Generally, the most significant
challenge is to reach equilibrium, with high reliability often leading
to low external validity, and highly reliable experiments often
being low in generalisability to situations outside of the laboratory
(Carter et al., 2013).