The language switching performance of two sets of bilinguals was tested using a task that involved a sentential context. Speakers of the first group (Spanish-English bilinguals) were tested in both a voluntary and a cued parameter, whereas the second group (Arabic-English bilinguals) were tested only in the voluntary parameter. The first group delivered a reduction of switch costs in the voluntary parameter, and the second, an elimination of costs suggesting that more ecologicaly valid experimental setups such as the one used, contribute to a reduction of switch costs.