Recent research has called for eliciting plural values of nature, yet little is known on how the choice of methods impacts the different values elicited. Drawing on the notion of methods as value-articulating institutions and using grasslands restoration as a case study, we explored how different elicitation methods influence people's value expressions towards grasslands. We did so in three different ways: (i) comparing values between elicitation methods (i.e., open-ended questions, Likert-Scale survey, rating exercise), (ii) comparing common discourses that emerged using multivariate statistics (i.e. multiple correspondence analysis (MCA) and principal component analysis (PCA), and (iii) tracing how interviewees' expressed discourses varied between methods. Our results showed that different elicitation methods not only elicited distinct values and discourses but also influenced the discourse that respondents endorsed during the same interview. These findings demonstrate that elicitation methods act as value-articulating institutions by defining which values could be expressed and how. While the Likert-Scale and rating exercise strongly framed and limited which values could be expressed by respondents, the open-ended questions loosely outlined and guided value expression. This study posits that values can only be understood in light of the methods used to elicit them and further, that using only one method for the elicitation of plural values might lead to neglecting or overlooking of particular values because of the methods conduciveness to eliciting or articulating them. Thus, plural valuation necessarily requires the application of multiple, complementary methods to unleash its full potential to elicit plural values.