Ecological Economics 28 (1999) 337 – 345

METHODS

Seeking sustainability discourses with Q methodology

John Barry a, John Proops b,*

a
Department of Politics, Keele Uni6ersity, Keele ST5 5BG, UK
b
Department of En6ironmental Social Sciences, Keele Uni6ersity, Keele ST5 5BG, UK

Received 9 July 1997; received in revised form 13 January 1998; accepted 30 April 1998

Abstract

This paper outlines the potential benefit of Q methodology for the study of environmental issues within ecological
economics, based on research into LETS (Local Employment and Trading Systems) in the United Kingdom. Q
methodology is a qualitative but statistical approach to enable the discovery of a variety of discourses concerning how
individuals understand their behaviour, and how they understand the social and environmental worlds in which they
live. Q methodology thus has the capacity to allow a more effective form of policy making and implementation
process. The paper introduces some of the basic tenets of Q methodology, illustrates these with results drawn from
the study mentioned above, and concludes by indicating the potential this methodology holds for ecological
economics research. © 1999 Elsevier Science B.V. All rights reserved.

Keywords: Q methodology; LETS; Sustainability; Environmental policy; Discourses

1. Introduction introduction of carbon taxes); (3) implement these

policies.
Environmental policy making has three main In ecological economics we are rather good at
stages: (1) identify the problem (e.g. ozone layer steps (1) and (2), but step (3) gets relatively little
depletion, global warming); (2) use theoretical treatment in the literature. In particular, the issue
analyses to find potentially effective responses to of what makes environmental policies socially and
these problems (e.g. banning the use of CFCs, the politically acceptable is almost unexplored. This
brief paper is a contribution to filling this lacuna.
* Corresponding author. Tel.: + 44 1782 583160; fax: + 44 Our focus is on a vital issue in environmental
1782 584144; e-mail [email protected] policy; that of identifying how individuals ‘think

0921-8009/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved.
PII: S0921-8009(98)00053-6
338 J. Barry, J. Proops / Ecological Economics 28 (1999) 337–345

about’ environmental issues. We regard this as and correlated, so as to extract ‘idealized’ forms
of central importance, because until we know of discourse latent within the data provided by
the ‘discourses’ people use about the environ- the individuals involved in the study. (A first
ment, it will be very hard to judge what, and step to exploring such discourses in the context
whether, environmental policies will be socially of ecological economics has been made by
acceptable, and therefore capable of being im- O’Hara (1996), with an exploration of the role
plemented. Indeed, finding out how people un- of ‘discursive ethics’ in environmental policy-
derstand an issue is essential to the whole making and environmental valuation.)
process of ‘problem identification’, both norma- A recent discussion of environmental dis-
tively and politically. We outline one very effec- courses is by Dryzek (1997), who offers a basic
tive approach (Q methodology) to exploring four-fold categorization, based upon the two di-
such environmental discourses, and give a brief chotomies: Reformist–Radical, and Prosaic–
overview of some of the discourses we discov- Imaginative. From these he derives the four
ered in a recent project (Barry and Proops, categories shown in Table 1.
1997). Dryzek extends his analysis to offer a range
of sub-categories in his analysis of environmen-
tal discourses, including: Prometheanism, Ad-
2. Discourse analysis ministrative Rationalism, Democratic Pragma-
tism, etc. However, Dryzek establishes these cat-
By a ‘discourse’ we mean a way of seeing and egories of discourse on the basis of, principally,
talking about something. For example, some historical and political theoretic discussion,
people may see environmental problems as being rather than empirical analysis with Q methodol-
caused by large-scale and impersonal forces, ogy.
over which they have no individual control; i.e.
they feel ‘disempowered’. Others may see the
world as being more responsive to their inputs 3. Q methodology
to it, feeling that ‘they can make a difference’.
Others again may feel that environmental prob- Q methodology is a relatively little-known
lems are not really very important, with other form of research methodology within social sci-
social ills, such as unemployment and urban ence, even though it has been established for
problems, being far more pressing. All three of over 60 years. It was invented by the psycholo-
these are very simple examples of ‘ways of see- gist William Stephenson in the 1930s, and most
ing’, of ‘world views’, or of ‘discourses’. The applications of Q methodology have been within
essence of these discourses is that they are indi- psychology (Stephenson, 1953), although Q
vidual and subjective; they represent the way a methodology has been increasingly used in other
particular individual, in particular circumstances disciplines, such as political science, particularly
and at a particular time, relates to, and forms in the USA (Brown, 1980). In essence, Q is an
conceptions of, certain aspects of the world. Of attempt to analyse subjectivity, in all its forms,
course, over time individuals’ discourses may al-
ter. Also, between individuals with shared expe-
riences and personal attributes, there may be Table 1
some degree of commonality or sharing of dis- Dryzek’s classification of environmental discourses
courses. Indeed, discourse analysis is not mainly
concerned with individuals’ discourses, but Reformist Radical
rather with the nature of shared perceptions; i.e.
Prosaic Problem solving Survivalism
with social discourses. Imaginative Sustainability Green radicalism
The strength of Q methodology is precisely
that it allows individual responses to be collated Source: Dryzek (1997) (p. 14).
 J. Barry, J. Proops / Ecological Economics 28 (1999) 337–345 339

