European Journal of Political Research 31: 147–157, 1997.

147

c 1997 Kluwer Academic Publishers. Printed in the Netherlands.

NOMINATIONS AND REFLECTIONS

Left-right political scales

Left-right political scales: Some ‘expert’ judgements

by Francis G. Castles, The Open University, UK &
Peter Mair, European University Institute, Florence, Italy
EJPR 12 (1984): 73–88.

Abstract. Although left-right scales are an inherent feature of much cross-

national research, they have necessarily been created on a somewhat ad hoc
basis, since the empirical foundation for valid cross-national scales rarely
exists. This paper seeks to provide such a foundation by using judgements
of party ideological position which are both explicit and non-idiosyncratic
across a wide range of countries. These judgements derive from a so-called
‘expert’ survey of leading political scientists in Western Europe, the USA,
and elsewhere. It is our hope that the scales which we derive in this way may
prove useful in a wide variety of contexts of comparative research.

Nomination:
‘Some expert judgements’ live on
MOGENS N. PEDERSEN
Odense University, Denmark

When I nominate an article that has made ‘a particular contribution to political

science’, I do it after serious deliberations – with myself. As the founding
editor of this journal has pointed out, I happen to be the only living scholar who
has been actively involved with the EJPR from the very beginning and till this
very day. Therefore my memory must be longer than the memory of all of the
other good colleagues who take part in this academic Voxpop. Fortunately
that is not the case. An editor reads manuscripts, hundreds and hundreds,
makes decisions and communicates with authors – and tends to forget about
the rest of the process. I certainly did not re-read all the manuscripts when they
appeared in print. The journal issue found its place on the shelf together with
other tools of scholarly work. Therefore I mostly remember some difficult
148
cases and borderline rejects, some cultural clashes, and exchanges with a few
authors who did not accept entirely the wise editorial decision.
Browsing through the pages today I am struck by many good articles, which
never really caught the attention of the profession – partly, probably, because
our American colleagues did not read the journal as much as they ought to.
Next, my eyes fall upon some odd articles, like e.g., Weber’s content analysis
of British Speeches from the Throne over several centuries (Weber 1982)
and Barricelli’s radical proposal for a new type of government (Barricelli
1985). These were articles taken in exactly because they were not only good,
but also somewhat removed from mainstream political science. A number
of special issues, which brought European political scientists together for a
while, deserve mention, if not nomination. Rose & Mackie’s annual electoral
data reports also stand out as a useful, indeed highly valuable, contribution –
but hardly eligible in this context. I end up with a small number of candidates
for the nomination, but also with a sad feeling, because I see the point that
was once made by Gunnar Sjöblom about the lack of cumulation (Sjöblom
1977). The EJPR has over its span of existence provided an outlet for articles
on electoral systems, parties and party systems, political economy and a few
other issues, but apart from these fields it is in retrospect difficult to see a
policy in operation, leading to cumulation. On the other hand, the official
policy of the journal did not encourage specialization – the old editor may
just have a fit of spleen. : : :
A few contributions stand out, because they created a lively discussion. Oth-
er contributions, because they were useful for comparative scholars. Strangely
enough one author has done more than others to stimulate discussion and also
provided one of the most useful pieces that I know of. This author has spent
the last many years ‘Down Under’, and from his outpost there he has served
the journal in many capacities, including also delivering exciting articles and
polemics. Francis G. Castles is the name, and the list of articles written for
EJPR is impressive. At least ten articles, mostly on aspects of public policy
in advanced democracies, starting with the well-known ‘Does politics mat-
ter?’ (Castles & McKinlay 1979), and – until now – ending with a discussion
with Göran Therborn on the relationship between religion and public policy
(Castles 1994a, b).
My favourite among these articles appeared in 1984. It was a product
coming out of an EUI-project, directed by Rudolf Wildenmann, and the
authors were Frank Castles, then at the Open University, and a younger
EUI-scholar, Peter Mair. Around 1982 members of the ECPR Council and
assorted others received a questionnaire from these two colleagues, in which
