Suggested reading
Ballantyne (2004); Burnett (2000); Driver (2001a);
Fernandez-Armesto (2006).
exploratory data analysis (EDA) An attitude
to quantitative methods that encourages
and licences a ‘trial-and-error’ approach. The
term was popularized by the statistician
John Tukey, who recognized two approaches
to data analysis. Exploratory approaches
uncover patterns and anomalies in the data –
he likened this to numerical detective work
whereby evidence is gathered.confirmatory
data analysis, in contrast, equates tosignifi-
cance testingand probabilistic inference, as
in a trial where evidence is put in a formal
manner and a judicial decision made ‘beyond
reasonable doubt’. The exploratory approach
is based on the notion that ‘better a good
answer to a vague question than a precise
answer to the wrong one’ and that ‘by assum-
ing less you learn more’. It has encouraged the
use and development of smoothing proced-
ures that reveal patterns in data, of diagnostic,
often graphical, tools for exposing where
assumptions are not met, and of procedures
that are robust/resistant to anomalies (out-
liers) in data. Johnston (1986a, ch. 6) argues
that such data analysis should be an integral
part of a non-positivist,realistapproach
to doing geography. Exploratory spatial data
analysis (ESDA) extends EDA to detect
spatial properties of data. kj
Suggested reading
Cox and Jones (1981); Haining, Wise and Ma
(1998).
export processing zone (EPZ) A geograph-
ically delimited territoryproviding special
facilities for foreign branch plants, using
imported inputs to manufacture commodities
for export. Plants locating within the territory
are subsidized with some combination of
infrastructure, tax advantages, relaxed labour
regulations, and eased imports and exports.
EPZs provide incentives to attract foreign
investment to low-wage countries, but also
high-wage countries (there are 300 in the
USA). EPZs typically locate near the periph-
ery of countries, reinforcing external orienta-
tion. Workers are predominantly women in
non-skilled jobs, often under draconian labour
relations. Since the first EPZ in Ireland in
1956, there has been an explosion since the
mid-1970s to over 5,000 zones, with employ-
ment exceeding 40 million, in more than 100
countries. es
extensive research Research strategies
directed towards discovering common proper-
ties and empirical regularities and making gen-
eralizations about them. Sayer (1992 [1984])
argued that extensive research is typically con-
ducted under the signs ofempiricismorposi-
tivismand relies onquantitative methods,
including descriptive and inferential statistics
and numerical analysis, and onquestion-
nairesand formalinterviews. As such, it is
concerned with ‘representative’ studies or
samples and privileges a logic of replication:
Can the results of the study be repeated? Sayer
regarded extensive research as weaker than
intensive research, which is typically con-
ducted under the sign ofrealism, because
it elucidates formal relations of similarity
or correlation rather thancausalorstructural
relations. dgexternal economies Closely related to the
concept ofexternalities, external economies
are economic benefits that derive from sources
outside an organization, such as a firm. These
benefits, which can include the contributions
of specialist suppliers, subcontractors or
skilled workers hired from the locallabour
market, accrue to individual companies even
though they are generated elsewhere. In cases
where they are not available, companies must
bear the costs of producing them internally.
The benefits of external economies are often
captured locally, declining with distance. An
early formulation of this argument came in the
shape of thegrowth poletheory developed
by the French economist Franc ̧ois Perroux in
the late 1940s (see Darwent, 1969), which was
influential inregional policydebates in the
1960s and 1970s. Growth poles were geo-
graphical concentrations of economic activity,
often dominated by a single industry or a
closely related group of industries, in which
local firms yielded the economic benefits (or
positive externalities) of co-location with other
firms. In this context, the growth of individual
firms might stimulate business amongst sup-
pliers (through so-called backwardlinkages)
and/or amongst users of outputs or services
(through forward linkages). Driven by propul-
sive industries, successful growth poles are
characterized by mutually beneficial, cumula-
tive economic growth. These arguments recall
Alfred Marshall’s account of ‘industrial dis-
tricts’ in nineteenth-century England: in
Sheffield’s cutlery quarter, for example, firms
clustered together to capture benefits of
shared access to critical factors of production
such as skilled labour and technical knowledgeGregory / The Dictionary of Human Geography 9781405132879_4_E Final Proof page 234 1.4.2009 3:17pmEXPLORATORY DATA ANALYSIS (EDA)