Objective: Our study sought to explore, by using response time measures, the cognitive effort associated with comparative optimism expression (CO) and its modulation. More precisely, our aim was to decide between two opposite options: (1) expressing CO as a self-serving bias (presenting oneself as better than others) will require less cognitive effort than restraining CO; and (2) modulating CO depending on social context will be effortless, thus for example restraining CO as a normative self-presentational answer (for presenting oneself modestly) will require less cognitive effort than expressing CO. Methods: Participants answered a CO questionnaire in two social contexts in which CO is socially valued (professional domain) or not valued (friendship domain). They had to answer either spontaneously or in order to convey a favourable or an unfavourable impression. Answers and times needed to respond were recorded. Results: The present data revealed that participants were slower when restraining CO to convey a negative image to an employer or to depict a modest self to a friend. Conversely, they were faster when expressing CO to convey a favourable image to a recruiter or a negative one to a friend. In the spontaneous condition, a same level of CO was expressed, but response times were lengthened in the friendship domain. Conclusion: Overall, these results suggest that in comparing one's own risk with that of an average other, restraining CO according to the socially valued self-presentational standard of modesty corresponds to a strenuous answer. Consequently, expressing CO might represent a more overlearned, automatised self-presentational answer.