respondents were asked to enter the parties of their own nation into a ten
– or was it eleven? – point scale which was called a ‘Left-Right’-scale. I
 149
still remember my own first reaction: ‘This is something that you cannot
do!’ Many colleagues had the same gut-feeling. But many of us did our
homework, since the two importunate scholars were such nice people, and
since they assured us that the contributions would not be used in an indecent
way.
The first premise for my nomination is audacity. Before the 1970s the
‘Left-Right’-continuum was treated in an impressionistic way (e.g., Lipset
1960). Some scholars would even warn against its use as a meaningful the-
oretical concept. This situation changed a bit during the 1970s, when survey
techniques made it possible to construct ‘Left-Right’-scales from electoral
data (e.g., Inglehart & Klingemann 1976). By means of other available tech-
niques it was also increasingly becoming possible to map dimensions of party
systems at other levels, most notably the level of the party system in parlia-
ment (see e.g., Pedersen et al. 1971). The emerging coalition theory needed
a dimensionality measure, and several scholars tried their hands during the
1970s. The burgeoning literature on policy outcomes also was in need of an
operationalization of the Left-Right dimension in order to find out ‘if poli-
tics matter?’. But still Giovanni Sartori, who eventually ended up building
much of his own analysis on a dimensional basis, would call the Left-Right-
continuum ‘a layman’s “index” of politics’ (Sartori 1976: 79) – and as such
‘a grand oversimplification resulting from a compound of fuzzy criteria’.
Comparisons across countries were in particular debatable. ‘Left’ and ‘Right’
were dubious categories!
Furthermore, in the early 1980s the systematic use of ‘expert’ judgements
was a novelty in political science, and it had never been used in party research.
Castles & Mair thus attacked a difficult methodological problem with a new
and mainly untested technical approach. But they dared to jump – and ended
up with enough responses to make an article feasible.
Paradoxically, my second premise is the caution of the authors. Not of
the editors, for although I hate to admit it, the truth is that the article was
commissioned and accepted without the usual use of external referees. For-
tunately the editors had no reason to regret their decision. The article is still
readable – and the reader is today struck with the caution and prudence of
the authors – from the very title, where ‘Expert’ is put in inverted commas,
and till the final expression of the hope that the scales ‘may prove useful in a
wide variety of contexts of comparative research’. The reader will note that
the data are reported and presented in a clear and transparent way. After a
decade of use and discussion it is obvious that some readers would have liked
to have had even more documentation. But ‘enough is enough!’ – especially
for ‘a somewhat extended research note’ as the authors modestly called their
contribution.
150
An implicit quality of the new technique was that it implicated a lot of us.
In a ‘high risk’-project it may have helped the two scholars that they were
‘spreading the unreliabilities and errors of expert judgements’, as it has been
said (Laver & Schofield 1990: 250). Nothing wrong about that!
Third premise: Usefulness. Since 1984 it has become much easier to use
the ‘layman’s “index” of politics’ in comparative analyses. One does not have
to start all over each time one addresses the dimensionality discussion. Part
of the reason for the success of the article may be found in the laziness of
scholars, but it would be unfair to stress such a point. It is more to the point to
quote the old say about ‘the proof of the pudding is in the eating’. For many
analytical purposes the crude scales of Castles & Mair have served us well in
comparative research. They tend to ‘work’ in the sense that the operationalized
dimension often correlates nicely with other important variables. They have
also made it much easier to discuss and analyze the phenomenon of Centre
parties in European politics. They have made it easier to group parties in
‘families’ across nations, and so on.
In passing it should be mentioned that the Castles & Mair findings have
been quoted and used so many times and in so many contexts that the article
has more than proved its usefulness. The 1984-scales have been used as
a baseline scale for other, similar scales (Laver & Schofield 1990: 252ff.).
Similar expert ratings have been utilized several times since then (e.g., Laver
& Hunt 1992). The 1984-study has even become replicated and improved
(Huber & Inglehart 1995). The Castles-Mair study thus is very much alive as
a contribution to the discipline of comparative politics.

