The identification and treatment of protest responses in stated preference surveys has long been subject to debate. We analyse protest responses while investigating ecosystem services providers’ preferences for incentive‐based schemes. We use a choice experiment for olive farmers’ preferences for agri‐environmental scheme participation in southern Spain. Our two main objectives are: first, to identify and discuss a range of possible motives for protest responses that emerge in a WTA context; second, we analyse the impact on WTA estimates of censoring serial non‐participation linked to protest or high compensation requirements (very high takers). Using a random parameter logit model in WTA space, we find that the inclusion or exclusion of serial non‐participants in the analysis can have a significant impact on marginal and total WTA estimates. Based on the findings, the paper makes recommendations on how to reduce the incidence of protest responses through survey design, regarding the identification of protesters as opposed to very high takers, and regarding the treatment of both groups of respondents for WTA estimation.