Statistically Speaking

So, the anecdotal eveidence that we are the only people in Britain who bought a new car has been confirmed by official reports of the latest car sales figures.

The news announced these as catastophic. I read more about it with interest in The Times today.

However, I discovered that (as all too often today) the stats are presented in a misleading and ultimately uninformative way.

I read that the sales of several marques are down 50 - 70% - they're listed out. But, those turn out to be the changes from July to August. With September (and new reg) just around the corner, obviously August will lag July.

Are these drop-offs much worse than usual ? Undoubtedly, but we have no idea by how much.

What we need are Year-on-Year figures, clearly, but these are not given anywhere.

We are told that the August figures compare in a shockingly bad way with August (so same month) figures from way back into the 1960s. That is certainly striking, but of course in the past, August was a bumper month as new registrations began on 1st August - it was the peak month. Now, presumably, it is routinely the lowest month as it is the month PRIOR to the new registration AND in the summer lull.

So, the comparison would have to be taken with great circumspection.

It's a shame that the real story - which most likely is bleak enough in plain fact - is so obscured by a clearly faulty presentation.

What I find even more frustrating is when journalists use completely false statistical or logical inference because they don't comprehend what is going on (whereas in this case, I am sure it was done for sensationalism which is at least understandable even if not necessarily desirable).

I think this may have overtaken my previous top irritation - reporting of news and especially of announcements before it has actually happened.