The length of time for which seed lots maintain their viability and vigour in storage has become a trait of interest for rice and other crop and model species. However, different research groups have taken different approaches to measuring this trait, including storing seeds under 'natural ageing', accelerated ageing, and controlled deterioration test conditions. There has also been a tendency to use only a single germination test result to assess the relative longevity of different genotypes. Here, we explain why we have taken a different approach in a genome-wide association study of seed longevity in Oryza sativa (Indica Group), and offer suggestions as to how, in the future, this trait should be assessed to make it possible to develop varieties with improved seed longevity, depending on the target storage environment.