Decisions and Data

I've recently been reading the book Blink by Malcolm Gladwell.  It's quite an interesting book and one of the ideas put forward is that when making decisions we gain more confidence by considering more data but we make poorer decisions at the same time.  If you can find the key elements then you can focus only on them and make quicker and more accurate decisions.  Finding those key elements, though, can be a real exercise in creative thinking.

Let's start by taking a quick tour of the first approach.  Why does it not work?  The thing is,  more data would allow us to create a better picture and from that make better decisions.  If we could interpret it properly.  The problem is,  usually we cannot.  Straight out the gate we're doing it wrong.  Typically we start by trying to reduce evreything to numbers.  We take qualitative features and conver them into quantitative values.

The problem with this is that it is a vague approximation at best and it also strips out the most imortant part,  the relationships.  If you really want to understand something, don't seek to measure it's characteristics, aim to see how interactions play out.  The outcome you are trying to predict is often the result of many, many interweaving factors, many of which are not obvious and require a creative mind to uncover. 

The Butterfly Effect is an idea in Chaos theory which poses the idea that large events can often be traced back to an initial minor and seemingly inconsequential event.  For instance a hurricane could be the eventual result of a butterfly flapping it's wings in a completely different part of the world a few months earlier.  Running backwards through the relationships involves constanty modifying your perspective, a key creative thinking characteristic.

Let's take an example.  Say we are wanting to find predictors for who will improve the most at a particular sport,  unicycling.  We have 10 people and none of them have ridden a unicycle before.  We are working for a company who wishes to offer a sponsorship to the person who they believe will be the best at the end of 10 weeks.  The sponsorship must be offered today so you only have 1 hour to make your decision.  What would you look for?

The typical approach may be to lots of features you think would lend themselves well to the task and seek to identify those.  You may identify balance, determination, how many other sports they've played, etc as things you want to measure (by asking them).  You could much data this way but can you model it all in a useful way.  For instance, does lots of natural balance lend well but mean nothing if they don't have determination?  Did they play a lot of sports simply because their parents wanted them to and not because they were interested?  Are there other factors that affect these factors in important ways that you are not measuring.  Eg. Is the level of balance not important but how they aquired it?  Natural vs practice.  Are people lying in their responses as they all really want the sponsorship? etc.

Let's take another approach.  They are all effectively starting at 0 in terms of unicycle ability so what I really want to measure is their ability to improve and their desire to improve.  I may decide that an indicator of this is how much they are willing to take control of their learning.  But learning is individual and so the person who is going to voluntarily modify what they are doing to suit them will go the furthest.  I would make my decision based on the following:

I would leave them in a place with several unicycles (enough for each person) in an area where they can play with them (flat ground, a wall and or poles to lean against etc.)  I would tell them I just need to go do something and will be back shortly,  we will conduct the assessment then.  Where I am really going is somewhere I can observe them without them noticing me.  Anyone who just stands and waits for my return is out.  The one I will pick is the person who takes a unicycle and tries the most varied approaches to figuring out the unicycle.  That is the person who will go the furthest in 10 weeks.

Now I may not be correct but I reckon I stand a much better chance than anyone who takes the first approach and I will do it with a lot less effort.  But I had to think creatively to figure out what I was really searching for and how I was going to effectively observe it (important distinction here, observing not measuring).

So next time you have to make an important decision, think about what factors you are basing that decision on and if there is a better way.

You're on your way to becoming a Highly Creative Person.