Hypothetical bias continues to be a major challenge for stated preference methods. Cheap Talk (CT) has been found to be an effective remedy in some applications, though empirical results are ambiguous. We discuss reasons why CT may fail to effectively remove specific types of hypothetical bias in Choice Experiments. We suggest augmenting CT in Choice Experiments with a so-called Opt-Out Reminder (OOR). Prior to each choice set, the OOR explicitly instructs respondents to choose the opt-out alternative, if they find the experimentally designed alternatives too expensive. In an empirical survey we find the OOR to significantly reduce total WTP and to some extent also marginal WTP beyond the capability of the CT applied without the OOR. This suggests that the CT practice should be adapted to fit the potentially different decision processes and repeated choices structure of the Choice Experiment format, rather than merely being adopted directly from Contingent Valuation.