There are some issues with the data in this report.  Inaccurate node counts.  Too low in many or all cases.  I think improper methodology was used to gather the statistics.