How do we ensure that research is reproducible?