--- - |2 Abstract Evidence that context changes measured risk preferences raises concerns about the practice of measuring financial risk attitude in one context to guide investment choice. We found that participants who first made choices between pairs of high (low) risk pension funds subsequently preferred higher (lower) risk funds when offered a choice from a wider range of funds. This effect was also observed when the riskiness of the initial pension funds was manipulated within subjects. Effects were not influenced by order, or attenuated by a bias warning. Tests across the domains of recreational and financial risk found that context effects are domain specific and that they influence both choices and judgments. Our results are consistent with theories of relative judgement. However, we also observed some evidence of sensitivity to absolute values. From an applied perspective, understanding such contextual sensitivity is important for recognising the limitations of risk profiling tools as part of a regulated financial advice process. - Journal of Behavioral Decision Making, EarlyView.