Should we present measures of statistical uncertainty when we have enumerated the entire population?