Recently, a relationship between bilingualism and enhanced social flexibility has been suggested. However, research on the subject is scarce and what little exists is limited by several conceptual and methodological concerns. In the current study, we attempted to (a) replicate the findings from a study by Ikizer and Ramírez-Esparza (Bilingualism: Language and Cognition 21:957–969, 2018) by using the scales that the authors developed, and (b) test the concept of social flexibility experimentally with a switch-task using socially relevant stimuli. In the first part, participants (n = 194) filled out the scales developed by Ikizer and Ramírez-Esparza. We could not find that bilingualism leads to enhanced social flexibility. We did, however, find that higher level of education led to higher scores on the social flexibility scale. In the second part, a subsample (n = 74) from Part 1 completed a task where they were asked to identify the congruency between a face and a voice based on either gender or emotion, and to switch between these two tasks. The experimental task did not show an advantage for the bilingual participants. On the contrary, higher proficiency in a second language led to lower accuracy in the congruent emotion condition, while level of education led to higher accuracy in that same condition. We suggest that factors other than bilingualism, such as level of education and biculturalism, most likely drove the effect found both in the current study and originally by Ikizer and Ramírez-Esparza.