in a structured and statistically interpretable form. 5. Implementing Q methodology

It represents an attempt to study subjectivity in an
organised manner. Q methodology, as usually practiced, is in six
The basic distinctiveness of Q methodology is stages:
that, unlike standard survey analysis, it is inter- First, one identifies the areas of ‘discourse’
ested in establishing patterns within and across which one wishes to explore, and the relevant
individuals rather than patterns across individual population. In our case, this was the attitudes
traits, such as gender, age, class, etc. As Stainton- towards the environment by members of several
Rogers (1995) (p. 180) notes ‘‘It is not, however the Local Employment and Trading Systems (LETS)
‘constructors’ —the participants — who are the fo- groups, as discussed below.
cus of the approach but the ‘constructions’ them- Second, one holds structured interviews with a
selves’’. What Q methodology attempts to elicit are sample of the relevant population. From these
the variety of accounts or discourses about or interviews one obtains a series of statements about
around a particular discourse domain, theme, issue the area of interest. These statements are the basis
or topic. for the later analysis, perhaps supplemented by
It is not generally the case that there are as many statements taken from other media (e.g. newspa-
discourses as there are participants; Q operates on pers). This approach to statement generation is
the assumption of ‘finite diversity’. Q allows the taken so that the research is focused on issues
researcher to see if there are any patterns shared which are mostly or wholly raised by the partici-
across individuals, and what are the diversity of pants, rather than the researcher.
accounts, without this resulting in chaotic multipli- Third, one makes a selection from these state-
cation. There are a limited number of ordered ments, for use in the ‘Q sorts’; i.e. one establishes
patternings within a particular discourse domain, a set of statements to which participants are asked
and Q works on this assumption and attempts to to respond, as discussed below. One has to decide
reveal those ordered patternings (factors or dis- upon the number of statements to present to
courses) in a structured and interpretable manner. participants, and we found 36 statements to be
It is, therefore, particularly suited to studying those manageable, both for the participant and the re-
social phenomena around which there is much searcher. (Details of the statement selection pro-
debate, conflict and contestation, such as the envi- cess are given below.)
ronment, for its express aim is to elicit a range of Fourth, the participants are asked to rank the
voices, accounts and understandings. statements on the scale ‘Agree with most strongly’
to ‘Disagree with most strongly’. This set of ranked
statements constitutes the ‘Q sort’ for that individ-
4. Sources on Q methodology ual. (Details of the statements used for the Q sorts
in our study, and the ranking procedure used for
The basic text is Stephenson (1953). A very these statements, are given below.)
accessible introduction (mainly for political scien- Fifth, from these Q sorts, statistical analysis
tists) is Brown (1980). A useful ‘how to do it’ book allows the extraction of a few ‘typical’ Q sorts,
is McKeown and Thomas (1988). There are also capturing the common essence of the several indi-
three websites worth exploring. There is a ‘Q vidual Q sorts. Each individual usually has aspects
Method Discussion List’ at QMETHOD@ of several ‘typical’ Q sorts contained in their
LISTSERV.KENT.EDU. A comprehensive ‘Q particular Q sort.
Method Page’ is at http://www.rzunibw-Muen Sixth, these typical Q sorts must be interpreted
chen.de/ p4bsmk/qmethod/, which has down- verbally, to give the social discourses uncovered by
loadable Q methodology programmes available. A the statistical analysis. These discourses are gener-
rather smaller ‘Q methodology archive’ at http:// ally not closely represented by any particular indi-
www/uww.edu/PoliSc/QINEX.HTM ctop is also viduals; instead, they represent a ‘pure’ or ‘ideal
worth exploring. type’ version of a way of seeing the world.
340 J. Barry, J. Proops / Ecological Economics 28 (1999) 337–345

