The Dictionary of Human Geography

Comp. by: VPugazhenthi Stage : Revises1 ChapterID: 9781405132879_4_S Date:1/4/
09 Time:15:23:39 Filepath:H:/00_Blackwell/00_3B2/Gregory-9781405132879/appln/3B2/
revises/9781405132879_4_S.3d

qualitatively complicates the contradictions,

crisesandclassstruggles that are character-

istic of capitalism. The spaces and places

produced throughaccumulationand compe-

tition dynamics become barriers for future

accumulation; conflict between places can

cut across and undermine class struggle; and

individual agents find it all but impossible to

undertake actions that are in their long-term

as well as their immediate interests. Again,

taking account of the spatial extension of

economic processes requires adjustments to

conventional, a-spatial political economy. In

this case, however, in contradistinction to

regional science and location theory, analysis

focuses on the dynamical dialectical relation

between economic processes and thespatial-

itythey shape and are in turn shaped by (see

dialectic), rather than on assumed spatial

structures and their impact on spatial eco-

nomic equilibria.

As economic geography has subsequently

moved to a conception of economy that

emphasizes the inseparability of the economic

from other societal and biophysical processes,

thereby calling into question any theory that

seeks to separate or prioritize economic rela-

tive to these other processes, so once again

the term ‘space-economy’ has fallen out of

favour (see also cultural turn; institu-

tional economics). es

space^time forecasting models Statistical

models that attempt toforecastthe evolution

of variables over bothtimeandspace(e.g. sets

of regions). These models are usually of the

general regression form and forecast the

future value of a variable and an observation

unit in terms of (a) lagged exogenous or

explanatory variables, (b) its own past values

and (c) the lagged values for neighbouring or

influencing spatial observation-units, thus

capturing the impacts of spatialdiffusion.

These models have been used to forecast

both economic and demographic changes,

and in studies ofepidemicsand the modelling

ofdisease. lwh

Suggested reading

Bennett (1979).

spatial analysis The application ofquanti-

tative methods in locational analysis

withinhuman geographyand sometimes used

as a synonym for that portion of the discipline

that concentrates on thegeometryof theland-

scape(cf.spatial science). O’Sullivan and

Unwin (2002) present spatial analysis as the

study of the arrangements of points, lines,

areas and surfaces on amap, and of their inter-

relationships. Analyses of those separate com-

ponents have deployed procedures adapted

from other sciences – nearest-neighbour

analysis and quadrat analysis, for point

pattern analysis;graph theoryfor lines;

andtrend surface analysisfor surfaces, for

example. Whereas many geographers have

undertaken analyses of the interrelationships

using techniques from within thegeneral

linear model, others have argued that spatial

analysis poses particular statistical problems

because of the nature of spatial data (cf.

spatial autocorrelation), thus requiring

special techniques.

The development ofgeographic informa-

tion systemsis rapidly facilitating advances in

spatial analysis and the greater power of com-

puters, together with software developments,

has significantly increased geographers’ ability

to work with large and complex spatial data

sets (cf.geocomputation). rj

Suggested reading

Bailey and Gatrell (1995); Haining (1990).

spatial autocorrelation The presence of

spatial pattern in a mapped variable due to

geographical proximity. The most common

form of spatial autocorrelation is where similar

values for a variable (such as county income

levels) tend to cluster together in adjacent

observation-units orregions, so that on aver-

age across the map the values for neighbours

are more similar than would occur if the allo-

cation of values to observation-units were the

result of a purely random mechanism. This

is positive spatial autocorrelation. Negative

autocorrelation is where neighbouring regions

are significantly dissimilar; more general and

complicated forms of autocorrelation can also

be defined. The presence of spatial autocorre-

lation is very widespread and indeed may be

said to lie at the core of geography,as

expressed in Tobler’s (1970) – light-hearted –

First Law of Geography: ‘everything is related

to everything else, but near things are more

related than distant things’.

However, the presence of spatial auto-

correlation violates a basic assumption of

independence in many standard statistical

models. Thus for regression, there is an

assumption that the residuals are not autocor-

related. The issue of spatial autocorrelation

was recognized early in the history of inferen-

tial statistics, but it was not until the work of

Moran and Geary in the late 1940s and 1950s

Gregory / The Dictionary of Human Geography 9781405132879_4_S Final Proof page 711 1.4.2009 3:23pm

SPATIAL AUTOCORRELATION