Crowdsourcing is a widespread method for outsourcing tasks to a crowd of contributors. These simple tasks are often formulated by multiple choice questionaries (MCQs) to which the contributor has to give a precise answer. We hypothesize that offering the contributor to fill in an imperfect answer and to evolve it is more profitable than a single precise answer. In this paper, we propose a model for the evolutionary answers of contributors to MCQs in crowdsourcing platforms. In order to realize this modeling we use the theory of belief functions. The model and experiments conducted on real data from crowdsourcing campaigns are presented in this paper. Our experiments show that modelling evolutionary responses using consonant mass functions improves the quality of the results obtained when aggregating responses compared with majority voting.