6. Our study ing (but of course not completely eradicating)

researcher bias. The broad categories used were
The details reported here derive from a study ‘discourse element’ and ‘type of claim’, with each
we made of the attitudes of members of Local being further subdivided as follows:
Employment and Trading Systems (LETS) in the
UK. Our study involved three distinct Q sorts, 7.1. Discourse element
exploring attitudes towards: LETS systems them-
selves; ‘Citizenship and Community’; and ‘Envi- Ontology: refers to sets of entities
ronmental Concern, Awareness and Sustain- (states, nations,
ability’. In this paper we concentrate on the envi- individuals, classes)
ronmental part of our study. (Fuller details of the Agency: refers to degrees of
overall study are available in Barry and Proops agency of these entities
(1997).) We recognise that members of LETS Motivation: such as self-interest,
groups are likely to hold views towards environ- public-spiritedness
mental issues rather different from the UK popu- Natural/unnatural taken-for-granted
lation at large or from, say, a representative relationships: relationships
group of bankers. The discourses we uncovered,
as discussed below, should not, therefore, be in-
terpreted as being of general applicability; they 7.2. Type of claim
are specific to this study-group.
Definitive: concerns the meaning of terms
Designative: issues of fact
7. Statement generation and selection Evaluative: expressions of the worth of
something
Having identified ‘the environment’ as an area Advocative: something that should or should
of interest for our research, we next needed to not exist
generate a series of statements on this topic which
respondents would be asked to rank by the (rela- The concourse matrix is shown in Table 2.
tive) extent to which they agreed or disagreed An example of an ontological definitive state-
with the statements. The overwhelming source for ment would be statement 4 in Table 3: ‘‘LETS is
statements was the interviews, but statements a new type of economy in which living sustainably
were also used from academic and popular litera- is a key’’. An example of an motivation–advoca-
ture; this generated hundreds of possible state- tive statement would be statement 24: ‘‘People
ments. After some initial piloting, we found 36 should take things into their own hands, like at
statements to be manageable, both for the Q the Newbury by-pass’’.
sorter and the researcher administering it. To Using the above approach we derived 36 ‘envi-
reduce our sample to 36 statements, we followed ronmental’ statements relevant to members of
Dryzek and Berejikian (1993) in employing a 4× LETS schemes. These formed the basis of the Q
4, 16-cell ‘concourse matrix’ to sample the avail- sorts, administered to our 25 participants. These
able statements. The aim of using this matrix was statements are shown in Table 3.
to filter the statements so as to leave us with what
can be termed the quintessential statements about
the discourse area we were interested in investigat- 8. The nature of the Q sort
ing. This filtering device (also commonly em-
ployed in political discourse analysis) not only Participants were required to sort the 36 state-
reduced the available statements into a manage- ments in Table 3, according to how strongly they
able Q sort, but also acted as an independent agreed or disagreed with them. A nine-point scale
method for choosing statements, and so minimis- was used, viz.: −4 − 3 −2 − 1 0 +1 + 2 + 3
 J. Barry, J. Proops / Ecological Economics 28 (1999) 337–345 341

Table 2
The concourse matrix

Type of claim Discourse element

(1) Ontology (2) Agency (3) Motivation (4) Natural/Unnatural

5. Definitive
6. Designative
7. Evaluative
8. Advocative

Source: Dryzek and Berejikian (1993) (p. 52).

