AbstractThis paper discusses two types of response‐scale heterogeneity, which may impact upon the EQ‐5D. Response‐scale heterogeneity in reporting occurs when individuals systematically differ in their use of response scales when responding to self‐assessments. This type of heterogeneity is widely observed in relation to other self‐assessed measures but is often overlooked with regard to the EQ‐5D. Analogous to this, preference elicitation involving the EQ‐5D could be subject to a similar type of heterogeneity, where variations across respondents may occur in the interpretations of the levels (response categories) being valued. This response‐scale heterogeneity in preference elicitation may differ from variations in preferences for health states, which have been observed in the literature. This paper explores what these forms of response‐scale heterogeneity may mean for the EQ‐5D and the potential implications for researchers who rely on the instrument as a measure of health and quality of life. We identify situations where they are likely to be problematic and present potential avenues for overcoming these issues. Copyright © 2016 John Wiley & Sons, Ltd.