Monthly Archives: November 2010

Lies….Damned Lies….and….

I’ve resurrected this to have a little rant about something I was reading on the internet today, now a lot of you won’t be interested, some of you may be confused by it, and many of you will think.”how sad” but meh.

It’s one of those little things that really irritates me.

I refer to the misrepresentation of statistics and the twisted manipulation some people will employ to try to back up a losing argument.

Now I’m not going to start spouting reams of data and picking flies over the tiniest discrepancy, I’m just going to vent about the lunacy spewed forth by someone who should have known better. Some of you may know what I’m talking about……most won’t.

So….to begin…..

Percentages……they mean nothing whatsoever unless you know what they relate to.

lets say, for arguments sake, you want to compare 2 things. One of them you say has only had an increase of 5%, the other an increase of 45% and therefore the second one is much more successful….seems logical….but 5% and 45% of what?

Lets add some figures, the first had an increase of 5% on 15,000….thats an increase of 750, the second had an increase of 45% on 30….thats an increase of 13.5….now which is more successful?

So point one…..Percentages mean feck all on their own, you have to justify them with some form of figures for them to have any relevance.

Trends…..a trend is only a trend if you have enough of it to measure.

So lets use the figures above shall we, both of them are increasing, the first 5% increase on 15,000, the second a 45% increase on 30. This isn’t a trend! this is just a simple increase. If the previous measurement had shown similar increases then it would be a trend, but to say that this month there has been an increase therefore the trend is that its sucessfull is bunkum! if the previous measures all show drastic decreases then the “trend” is just the opposite, this one isolated increase is an anomaly!

So point two…..trends are a discernible pattern in data over a series of events. Not a one-off observation of a small part of the available data.

Performance… a whole is related to a range of things.

Saying one thing is performing better than another is not as easy as saying “A sells more items than B therefore A performs better” what if A sells a thousand penny chews but B sells 2 Ferrari’s? who’s performed the best? Ok bad example, lets say A and B do facepainting, A paints 6 faces an hour, B does 5 faces an hour, and they both charge £5 a face. Who performs best? On the surface it’s A, but what if you take into account that with materials, stall rent etc it  costs A £3.50 per face painted, but B only has to pay out £1.50 per face…who performs best now? profit per hour for A is £9, and for B is £17.50 so looks like B performs better. But what about customer satisfaction? 95% of A’s customers are happy with their faces, but only 65% of B’s are… can see where I’m going with this.

Point three….You can’t make wild accusations that one thing performs better than another if you haven’t got all the facts, if you do you just make yourself look like an imbecile.

I think that’s enough, I have vented. I needed to do that, with all of the statistics being banded round in the news recently this little bug bare of mine has been over worked, and the straw that broke the camels back was an article I read to slating a councils proposal, ridiculing it by “picking holes” in its suggestions and totally misrepresenting the original report. Unfortunately the analytical prowess of the author of the article was lacking, the above are examples of the “facts” he was claiming, and even without all the data a dullard could tell that his arguments were flawed, misrepresented and ridiculous.


So there…Rant over.

As was falsely accredited to Disralei “there are three kinds of lies: Lies, Damned Lies and Statistics” I would add to that though “There are four kinds of lies: Lies, Damned Lies, Statistics and the unfathomable ramblings of fuckwits who think they are more intelligent than they are” 😉