+ 4. Here −4 corresponded to ‘‘Disagree with methodology. In its use one enters each Q sort in
most strongly’’ and +4 to ‘‘Agree with most the study as data. The package then correlates
strongly’’. It should be stressed that this is a each Q sort with every other Q sort. This intercor-
relative, not absolute, scale. It may be the case relation matrix is then factor analysed by the
that a participant agrees with all of the state- centroid procedure. The resultant factor analysis
ments; even so, a ranking is still possible. The is rotated to a ‘simple structure’ (using varimax
ranking of the statements by an individual is rotation) to extract factors which are significant
known as that individuals ‘Q sort’, and reflects the by the protocols of Q methodology. The package
individual’s response to the statements. extracts all significant factors, and re-expresses
There was a limit placed on how each statement them as the ‘best estimate’ of the Q sorts that
was scored. In particular, the number of state- represents them. Thus the analysis produces an
ments corresponding to each strength of agree- ‘ideal type’ around which those Q sorts which
ment was fixed beforehand. The number of come closest to this ideal are listed. Thus each
statements that had to correspond to each element factor is represented by an ‘ideal-type’ or ‘simple
of the scale was as follows: structure’ Q sort.

Scale score −4 −3 −2 −1 0 +1 +2 +3 +4
No. of statements 2 3 4 5 8 5 4 3 2

The ‘pyramidal’ structure of this Q sort is typical, The significance of a factor can be determined
and is used to encourage the participants to give statistically by employing the eigenvalue criterion.
careful consideration to the ranking they wish to By convention, factors with eigenvalues greater
achieve, while not demanding that the ranking be than 1.00 are considered significant.
complete. It should be stressed that the imposition The four ‘ideal type’ Q sorts we found are
of the above structure on the ordering is not shown in Table 3. For example, the first extracted
required by the statistical technique, nor, it seems, Q sort (marked ‘A’) has statement 1 registered as
are the elicited discourses much influenced by the + 2 (i.e. agree moderately), while the second ex-
precise scoring structure (McKeown and Thomas, tracted Q sort (marked ‘B’) has statement 1 regis-
1988). tered as + 1 (Agree Weakly), etc.
All of the extracted Q sorts expressed a view
that multinationals were an environmental threat,
9. PCQ and factor analysis and that society was profligate and consumption-
driven, felt the government ought to do more
To analyse the Q sorts, and extract the underly- about environmental problems, and that global
ing ‘ideal type’ Q sorts, we used the PCQ software environmental issues are of individual concern
package which is specifically designed for Q even if they are not under one’s control. This
342 J. Barry, J. Proops / Ecological Economics 28 (1999) 337–345

