We test convergent validity of contingent valuation and choice modeling with field and eye-tracker data in an application for stone pine reforestations. Field results yield significantly different structural models and higher willingness to pay values in choice modeling. Eye-tracker results show that respondents devote relatively more time to attributes and the bid in choice modeling and that there are no significant differences in the total time used to answer. Divergences remain for modified formats that minimize visual and cognitive differences between methods. Using an attribute-stimulus format in contingent valuation or including dominant alternatives in choice modeling does not change results.