Analysis of the output of a process has suggested that the variability is nonrandom on several occasions recently. However, each time an investigation has not revealed any assignable causes. What are some of the possible explanations for not finding any causes? What should the manager do?