Findings on children's proportional reasoning abilities strongly vary across studies. This might be due to the different contexts that can be used in proportional problems: fair-sharing, mixtures, and probability. A review of the scientific literature suggests that the context of proportional problems may not only impact the difficulty of the problem, but that it also plays an important role in how children approach the problems. In other words, different contexts might elicit different (erroneous) thinking strategies. The aim of the present study was to investigate the role of context in third graders' (n = 305) proportional reasoning abilities. Results showed that children performed significantly better in a fair-sharing context compared to a mixture and a probability context. No evidence was found for a difference in performance on the mixture and the probability context. However, the kind of erroneous answers that were given in the mixture and probability context differed slightly, with more additive answers in the mixture context and more one-dimensional answers in the probability context. These findings suggest that the type of answers elicited by proportional problems might depend on the specific context in which the problem is presented.