Table 3
Statements, with scores on each of the four extracted discourses

Statement A B C D

1 I try to be ‘green’ in the things I buy +2 +1 +2 −1

2 I would not describe myself politically as green 0 +2 −3 +1
3 LETS should aim to create ‘sustainable livelihoods’ +2 −1 0 +2
4 LETS is a new type of economy in which living sustainably is a key aspect +2 −1 +1 0
5 We need to overcome the distinction between ‘work’ and ‘leisure’ if we are to create a more −1 0 +3 +4
balanced, humane and sensible economy
6 The root of the ecological crisis is greed and therefore money 0 +4 +3 0
7 The problem in our society is that we waste resources just because people can afford to do so +1 +1 0 −1
8 I am not very concerned about the environment −4 −2 −4 −2
9 Multinationals are not a grave threat to an environmentally sustainable society −3 −4 −4 −3
10 I would not be willing to pay an extra penny in tax to pay for environmental improvements −4 −2 −2 +1
11 Development has gone too far in the UK +1 +1 0 0
12 Cars are the biggest cause of environmental and health problems −1 +1 +1 0
13 I do not recycle as much as I should 0 0 −2 +3
14 Our society is profligate and consumer-orientated/consumption-driven +3 +3 +3 +2
15 I do not see LETS as a way of putting green ideas into practice −2 +2 −2 −1
16 The majority of people on LETS are ecologically minded −1 −1 −1 −2
17 In the country you are more aware of how dependent we are on he natural world 0 0 −1 +1
18 The world can support a lot more people than it does at present +1 0 −1 −3
19 People need not feel more concerned about global environmental issues as these are not under −3 −3 −3 −2
their control
20 There are a lot of environmental cover-ups 0 0 +2 +1
21 People are taking a short-term view: they’re not thinking about the long-term effects of what +2 +3 +1 +2
they’re doing
22 You can’t look at one part of the planet, because all the parts interact +4 −1 +1 0
23 It’s not necessarily in the nature of humans to want more and more +1 0 0 −3
24 People should take things into their own hands, like at the Newbury by-pass +1 −3 +2 0
25 LETS have no environmental benefits −2 0 −3 −4
26 The cause of the environmental crisis is big business in all its forms 0 +4 +4 −4
27 The government should take responsibility for legislating on environmental issues a great deal +3 +3 +4 +4
more than it does
28 It’s barbaric to breed animals to eat −1 +1 0 −1
29 I see technology as progress −2 −4 −1 +3
30 In the future people are going to have to lower their material standards of living due to resource +4 0 0 +2
shortages and other environmental pressures
31 The damage we are doing to the planet is beginning to come to the fore in public awareness 0 −2 +1 +3
32 We all have to take responsibility for environmental problems +3 +2 +2 +1
33 The environmental benefits of LETS are important to me −1 −2 0 0
34 I’ve not been made more aware of environmental issues since joining LETS −2 +2 0 0
35 I believe most future environmental problems will be solved by technology −3 −3 −2 −1
36 I think there is a trade-off between unemployment and environmental quality 0 −1 −1 −2

‘global concern’ is, of course, a long-standing random sample of UK residents.

‘green’ or ecological position, perhaps indicating
that, in various ways, this study found a global
sense of environmental responsibility, as well as a 10. The four discourses on the environment
sense of environmental responsibility in which
government has a large part to play. Clearly, The four discourses extracted from the environ-
members of LETS groups are likely to have dif- mental part of the study, as we interpret the data
ferent perspectives on environmental issues than a from the statistical analysis, were as follows.
 J. Barry, J. Proops / Ecological Economics 28 (1999) 337–345 343

