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
that tests were devised, and it was the devel-
opments by Cliff and Ord that brought the
topic to prominence. A mathematical repre-
sentation of ‘neighbours’ was a stumbling-
block to computing the tests, and Cliff and
Ord reformulated Moran’s test to employ a
W-matrix of sizeNN, whereNwas the
number of regions or observations, and a cell
valuewijof 1 indicated that regionsiandjwere
neighbours, and 0 if they were not (Cliff and
Ord, 1973). This idea opened up a whole field
of research, and spatial autocorrelation is now
a theme in many social sciences. Moreover, as
Tobler’s Law suggests, spatial autocorrelation
should not just be seen as a problem but also
as a reflection ofspatial interactionand a
central aspect of spatial modelling, and this
has been the basis for the development of
spatial econometrics. The Moran test is a
global one, detecting whether there is autocor-
relation on average across the set of regions; a
more recent development has been the disag-
gregation of this measure into local measures,
such as Anselin’s LISA (Local Indicators of
Spatial Association) to detect local clusters of
positive and negative autocorrelation (Anselin,
1995). (See alsolocal statistics.) lwhSuggested reading
Odland (1988).spatial econometrics An interdisciplinary
research field being developed by economists,
geographers and statisticians. The term was
invented by Jean Paelinck (Paelinck and
Klaassen, 1979), and the research focuses on
the construction and application of statistical
modelsand tests explicitly designed for spatial
(geographical) data, building on initial work
onspatial autocorrelationby Cliff, Ord
and others.
The field of econometrics was developed
throughout the twentieth century by econo-
mists to deal with the special statistical mod-
elling issue posed by the non-experimental
context of economics and the social sciences,
often very different from thelaboratoryand
agricultural field-trials context for which much
statistical theory was developed. In particular,
econometricians constructed methods to deal
with time-series autocorrelation, time-lags
and dynamics in economic relationships and
simultaneity andfeedbackbetween different
equations in macroeconomic models.
However, it is only in recent decades that
equivalent techniques for spatial data have
been developed. Amongst many contributors,Luc Anselin’s work, and his computer package
SpaceStat, have been particularly influential in
developing both the methods and their appli-
cation. The focus of the field has moved from
testing for spatial autocorrelation towards the
modelling of spatial interaction through
spatialdiffusion, spatial lags or spillovers,
andendogeneityis a major focus in the con-
struction of the estimators and tests. lwhSuggested reading
Anselin (1988, 2002, 2003).spatial fetishism Any approach that treats
spaceas sufficiently autonomous to social pro-
cesses that ‘no change in the social process or
spatial relations could alter the fundamental
structure of space’ (Smith, 1981b, p. 112).
The term was developed in the course of inter-
actions between human geography and
marxism; by analogy with Marx’s critique of
commodityfetishism, in which the conditions
of production of a commodity are made invis-
ible and theeconomyis reduced to exchanges
between objects rather than seen as a series of
unfolding social relations and practices, so it
was argued that approaches such asspatial
science conceptualized space as somehow
exogenous to and separate from social pro-
cesses (cf. Soja, 1980) (See also spatial
separatism.) esspatial identity The ways in which
identities are constituted, articulated and
contested in relation to space andplace.
A wide range of research has explored the
territorialized spatialities of identity in relation
to, for example, a region,‘homeland’or
nation. Other research traces the deterritorial-
ized – and reterritorialized –spatialities of
identity in terms ofmigration,diasporaand
borderlands. Whilst some research has
viewed space merely as a container foriden-
tity, or as a stage on which identities are
played out, other research – particularly since
the 1990s – traces the mutual constitution of
both space and identity and their multiple and
contested articulations (including Keith and
Pile, 1993). Rather than view spatial identities
inessentialistterms, this research – often
inspired by post-structuralism, psycho-
analysisandpost-colonialism– traces their
politicized differentiation. Although spatial
metaphorshave been particularly important
in conceptualizing identity, geographers have
been concerned to ground and locate the
material spatialities of identity, which are often
closely bound up with ideas aboutsituatedGregory / The Dictionary of Human Geography 9781405132879_4_S Final Proof page 712 1.4.2009 3:23pmSPATIAL ECONOMETRICS