Limitations of GIS Health Research

Modifiable Areal Unit Problem

Formal name to describe the phenomenon of how statistical effect/outcome and interpretation changes depending on how neighbourhoods are defined and partitioned, i.e. how a map’s boundaries are chosen affects the subsequent analysis.

[There is a] demonstrable affect [sic] of boundaries on empirical models of health inequalities. Statistical conclusions are susceptible to the magnitude of data aggregation and the way in which the units are subdivided whenever researchers work with data that was partitioned by administrative fiat. (p. 145[1]).

Specifically there are two (related) issues:

  • Scale Effect
  • Zooming Effect

Scale Effect

“[D]ifferent statistical results are obtained from the same set of geographic units when they are organised into an increasingly higher (or lower) spatial extent.”(p. 145[1]).

Further reading: Openshaw (1984) ‘The Modifiable Areal Unit Problem’ in Concepts and Techniques in Modern Geography.

Zooming Effect

“[E]ffect of basing a hypothesis from areal geographical units, which, if subdivided differently at the same spatial extent, may or may not lead the investigator to conclude differently.” [original emphasis] (p. 145[1]).

Further reading: Haynes et al (2007) ‘Modifiable Neighbourhood Units, Zone Design and Residents Perceptions’ in Health and Place 13(4).

Inadequate Level of Measurement

GIS research in population health is commonly limited by being dichotomous, so can’t offer any insight in to the graded nature of the social health gradient (p. 142-3[1]).

References

1. Schuurman, N. and Bell, N. (2011) ‘GIS and Population Health: An Overview’ in Nyerges et al (eds), The Sage Handbook of GIS and Society, London: Sage

Advertisements