A wide range of crimes are not included in the BCS
We're all guilty - media, politicians, campaigners.
Guilty of simplifying statistics and presenting them in black-and-white terms.
So, when the crime figures come out, they're either "up" or "down" or "stable".
Never mind that the data may be incomplete, or give only a partial view of the reality.
It's this "reality check" that the Crime and Society Foundation is offering in its new report.
The foundation has examined the basis for the key claims made by ministers about crime - and persistent offenders - and found that they do not stand up to scrutiny.
The report is not alleging that some crimes have not fallen.
It acknowledges, for instance, that there is sound evidence for saying that burglaries and car crimes are down - because these offences appear to be measured effectively and consistently by the British Crime Survey (BCS).
But there are a wide range of other offences - drug crimes, sexual assaults, fraud, murders, environmental crime - not included in the BCS, which is cited by ministers as the most reliable indicator of crime trends.
According to a Home Office study, quoted in the Crime and Society Foundation report, there were more than 60 million offences committed in the year 1999/2000.
That's five or six times the number reported in the BCS.
The implications are clear: if the government's policy is to cut crime, and if the BCS figures are relied upon to judge the success of its policy, then the focus need only be on those crimes included in the BCS, and that claims that overall crime is falling may not be true.
The report makes a similar point with regard to the assertion that "100,000 persistent offenders commit half of all crime" or that "5,000 prolific offenders are responsible for one in ten crimes".
These phrases have become common currency, yet they appear to be based on data relating to convictions.
Quite apart from being a serious statistical misinterpretation, the report argues that a policy to target this hardcore of criminals will end up focusing on the "usual suspects" - those known to the authorities who are convicted most often, and not necessarily those who offend most often.
To some, it may seem like hair-splitting, but ministers have devised an entire initiative - the Prolific and Priority Offenders Strategy - on the principle that if this group is dealt with then crime can be cut and harm to communities reduced.
The new report is unlikely to deflect the Home Office from this approach, but it does present a plausible reason as to why it may not work.