The Dictionary of Human Geography

Comp. by: VPugazhenthi Stage : Revises1 ChapterID: 9781405132879_4_M Date:1/4/
09 Time:15:19:41 Filepath:H:/00_Blackwell/00_3B2/Gregory-9781405132879/appln/3B2/
revises/9781405132879_4_M.3d

and indirect costs linked with modernization

processes, attention has also moved to forms

of contestation andresistancethat accom-

pany and often undermine modernization

processes (Scott, 1985; Nabudere, 1997).

More recently still, the term has been recast

asreflexive modernizationin the work of Ulrich

Beck and others (cf.reflexivity). Rather than

following a preset and, indeed, general path

(see modernity), modernization is now

thought to entail the ability constantly to

adapt and thus to create a ‘second modernity’.

Where older modernization discourses argued

with reference to timeless and abstract prin-

ciples, a newer and reflexive practice seeks

constantly to review its goals and practices in

dialogue with incoming information. Key in

this attempt is the geographical term ‘bound-

ary’, which is seen less as a fixed entity than a

conscious practice (Shields, 2006): just as

boundaries emerge from processes of negoti-

ation and remain open to future negotiations,

reflexive modernization both relies on the

boundedness of its practices while acknow-

ledging their potential for change. Structurally

similar to recent attempts to broaden the

notion of ‘modernity’ by incorporating non-

Western practices, reflexive modernization

thus constitutes a discursive process, rather

than marking the outcome of discursive prac-

tices, as older notions of modernization had

attempted to do (seediscourse). It is hence

no surprise to see the term ‘reflexive modern-

ization’ often being conflated with the notions

of ‘third way’ or ‘third (or ‘new’) modernity’.

Likewise, the realization that modernity may

well have become a problem in its own right

shares the key assumption of a required dialo-

gical openness with reflexive notions of mod-

ernization. Principally, such problematization

takes place through the recognition that unin-

tended consequences within industrial soci-

eties, which proliferate asrisksbeyond the

accepted certainties, are created within mod-

ernity. In this form, ‘reflexive modernization’

has been heralded by many as an agency-

orientated alternative to postmodern dis-

courses and attitudes (Beck, Bonss and Lau,

2003) and is related tostructuration theory

(Beck, Giddens and Lash, 1994). us

Suggested reading

Alexander (1996); Beck (1997); Galloway

(2005); Gleeson (2000); Nabudere (1997);

Shields (2006).

modifiable areal unit problem (maup) A

particular form of ecological fallacy

associated with the analysis of spatial data

sets, in particular those in which data on indi-

vidual observation units (such as households)

are aggregated into areal units (such ascensus

tractsand counties) for analysis.

Robinson (1950) provided the classic

expose ́of the maup for social scientists (see

also Gehlke and Biehl, 1934), in which he

showed that a high correlation between two

variables (the percentage of the population

who were African-American and the percent-

age who were illiterate) at the statescale

within the USA was not replicated at the indi-

vidual level; in other words, whereas the aggre-

gate data analysis showed that states with more

African-Americans also had more illiterates,

leading to the ecological inference that African-

Americans were more likely than non-African-

Americans to be illiterate, this strong claim

couldnotbesubstantiatedby analysesofindivid-

ual level data. (Thecorrelationcoefficient for

the state-level analysis was 0.946, but for the

individual level it was much smaller, at 0.203.)

The maup was introduced to the geographi-

cal literature by Openshaw (1977; see also

Openshaw and Taylor, 1979), whose empirical

studies corroborated and extended Robinson’s.

With one data set, for example, they showed

that different aggregations (or regionalizations,

since most geographical examples involve aggre-

gations into territorial blocks) could generate

correlation coefficients covering almost the

entire potential range from1.0 to þ1.0 (in

one example, the range found was 0.73 to

þ0.98), although most distributions were lepto-

kurtic, with the majority of observed coefficients

close to the median value. They showed that

the problem is made up of two components.

With thescale effect, there is a tendency for larger

correlations to be associated with larger (and

thus almost invariably a smaller number of) spa-

tial units. Theaggregation effectrefers to the large

number of different ways in which individual

units can be combined into a set of areal units –

usually with constraints such as size and contigu-

ity (Openshaw, 1982).

The maup is important in many areas of

spatial scienceusingquantitative methods,

because it indicates the need for caution in

inferring a relationship between two variables

based on a single aggregation at a particular

scale: a result identified from one such analysis

may not be replicated exactly in another.

Openshaw and Taylor (1981) identified three

possible responses to the maup:

 it is aninsoluble problem, and so can only be

ignored;

Gregory / The Dictionary of Human Geography 9781405132879_4_M Final Proof page 475 1.4.2009 3:19pm

MODIFIABLE AREAL UNIT PROBLEM (MAUP)