Accuracy vs Precision

Today's XKCD;

I'd be inclined to take this a step further; when you say "people are stupid compared to your expectations", what you are really saying is that my expectations of how smart people are is constantly and consistently wrong, yet I am unwilling to change my expectation to accomodate this information.

One of the things that stuck with me from my degree

¹ is the idea of "Accuracy" vs "Precision."

Precision is essentially about consistency – a repeated measurement will give very similar values. Accuracy is about a measurement being close to a "true" value. So, a measurement using an incorrectly calibrated set of weighing scales can be very precise (always reporting the same weight), but not very accurate. Or reporting a figure to a large number of decimal places can be a very precise measurement – but that doesn't mean that it is an accurate one.

So, the "people are stupid" statement could be taken as saying that your views are not accurate, you know that they are not accurate, you know how you could change your views – but you refuse to. Perhaps to maintain your belief in your own level of relative 'smartness'.

Which, I guess, is pretty stupid…

---

It also reminds me of my favourite George Carlin quote – Think about how stupid the average person is. Then realise that half of the population is more stupid than that.

²

1. I studied Chemisty with Environmental Chemistry. On one hand, I probably should have switched to a topic I found more interesting when I realised that I wasn't going to maintain a keen interest for 3 years. On the other hand, I'm not sure I would have necessarily made a better choice. ↩

2. Of course, if you interpret "average" to mean "mean", then this assumes that stupidity/smartness is normally distributed, and the mean and median are the same – which is not necessarily true. ↩

Twitter and GfK announce TV measurement partnership →

Twitter and GfK have announced a partnership to "introduce GfK Twitter TV Ratings in Germany, Austria and the Netherlands. The new service will provide insights into the frequency and reach of messages from Twitter users associated with television programs and campaigns."

Those watching the "social TV" industry will recall the deal Nielsen announced with Twitter at the end of 2012 (I wrote about the deal and implications of this kind of measurement for my work website at the time.)

And those wondering why the UK wasn't included in the deal may want to cast their mind back to last August, when a similar partnership between Twitter and Kantar (with SecondSync providing data) was announced.

The big question from my point of view is about how this "reach" measurement is being measured.

Will it be based on inflated counts that assume that every follower sees every tweet, and doesn't account for the fact that people might follow more than one person who tweets about a programme?

Or will it be based on actual data from Twitter, who presumably have the ability to know how many people actually see each tweet (given that they have to do the work of putting it in front of them.)

Sadly, I'm expecting the former…