A place in history: a guide to using GIS in historical research

CHAPTER 7: SPATIAL ANALYSES OF STATISTICAL DATA IN GIS

7.5 Conclusions

Although a significant amount of progress has been made in determining the research agenda for quantitative analysis of statistical data, there is still some way to go before this functionality becomes a standard part of the GIS toolkit, and even further before it becomes a common part of historical analyses. There is a large amount of potential for the statistical analysis of spatially referenced data, but the limitations of the data must be remembered. This means that the researcher must select techniques that do not rely on the assumption of independent randomness, and must bear in mind the limitations of their data, in particular: its quality; potential problems caused by error propagation; the possible impact of modifiable areal units; and the potential for incorrect inference due to ecological fallacy.

As the use of GIS among historians becomes increasingly sophisticated, it is to be expected that increasing use will be made of spatial statistical techniques. The fact that these are not well integrated into GIS software as yet provides opportunities for the researcher to devise the techniques that are most suitable for their own data, although these must be devised within the limitations of the source data. The impact of the spatial nature of the data must be borne in mind here and, in particular, unnecessary aggregation should be avoided not only through space, but also through attribute and time. Such techniques should not just produce a single summary statistic for the entire study area but should show how the impact of the phenomenon varies across the map. The results of the technique may be map-based rather than using traditional numeric summaries.