Comp. by: VPugazhenthi Stage : Revises1 ChapterID: 9781405132879_4_Q Date:31/
3/09 Time:19:14:09 Filepath:H:/00_Blackwell/00_3B2/Gregory-9781405132879/appln/
3B2/revises/9781405132879_4_Q.3d
voting, because of the many other contingent
factors (some of them spatial in nature) that
influence electoral behaviour.
This example illustrates the dominant focus
within contemporary quantitative human
geography, not – as it was 50 years ago – on
the orderly arrangement of points and lines in
the landscape, and of people, ideas and com-
modities flowing through those spatial struc-
tures (as postulated by Haggett, 1965), but
rather on common patterns of behaviour in
similar contexts – or common realizations of
human agency within particular contextual
settings. The emphasis has shifted between
two of geography’s fundamental concepts,
from a focus onspaceto one onplace. Within
that changed emphasis has come the adoption
of new quantitative procedures that emphasize
the important role of place as context – such as
geographical analysis machine,geograph-
ical explanation machine,local statistics,
geographically weighted regression,
multi-level modelling, the modifiable
areal unit problem,ecological inference
and theecological fallacy.
Contemporary use of quantitative methods
in human geography entails exploring
patterns and processes involving large numbers
of decision-makers, seeking order in often
complex situations (Haggett, 1990). Even
when data are available on individual decision-
makers, the goal is to distil general patterns – as
in Rushton’s distinction between behaviour in
space and rules of spatial behaviour. Unlike
other approaches to human geography, there-
fore, spatial science is macro/meso inscale,
rather than micro – though it may deploy
micro-scale data in its search for macro/meso-
scalepatterns. Its findings might pose ques-
tions that can only be addressed by studying
individuals, but its core methodologies focus
on wholes rather than parts.
This approach is the only one possible given
certain subject matter. Patterns of (changing)
regional development and underdevelopment,
for example, can only be studied as statistical
aggregates – as can many indices linked to this
concept, such as inflation, unemployment
rates, land values, house prices,productivity
and so on. Aspects of our worlds are only ac-
cessible in such formats, and are sensibly ana-
lysed quantitatively. In other cases, whereas
individual-level data are available (or can be
obtained through bespokesurveys) – on ill-
ness, turnout at elections and so on – much
can be gained from aggregating them and
studying general patterns, of morbidity and
mortality rates, for example. Much work in,for example, population,medical,social
andelectoral geographyis of this type –
though such sub-disciplines are not con-
strained to spatial scientific approaches.
A further justification for macro-/meso-
approaches lies in the probabilistic nature of
much of our understanding and representa-
tion of the world. It is generally accepted that
smoking causes lung cancer, for example, but
not all smokers get lung cancer and not all
people who get lung cancer have smoked. It
is a probabilistic relationship. Part of the rea-
son why we cannot say that smoking causes
lung cancer without qualification is because
other variables can either accelerate or decel-
erate, even block, the processes involved;
part also reflects our incomplete knowledge
(constrained because of the difficulties of con-
ducting controlled experiments). Similar argu-
ments apply to a whole range of behaviours
studied by (physical and well as human) geog-
raphers: the processes involved are so complex
and difficult to unravel – because we cannot
conduct experiments (although a range of new
techniques may be of value in such situations:
Sherman, 2003; Dunning, 2008). Hence work
in spatial science is almost of necessity condi-
tional: it represents the state of our current
understanding and can only be phrased in
probabilistic rather than deterministic terms.
The meso-/macro-patterns in many aspects
of contemporary society that are quantitatively
described and analysed by social scientists are
relevant to people’s daily lives and experiences- the ‘meaningful nature of life’ explored in
other geographical practices. Levels of ethnic
segregation in urban neighbourhoods and
schools (seeurbanization) provide contexts
within which not only lives are (partially)
lived, but also people’s life chances and
relationships are influenced. Geographical
concentrations ofpovertysimilarlystructure
manypeople’s life chances, while biased elec-
tion results reflect the operation of the aggre-
gation and scale issues underpinning the
modifiable areal unit problem (Johnston and
Pattie, 2006).
Alongside their intrinsic interest and im-
portance to understanding spatial patterns
and behaviour, therefore, meso-/macro-scale
work is relevant topublic policies– either
studies of their impact on geographies or
analyses of geographical patterns that call for
public intervention. Geographies ofmortal-
ity, for example, may identify areas where
further investment of medical resources is
warranted. Much public policy has impacts –
direct and indirect – on topics of interest to
Gregory / The Dictionary of Human Geography 9781405132879_4_Q Final Proof page 610 31.3.2009 7:14pm Compositor Name: ARajuQUANTITATIVE METHODS