Education authorities in Georgia took what apparently is a very unusual step: They made some routine checks to see if there was cheating on standardized reading tests. They found that cheating might have occurred at 1 in 5 schools, The New York Times reports.The same paper also reported recently that a survey taken by two criminologists found that more than 100 retired NYPD officials, ranked captain and higher, said that the intense pressure to lower crime rates had led police supervisors to manipulate statistics.Both articles reveal that there is a great reluctance on the part of public officials to consider the possibility that fraud has tainted data they rely on so much to show the public what a great job they are doing. Georgia education officials, it turns out, were exceptionally conscientious. But few other states have made use of a readily available computer tool that would identify cases in which the students' tests have an unusual number of erasures. In New York, the Police Department scoffed at the criminologists' survey and insisted internal audit controls were working fine. But if you talk to police officers, you will find that many are quietly offended by the manipulation of crime reports they see. The NYPD's reliance on data through its heralded Compstat system has been copied in many other cities.Crime in New York has dropped, for sure, and perhaps the national push for schools to focus on reading scores has produced some modest educational improvements (a debate in itself). But that doesn't mean data should be set up as a sort of Wizard of Oz-like oracle. Beware the man behind the curtain. Given the increasing reliance on statistical performance indicators in government, falsifying data is a serious threat to government integrity and needs to be monitored and investigated independently.