Purpose: This paper aims to test if team reflexivity is a countermeasure to the detrimental effect of team virtuality on team performance improvement, in decision-making teams. Design/methodology/approach: Study 1 regarded 210 individuals (N = 44 teams) executing five decision-making tasks. Study 2 regarded 60 individuals (N = 20 teams) executing four decision-making tasks. Study 1 was longitudinal, with no experimental manipulation. Study 2 had an experimental longitudinal design comprising two between-team manipulations: medium of communication and team reflexivity; the outcome was team performance improvement. Findings: Study 1's results show that team reflexivity positively moderates the effect of virtuality on team performance improvement over time. Study 2's results shows that a reflexivity manipulation benefits face-to-face teams more so than virtual teams, probably because team reflexivity is more effective when media richness is high. Originality/value: The implications of reflexivity's lack of effect in low virtuality (Study 1) and high virtuality (Study 2) teams are discussed. This study contributes to the team learning and virtual teams' literatures by expanding current knowledge on how team reflexivity can facilitate team learning under face-to-face versus virtual communication conditions.