Thursday, April 21, 2011

Nuclear accidents, house prices & A levels

The nuclear accident in Japan has now been rated 7 on the International Nuclear and Radiological Event Scale which puts it in the same category as Chernobyl. Experts feel that this is in some sense misleading (see for example Understanding Uncertainty ) & therefore unhelpful in the sense that Sir Humphrey would use that word.

The problem is that the scale has nowhere else to go, no way of making any finer distinction between the serious, the very serious & the end of the world as we know it.

This is the kind of problem that quite commonly arises with scales & coding frames, just another example of man’s inability to foretell the future, to imagine all the possible outcomes that need to be catered for. And after all we, fortunately, do not have much experience of nuclear accidents to go on.

I remember a problem which arose with the old government house price index. This was one of the earliest computer projects, first designed in the 1960s. In those days space for both processing & storing data was ridiculously limited, compared to what we have today, & so statisticians had always to compress their data as much as possible before analysis began. (This of course is the same problem which led to programmers to omit, wherever possible, the 19 from 20th century dates, thus exposing the world to another threat to civilisation as we know it (allegedly) as the millennium approached).

The forms which came in to the department recorded the actual price at which houses were sold but these were collapsed before being put on to the computer into codes such as 1= less than £1,000; 2= £1,000 to £2,499 … 8= more than £20,000.

By 1975 it was clear that the time was rapidly approaching when no house would be sold for less than £20,000, & the whole system had to be redesigned.

And of course in the field of education we now have the situation where soon everybody will have a clutch of A*s at A level.