Data Rich - Information Poor : Furnished Lets

Data Rich - Information Poor :
Categories out of context - furnished lets
© Mahes Visvalingam, 2000

Croplogo.jpg (11317 bytes)

Most statistical definitions of social indicators, as well as area classifications, treat variables as if they were dichotomous variables. For example, the renting of accommodation from private landlords and from the local authority will be treated as two separate deprivation variables. I was interested in studying this dubious practice of plucking out of context just the apparently relevant categories from multi-category data. I will focus on two such categories which have been used as dimensions of deprivation – namely furnished lets and elderly respectively.

Here is a 3D scatterplot showing the chi-score distribution for the 4 main house tenure types, namely owner occupied in blue, council-owned in red and private renting in green.

This map uses the same colours and symbols to show places where the various types of tenure were concentrated.

Let us just focus on furnished lets, which are shown as open green circles. No doubt many deprived people lived in furnished rented accommodation, from which they could be more easily evicted if they were in rent arrears, for example.

However, the map shows that areas with an excess of furnished lets were far from deprived. The map was picking out locations where air and armed forces were stationed (e.g. the two cricles above Hull) and also where Victorian dwellings within Hull had been divided into furnished lets for professionals and students. So, dichotomous variables can be misleading.