10.1. Discourse A: techno-sceptical, non-green distinctive from the others in that it expressly
holism disagrees with a ‘direct action’ form of political
activism. Perhaps this account sees the role of
For this discourse, the following statements are government as keeping big business in check, to
particularly important: agreement, 14, 22, 27, 30, make sure it does not create environmental prob-
32; disagreement, 8, 9, 10, 19, 35. What is distinc- lems and threaten ecological sustainability.
tive about this account is its holistic, intercon-
nected, ecological view. This discourse displays a 10.3. Discourse C: political ecologism
strong concern about the environment. There is
also strong agreement that in the future there will The statements most influential for this dis-
have to be a lowering of material standards of course are: agreement, 5, 6, 14, 26, 27; disagree-
living, related to our profligate and consumption- ment, 2, 8, 9, 19, 25. This discourse expresses an
driven society. It is committed to environmental anti-capitalist ecologism, is anti-multinational,
issues, yet has no strong political identification as like Discourse B, and sees big business, money
‘green’. There is also strong support for the idea and greed as the causes of the environmental
that, while we all have to take responsibility for crisis. What is distinctive about this discourse is
environmental problems (perhaps in the sense of that this is the only one that can be characterized
moral responsibility), the government has a statu- as politically ‘green’. It is less anti-technology
tory responsibility for taking care of environmen- than the previous two discourses, but only by
tal problems, a responsibility which it neglects. degree. It is very strong on the idea that the
LETS are seen as having environmental benefits government should take responsibility for legislat-
directly, but also indirectly, in terms of educating ing on environmental issues more than it does, as
people. Most future environmental problems will is Discourse D. Also, this discourse stresses the
not be solved by technology, and technology is need to overcome the distinction between ‘work’
not seen as progress. and ‘leisure’ if we are to create a humane and
balanced economy. It stresses the environmental
10.2. Discourse B: anti-capitalist, benefits of LETS. Unlike the previous discourse,
techno-scepticism, non-green ecologism this discourse is in agreement with the suggestion
that people should take things into their own
For Discourse B we find the following state- hands.
ments to be particularly important: agreement, 6,
14, 21, 26, 27; disagreement, 9, 19, 24, 29, 35. This 10.4. Discourse D: pro-technologism,
discourse emphasizes the causes of the ecological acquisiti6eness
crisis. Overall, this account lays the blame on
business and the current economic system. The For Discourse D, the most important state-
root of the ecological crisis is greed and therefore ments are: agreement, 13, 27, 29, 31; disagree-
money and big business in all its forms; multina- ment, 9, 18, 23, 25, 26. This discourse stresses the
tionals are seen as a grave threat to an environ- notion that the government should take more
mentally sustainable society. While this discourse responsibility for environmental issues, and also
is pro-ecological, it cannot be described as a the need to overcome the distinction between
politically ‘green’. LETS is not seen as a way of work and leisure if we are to create a balanced
putting green ideas into practice, and the environ- and humane economy. However, what stands out
mental benefits of LETS are not important to about this discourse is that, unlike any of the
those who hold this view. This account expresses other accounts, this one agrees with the statement
an even more extreme techno-scepticism than Dis- that technology is progress. At the same time it is
course A, and also displays strong disagreement not believed that future environmental problems
with the claim that future environmental prob- will be solved by technology. So while technology
lems will be solved by technology. This account is is progress, this does not mean that it can over-
344 J. Barry, J. Proops / Ecological Economics 28 (1999) 337–345

come all problems. It also stands out because it 36 statements, the minimum number of judge-
strongly disagrees with the statement that the ments necessary to provide a partial ordering on
cause of the environmental crisis is big business. the nine categories defined above is of the order of
This discourse articulates a view of humans being 86.1 Thus most of the data in Q methodology
naturally acquisitive. It prefers to leave the derives not from the number of participants, but
achievement of environmental improvements to from how much information is implicit in each
individual actions. participant’s Q sort.
The final benefit is that, as far as we are aware,
Q is unique in being largely driven by the partici-
11. The costs and benefits of Q methodology pants’ responses (remembering that the original
statements derive from the participants),
Having outlined Q methodology, and given a while still generating statistically rigorous
brief indication of one environmental application, results.
we now outline the costs and benefits of this
approach.
The first cost of Q is that it is time intensive for 12. Q methodology and ecological economics
the researcher. Having identified the area of inter-
est, and the corresponding population, one has to As mentioned in Section 1, ecological econom-
spend considerable time on the initial stage of ics is not strong in the analysis of how environ-
structured interviewing to elicit statements that mental policies can be implemented. The vital
can be considered for inclusion in the set of aspect of how the ‘public’ views environmental
statements for the Q sort. The process of extract- issues and policies has been barely explored, and
ing the appropriate number of statements, satisfy- it is here that we believe Q methodology can have
ing the criteria mentioned above, is also time an important role to play. Using Q, researchers
consuming. Finally, the Q sorts also take time, as can seek to explore people’s attitudes towards
participants usually need quite detailed advice on nature, in a way that is responsive to the attitudes
strategies for sorting the statements. held by the respondents, rather than the re-
A benefit of Q is that relatively few partici- searchers, while still having a rigorous statistical
pants, and Q sorts, are needed to give statistical basis for the extraction of the discourses within
significance. As few as 12 participants can gener- the population.
ate statistically meaningful results, in terms of the It is on the basis of such ‘environmental dis-
range of implicit discourses uncovered. The rea- courses’ that we might hope to construct socially
son for this is that each participant’s Q sort acceptable and effective environmental policies. If
provides a very large amount of information. it is possible to identify that certain groups have
With the 36 statements we used, the ranking from discourses about nature that are markedly differ-
‘most agree’ to ‘most disagree’ requires that each ent from other groups (e.g. urban/rural; male/fe-
statement be, at least implicitly, ranked against male; rich/poor; etc.), then policy makers will
every other statement. If one takes an arbitrary know that policies acceptable in one locality, or
statement, one can judge for all the other state- stratum of society, may be ineffective or even
ments whether one agrees or disagrees with it unworkable elsewhere. In addition, Q methodol-
more or less strongly. This involves 35 judge- ogy can also be part of the process of delivering
ments, and divides the set of statements into two ‘better’ environmental policies. If it allows respon-
disjoint sets: those agreed with more strongly than dents who ultimately have to live with the conse-
the original statement, and those agreed with less quences of environmental policy to be included in
strongly. Taking another arbitrary statement from
either of these disjoint sets, one then judges it
against all other statements in this set, etc. Using 1
We are grateful to Mick Common for questioning the
optimal sorting theory, one can show that for our faulty analysis of the figure in an earlier draft.
 J. Barry, J. Proops / Ecological Economics 28 (1999) 337–345 345

