We examine the influence of perceived social norms on the adoption by French farmers of eco-schemes (ES), a new type of direct monetary compensation provided by the EU Common Agricultural Policy to farmers who voluntarily adopt sustainable agricultural practices. Using a representative large scale web-survey, we show that French farmers hold inaccurate beliefs about their peers on various dimensions of the ES. Farmers are then randomly exposed to norm-based informational treatments where we vary the type of truthful information released about their peers. In addition to being able to assess the direct causal impact of the treatments, our between-subject design allows us to analyze cross-learning effects: providing information on the beliefs of peers may not only induce farmers to update their beliefs about their object of interest, but may also change their beliefs about other outcome variables. We demonstrate that norm-based informational treatments may modify:(i) farmers’ personal opinion regarding the ES; (ii) farmers’ beliefs about their peers regarding the ES; and (iii) that cross-learning effects may matter. While changes in beliefs are shown to be consistent with Bayesian-updating, we demonstrate that their causal effects may strongly depend upon the specific nature of the belief under consideration.