How trustworthy are official statistics?

Recently here, the veracity of some official statistics has been called into question. This is surely not confined to the UK.

Some of the statistics that have been questioned include those generated by the police. For reasons that need no explaining here, the veracity of rape statistics has been the most controversial, with particular reference to the number of reported rapes recorded as “no crime”.

This is far from a new problem, and is by no means confined to crimes against the person.

Here is a report from December 2009 which alludes to such practices as cuffing, nodding, skewing and stitching in order to massage crime figures. Suffice it to say, these practices are not 21st Century inventions.

Obviously, some statistics are more reliable than others. One would expect the number of murder convictions to be 100 percent reliable, because this involves nothing more sophisticated than counting.

The actual number of recorded murders in any year should also be fairly reliable both because murder is an extremely rare crime and because it is always investigated seriously. On the other hand, one would not expect these figures to be as reliable as the recorded number of convictions because a few may slip through the net for various reasons.

When we begin talking about property crime or non-crime statistics related to such things as economic output, then clearly we are on shakier ground, because even without blatant fraud or errors caused by recklessness, such figures can only be estimated often to a very imprecise degree.

A recent article about the Chinese economy pointed out that little can be read into any purely statistical analysis. This is true to some degree for not only a country as vast as China but even for fairly small geographical areas and organisations.

When official statistics can be so untrustworthy, including those their compilers have no perceived interest in massaging, you should always exercise caution when relying on statistics produced by pressure groups and the like, who have every reason to be less scrupulous than the men from the ministry.

[The above was published originally February 4, 2014; the report from December 2009 has been replaced with a later link for technical reasons].


Back To Digital Journal Index