As Davidsson (2006, p.19) has pointed out, “One problem with the PSED design is that ... some cases may have been active start-up efforts for years before being sampled, while cases that are abandoned rather quickly after initiation are, in comparison, under sampled”. GEM surveys also suffer from this bias. They may contain more “tortoises” – long term or “slow” nascent entrepreneurs – and fewer “hares” – those “fast nascents” who sprint to a start-up or abandonment of the idea. In this paper, we assess the extent of this bias, and test whether time to start a business might predict future potential.