Saturday, December 05, 2009

A problem common to parents of young children everywhere

Any data set needs to be cleaned up, manipulated if you like, before proper analysis can begin. This applies even to a comparatively well controlled collection exercise, limited in its coverage in both space & time, such as a census of population.

Can a 10 year old girl be married, or a 17 year old boy a doctor, or should the records be amended?

How to deal with missing data?

What groupings may we safely or properly use to collapse the detail into useful categories without sacrificing or concealing important information?

Decisions have to be taken on matters such as whether forced consistency - a One Number Census - overrides all other considerations.

All of these discussions can - & do – lead to heated debate, some of which may be intemperately expressed. Life long enmities may be formed.

The answers adopted will always, to a greater or lesser degree, depend on how you think the world really is – there is just no getting away from that. Even when you are dealing with proper science (ancient tree rings & where modern thermometers are placed) & not just the softy social kind.

The least you can do is to be as honest & open as you can be about the process & to keep it well documented (though that can also result in just another indigestible pile of records).

If you can collapse all the important information about the earth’s atmosphere & climate over the last several millennia & extract from that one indisputable equation which is stable enough to project over the next N years to tell us what the future WILL be – well congratulations.

But please, there must be questions to be asked about the details.

Boys, boys: stop fighting. You should learn to share.



Related posts