Abstract Aim: To examine the associations of midlife social participation, a potentially modifiable factor, with physical performance in early old age, a marker of healthy ageing.Background: Previous studies among older adults have shown that low social participation is related to adverse health outcomes such as disability. However, whether social participation is also related to objective measures of physical performance which capture variation in ability across a wider spectrum is not yet widely accepted.Method: Approximately 1,600 men and women from the MRC National Survey of Health and Development reported their involvement in a number of social, secular and religious group activities at 43 and 60-64 years and frequency of participation was classified into three groups: low, moderate and high. Grip strength, standing balance, chair rise speed and timed get-up-and-go (TUG) were assessed at 60-64 years. Associations between participation and performance were estimated using linear regression models.Results: Compared to those reporting low levels of social participation at age 43, individuals reporting high levels had faster chair rise (1.36 stands/min, 95% CI 0.39-2.34, p <0.01) and TUG speeds (2.45 cm/s, 95% CI 0.25-4.65, p= 0.03) at 60-64 years. Benefits were only observed if social participation persisted over time; any protective benefit was lost for those with reduced levels of participation at 60-64 years.Conclusion: Midlife social participation is associated with better physical performance in early old age, but only for those who maintained high levels of social participation, suggesting continuation of social participation may be protective or associations are bi-directional.