Taxing Statistics: An error of less than 1%

A ‘human error’ led to the government accidentally sending tax forms to ‘fewer than 100’ children – including a nine-year-old.

Now the last census in 2011 said that there were 16,213 children under the age of 16.. There are probably more now.

But that means that in putting all the island children on the tax system, there was an error of 0.62%. That's less than 1%.

With any new system, there are bound to be the odd glitches, either in the program or in the input, but an error of less than 1% seems pretty good to me, especially given the time pressures.

It's perhaps not the headline screaming from the media - "nearly one hundred children" - and one might add, no point of comparison with the number entered on the system.

Now the obvious thing is to make the program idiot proof so that date of birth is checked against age before the checkbox resulting in the error can be ticked (or not ticked). But that's easy to see in hindsight. Talk to anyone who programs and they will tell you that no program can easily be made - what we in the trade call - "idiot proof", because what you think is obvious turns out to be not the case.

Once you know that this kind of mistake can occur, you can program extra checks to make sure it cannot happen or flag up warnings. Making sure it can't happen is best, but on occasion when a complex judgement is needed, a warning is the only way. That works fine as long as someone reads the warning - which is alas not always the case. There's a click culture which doesn't read stuff properly - just think how many terms and conditions you have spent carefully reading before clicking "yes, I have read them".

So the next time you read a headline which gives a statistic, think about how that figure fits in the context of correct entries, and ask if the media have decided to use this wonderful opportunity to make a whipping boy of the tax department.

I don't think the tax department are doing everything right, and I think they underestimated the amount of work to transition to the new system, but if Kristina Moore is going to ask questions, whether they are adequately resourced in terms of staff should be at the top of her agenda, rather than asking where the buck stops.