Social support is demonstrated to have mixed effects on both pain and related physiological arousal. In this study, a meta-analysis was conducted to characterise these effects. A total of 2416 studies were identified in a systematic search, among which 21 were eligible for the quantitative review. The mere presence of another person was not sufficient to modulate pain perception. However, the presence of a stranger was identified to decrease pain-related arousal (SMD = −0.31), and the presence of a significant other increased facial expression of pain (SMD = 0.21). We further found verbal support to decrease pain (SMD = −0.69) and arousal (SMD = −0.99), and we demonstrated moderate to large analgesic effects of intimate relationships through touching (SMD = −0.95) and viewing (SMD = −0.60) of a romantic partner. Finally, we presented evidence of publication bias for pain-related arousal but not for behavioural pain outcomes. Together, our findings suggest that the impact of social support on pain is context-dependent with the verbal communication of support and intimate relationships being of particular importance.