generated, the modal salient beliefs, and the idio‐graphic methods, differing in their level of cueing, were compared. Results suggest that when different attribute sets are generated, omission‐inclusion effects can dramatically affect interpretation of findings. Context effects did not appear to influence attribute ratings. Group rating effects led to differences in valence ratings of some attributes. Use of different data collection strategies, varying in their level of cueing, would lead researchers to different conclusions about what was important to recre‐ationists in their site choices.