Data Rich - Information Poor : Problem
© Mahes Visvalingam, 2000

Croplogo.jpg (11317 bytes)

    The powerpoint presentation and published papers are available.from m.visvalingam@dcs.hull.ac.uk.

problem.jpg (82731 bytes)  
Area-based programs for regenerating deprived areas need to identify the worst areas. Why is the identification of these worst areas a problem?
  • For a start, deprivation (like market potential) is a vague concept.
  • We cannot measure deprivation directly. Personal details, such as income, are protected in UK.
  • To ensure confidentiality, personal data are normally aggregated to and released for a variety of functional areas, as explained in the excellent edited volume by Dale and Marsh (1993)
 
So, the occurrence of multiple deprivations is normally inferred from a set of proxy characteristics.  The character of deprivation varies geographically and over time. Conditions today are very different to those in 1971, let alone Victorian Britain.   Even so, unemployment has always been regarded as the key indicator.

What normally happens is that areas are scored on individual indicators, such as male unemployment.  These individual scores are then summed to derive composite indices. The indices are then used to rank areas on a good to bad scale so that the worst areas may be identified.

I will limit myself to problems relating to the scoring of areas on individual symptoms; i.e. to the statistical definition of the primitive indicators themselves.

 
data.jpg (31235 bytes)  
In my examples, I am going to use the 1971 population census data released for one kilometre grid sq areas.   I am going to use this data for two reasons:

Firstly, use of historical data will focus attention on methodological issues and avoid current political controversies.

Secondly, and more importantly, constant area units highlight a problem that had escaped attention for decades. 

 
I will initially use just 2 variables, namely:

    - Number of economically active males
    - Number of unemployed males

The problem is how do we use these two variables to rank geographic areas according to the severity of male unemployment? You are probably saying to yourself, surely it is a simple matter of using the rates of unemployment.

I will briefly explain why I have reservations about this.  If you want to know more, I will be happy to supply copies of papers, which are listed on the web

 

HOME / NEXT