determining or identifying a problem (which can 13. Conclusions

then be addressed by policy makers), it is more
likely that the policy will be acceptable (or legit- In summary, we believe Q has great potential
imate) and thus more likely to be effective. Thus for ecological economics. It allows a responsive
Q may have a role to play in the development but statistically rigorous approach to the subjec-
of a more ‘democratic’ form of policy formula- tive perceptions of human–nature relationships,
tion, in that it allows those that have to live and its results could be extremely useful in in-
with environmental policies some part in deter- forming environmental policy making. We hope
mining them. that this brief introduction and report will stim-
Therefore, we believe that information from Q ulate others to explore Q as another methodo-
could assist environmental policy making in two logical tool for ecological economics research.
ways.
First, it would identify for policy makers the
ways environmental issues are perceived by vari- Acknowledgements
ous groups, allowing the identification of com-
mon issues or perspectives in the population. The research reported in this paper was sup-
Policies directed towards any such widely shared ported by the UK’s Environmental and Social
concerns would be likely to receive good social Research Council (Global Environmental Change
and political support, and be effective. While Q Programme), under grant L320253192. We are
does not directly give an indication of the rela- grateful to Mick Common, Sabine O’Hara and an
tive strengths with which the various discourses anonymous referee for comments on an earlier
are adhered to by the population at large, the draft of this paper.
discourses identified can be related back to the
individuals participating, to give at least an im-
pression of what likely adherence there is in a References
wider population to each discourse.
Second, it would become apparent if different Barry, J., Proops, J.L.R., 1997. Local Employment and Tread-
groups in society had markedly different per- ing Systems: Linking Citizenship and Sustainability? ESRC
spectives on certain environmental concerns. Global Environmental Change Research Report
L320253192.
This would suggest what policies would be likely Brown, S., 1980. Political Subjectivity: Applications of Q
to receive support, and from whom, allowing Methodology in Political Science. Yale University Press,
the policies to be formulated in a way most Yale.
likely to generate wide acceptance. For example, Dryzek, J., 1997. The Politics of the Earth. Oxford University
Press, Oxford.
a segment of society may see problems of unem-
Dryzek, J., Berejikian, J., 1993. Reconstructive democratic
ployment as far more pressing than environmen- theory. Am. Political Sci. Rev. 87, 48 – 60.
tal protection. In that case, identifying McKeown, B., Thomas, D., 1988. Q-Methodology. Sage, Lon-
environmental policies which simultaneously re- don.
duce environmental damage and increase the de- O’Hara, S., 1996. Discursive ethics in ecosystem valuation and
environmental policy. Ecol. Econom. 16, 95 – 107.
mand for labour (e.g. replacing labour taxes by
Stainton-Rogers, R., 1995. Q methodology. In: Smith, J. et al.
environmental taxes) would help ensure the en- (Eds.), Rethinking Methods in Psychology. Sage, London.
vironmental policies received wider political sup- Stephenson, W., 1953. The Study of Behaviour: Q Technique
port. and its Methodology. University of Chicago Press, Chicago.