Researchers who announced the stunning arrest rates in 1967 were stumped by data deficiencies, such as their inability to tell whether the same person was being counted more than once — an often overlooked point the researchers made in their own report. Today, data problems in crime measurement persist. Reporting by local law-enforcement agencies is incomplete, and criminologists say local data aren’t calculated in a uniform way across the U.S.

To mitigate gaps and inconsistencies in the numbers, the Federal Bureau of Investigation extrapolates from the numbers it does have to get a nationwide total.

That 52% of American men will be arrested originated in a report from the President’s Commission on Law Enforcement and Administration of Justice. Ronald Christensen, then a graduate student in physics who held degrees in electrical engineering and law, was asked to crunch the numbers on arrests and convictions. He set out to calculate how many men and women at each age had been arrested, counting juvenile arrests. Mr. Christensen assumed that current arrest rates would hold.

For the entire article,  please go to http://online.wsj.com/article/SB125851115456653127.html?mod=WSJ_hpp_MIDDLENexttoWhatsNewsThird.

Write to Carl Bialik at numbersguy@wsj.com

Printed in The Wall Street Journal, page A15