The study of gentrification has never really specified a definition in the true sense of the

word; a statement that precisely delimits its nature. Rather, gentrification has been

studied in a way that has resulted in an emergent definition arising which has become

assumed rather than defined in such a way that a criteria for its existence may be

applied. This is cause for concern; the lack of some form of benchmark means that it is

possible to distort the meaning of the word and indeed question whether the word has in

fact any connotative meaning at all! Common elements to be found in the literature

reveal gentrification often to be a male, white, professional, owner occupier and inner

city process. This leads to an emergent definition which guides enquiry to these

processes and not others. It may be argued that the class replacement and displacement

dimension of gentrification has been left in the background while researchers have

examined those processes which most closely match the emergent definition of



By focusing on the nature of the process rather than typical symptoms of the process it is

possible to understand that gentrification may occur as easily in the suburbs or rural

environment as in the inner city, that it could be constituted of black middle classes as

white. What has actually happened is that gentrification has been defined in terms of its

most likely or frequent occurrence leading to a stereotypical theory of the ‘archetypal’

case rather than an understanding of its contingencies and varieties of its manifestation.


Levels of analysis in the study of gentrification

As Williams notes ‘gentrification is a complex and varied process which can be

conceptualised at a number of different levels’ (p65, 1986). Williams argues that the

dominant mode of analysis in the study of gentrification has largely been at an empirical

level which has lacked an appreciation of the processes involved.


Levels of analysis are more clearly demarcated when carrying out research proper, for

example the census has provided a key methodological tool (Galster 1986, Bourne,

1993) for gentrification research in the past. The smallest level of analysis in the British



census is the ED (enumeration district) that consists of only two hundred households but

at this level measurements of social variables such as class (which are ten percent

sample variables) may lead to high levels of inaccuracy due to the preservation of

anonymity. At an electoral ward level significance is stronger yet blunter in its

pinpointing of certain areas as size is much larger and variable. There is therefore an

interaction between levels of analysis and the validity and bias of research.


As an example, one can visualise a situation where the researcher hypothesises a picture

of low levels of gentrification activity while many individual households may be

moving undetected. This is an extreme hypothetical situation but worth bearing in mind,

it may be that contemporary gentrification is made up of a large number of individual

households that remain in obscurity while researchers complacently announce the death

of gentrification. While theory can remain aware of gentrification activity it may be

more difficult to operationalise a definition and provide empirical evidence for such



As can be seen in the discussion in the first part of the chapter, the definition of

gentrification is made at a fundamental micro level of analysis with specific reference to

an abstract household’s movements. This is not the same as saying that the study of

gentrification may only be carried out at this level, rather, it shows the microfoundations of what may form a much wider phenomenon. It is precisely this

examination of the basis of gentrification that may lead to a better understanding of the

forces at work behind its outward appearance and from which bigger units of analysis

may be built.


Recognising that ‘levels’ of analysis exist is important in structuring accounts and

theories of gentrification. Loretta Lees (1994) has noted that gentrification can be

studied at three distinct levels; nation, city and locality. Lees observed the areas of

property transfer in understanding national differences in gentrification between London

and New York at these different levels. What is particularly interesting about Lees’

work is that it demonstrates the way accounts may differ according to the level at which

analysis is carried out. Focusing on broad aggregated levels such as national data sets for

example may reveal very different and divergent pictures to research that looks at a

micro level. While the revealing of the dynamics and population involved may be better

understood within a locality this does not suggest that all research should be pinned at

this level.


It is clear that wide approaches cannot exist in isolation from an approach which

observes the phenomenon directly. Such levels of analysis and conceptual headings

could be extended and more widely applied in understanding differences and similarities

between other contexts. Problems do exist however in such analysis, Dangschat (1991)

has shown that it is immensely difficult to understand the interrelationships and

directions of causality between different levels of analysis since the interpretation of

those directions may often be open to question.


Levels of analysis may also be culturally bound, for example a regional level may be

more useful in a European context as used by Dangschat but would more likely be seen

as a city level in America or in Britain as used by Lees whose attention is directed

between these latter two countries.


Little consideration in the literature has been given about what scale of gentrification

activity should be considered a defining characteristic; if a middle class couple move in

to a working class home this by definition is a case of gentrification but as a single case

is unlikely either to be considered gentrification as popularly conceived (as a group

phenomenon) or identified through the methodological tools available to us. The area

needed to be able to study gentrification may often be bigger than the area needed to

fulfil the requirements of the definition since, by definition, one instance may count as

an ‘act’ of gentrification while certain research may need larger samples to observe the



If one defines gentrification in the way offered above it suggests that it may occur at any

level, from a micro to pan-global level at which contingent factors could be vastly

different. Writers have also acknowledged for some time that gentrification by

‘pioneers’ has paved the way for more cautious and investment seeking gentrifiers.