Feeding My Ignorance

"Its all just ones and zeroes" - not sure whether that quote comes from a single identifiable source, but it is often used as a way of breaking down the complexities of most computer systems to something "understandable". The problem is that although we understand "ones and zeroes" in their own right it requires rather more to interpret the meaning of all those "ones and zeroes".

...

Data presentation is hugely important in all sorts of situations. The types of mathematical model that I do demands that the output is rendered in a form that is understandable by the "client" - if I were to present a string of ones and zeroes then there would almost certainly be zero understanding!

The problem is that it is far too easy to "misrepresent" the data. I don't mean here by in any way manipulating it, or telling lies about it - I mean it is possible to present the data in such a way that it favours a particular understanding (and perception) on the part of those receiving the information. This is not a surprise, it is something that the media do all the time - and it plays (very much) on our inbuilt biases towards certain things.

One way of skewing things is to quote "bald" numbers - there have been 2503 more ... this year compared to last year. Doesn't matter what the "..." stands for - the point here is that the number is a "big number" and we will react accordingly - if "..." is something good then we rejoice - if it is something bad then we begin to panic.

There is no context - a bit like when I was told after a blood test that the result was "89" - sounds intuitively like a good number - unfortunately, in this case the minimum acceptable was around 200 and 'normal' wa more like 500. Without context a bald number is simply 'noise'.

I was thinking about this in terms of some modelling I have been doing recently - if I do x instead of y then (all other things being equal) I will be £2000 worse off in three years time. Sounds bad - better do "y". So - I have made a decision - I have optimised my outcome - utilising my "best knowledge" at the moment. You will recall earlier writings where I discuss my understanding that there is no "wrong" decision - we always make the "right" decision based on our current understading.

Wait - what about context? what exactly does "worse off" mean? for that matter are "x" and "y" really mutually exclusive - instead of either/or is there any room for both/and?

In this case there is a huge amount of context - it clearly matters whether the value in three years time is going to be £2010/£10 or (a rather unlikely scenario where I am concerned) £10264720/£10262720. In the latter case it is likely that there is a more compelling reason to do either "x" or "y" - the "worse off" result is not significant.

Another context is what are the outcomes other than my "value" in three years time? It may be that being worse off also results in a greater long term income - so if I were to examine ten years time or twenty years time I would be much better off.

We must also consider the "value" of doing "x" - if that is considerable then we may not be unhappy to end up worse off. This is, of course similar, to the previous point - we could be in some way investing the £2000 in the expectation of a high ongoing income.

In looking for a "real-life" example of this I came across an article on the BBC website. The headline might well have been "Bank of England forecasts growth of 2%". As stated in the article Minouche Shafik suggests that the 'whole story' is better represented by :

"If economic circumstances identical to today were to prevail on 100 occasions, the best collective judgment of the Monetary Policy Committee is that the mature estimate of Gross Domestic Product growth would lie above 2% on 50 occasions and below 2% on 50 occasions."

Yep - it needs context - it need explanation - it needs caveats. Even without "fake news" it is easy to mislead your audience - intentionally or unintentionally. I am sure that I have talked about it already - but cannot find the link - the act of communication is immensely complicated - in order to pass understanding from one person to another it is necessary to jump through hoops and penetrate a whole host of barriers seeking to block the clarity of the initial message.

So, as they say, what you thought you said is not what the other person thinks she heard!

Remember that - and the need for context - the next time you read a headline like "Unemployment falls by 20000".