How do we map joint actions we participate in onto joint actions we observe others performing, such as when a couple dancing tango observes another couple dancing tango? We investigated this question using a task where participants were instructed to perform individual or joint movements in synchrony with individual or joint movements observed on a computer screen. The observed movements started slowly and then continuously increased in tempo (from 1.75 Hz to 3 Hz). The results showed that, with regard to spatial parameters, joint performance was more accurate when observing joint performance than when observing individual performance. Individual performance was more accurate when observing individual action than when observing joint action. There were no systematic differences with regard to timing parameters. These results suggest that mechanisms of temporal coordination may be less susceptible to differences between individual and joint action than mechanisms of spatial matching.