Reflections:
Revisiting expert judgements
PETER MAIR1 & FRANCIS G. CASTLES2
1 Leiden University, The Netherlands; 2 Australian National University

Our original article (Castles & Mair 1984) was conceived in the European
University Institute in Florence, where we were both involved in a relatively
undisciplined but eventually very fruitful and productive research project on
‘The Future of Party Government’, directed by the late Rudolf Wildenmann.
Indeed, the fact that it proved possible to plan and execute such an expert
survey at that time owed a great deal to the somewhat chaotic network of
international contacts spawned by Rudolf’s project, as well as to the more
formalized networks which had been created and nourished by the ECPR
over the previous decade or so.
 151
Our mutual interest in generating the data on the left-right positioning
of parties was stimulated by two distinct concerns. On the one side, it had
become generally accepted that left-right differences between parties had a
major relevance for public policy outcomes, and a growing body of literature
in the late 1970s and early 1980s had been devoted precisely to the estimation
and measurement of this relationship (e.g. Hewitt 1978; Castles & McKinlay
1979; Cameron 1982; Castles 1982). On the other side, left-right differences
between parties were also seen as particularly relevant within the growing
policy-based coalition literature (e.g., de Swaan 1973), while variations in
the degree of polarization and ideological distancing were seen to be crucial
to the functioning of party systems more generally (Sartori 1976). At the same
time, however, as we pointed out in our original article, the various data from
which these differences could be assessed were remarkably inconsistent and
insensitive. They were either based on either sparse and often incomparable
survey data, or they rested on simple judgement calls, in which the individual
researcher would make an easy but often crude distinction between ‘a left’ and
‘a right’ or between one category of party family and another, and in which,
for example, all social-democratic parties could end up being assigned a
similar position in left-right terms regardless of the particular positions which
each adopted in practice.
What was needed, therefore, was a more systematic data base, in which
variations across a common cross-national scale could be compared, and in
which real differences between parties could be measured. Encouraged by
the international contacts available through both the EUI in Florence and the
ECPR, we therefore decided to elaborate such a data collection on the basis
of ‘expert’ judgements. The idea was simply to poll expert and experienced
political scientists on their own countries by use of a short questionnaire
which asked them to locate the various parties in their country on a common
left-right scale; that is, we would make the best of a bad lot by pooling as
many individual judgement calls as we could amass, and by ensuring that
those calls to be as precise as possible. The ultimate goal was to create a scale
which could be adapted for comparative cross-national research. Given the
paucity of comparable data which had existed up to then, most scholars who
needed to take account of left-right differences in different countries often
ended up by having to cull a very varied secondary literature, or by having to
write to various colleagues in pursuit of potentially idiosyncratic information.
The expert survey, on the other hand, could turn this sort of search into a finer
and more precise art, in which the same colleagues in the same countries
would locate the parties along the same scale, and thus provide the wider
community with a standardized set of judgements.
152
The approach was also reasonably novel at the time (see also Laver &
Schofield 1990: 246–248). As far as we knew, there had up to then been only
one other systematic attempt to collate expert judgments on party positions,
that carried out by Michael-John Morgan (1976) for a University of Michigan
PhD thesis, which we cited in the original article but to which we had no
access at the time. Since then, however, the procedure has become more
commonplace. Laver & Hunt (1992), for example, employed a very detailed
and comprehensive expert survey in order to probe the positions of party
leaders and party voters on a whole raft of issues, and also polled these
experts on the relative importance of the various different dimensions of policy
competition. More recently, Huber & Inglehart (1995) conducted a somewhat
more elaborated version of our own simple left-right survey across experts
in 42 countries, and here also an attempt was made to weight the various
dimensions of competition. Expert surveys have also since been employed on
a more limited basis to measure cross-national variation on themes ranging
from the extent of corporatism to party attitudes to European integration.
The growth in the use of expert surveys in the late 1980s and 1990s offers a
satisfying testimony to the validity of the approach which we adopted. what
has also been satisfying has been the extent to which the original data have
subsequently been employed in the scholarly literature, where, even now,
they are still regarded as providing an authoritative source for locating party
positions. Indeed, as is testified by its inclusion in this anniversary ‘roll of
honour’, the article eventually proved to have much more impact than we
could have imagined. Had we known it was likely to have had this impact,
however, and in much the same spirit as the parents of a one-time president
of the National Union of Students in Britain (who were once reported to have
said that ‘If we’d known he would become famous, we’d have called him
something different’), we would have done it somewhat differently. Or, at
least, we would have done it more carefully.
For example, and in this regard we have certainly been put to shame by the
recent exercise by Huber & Inglehart (1995), we simply did not make enough
effort to increase the number of countries/cases. Data for only 17 countries
are actually reported in the original article, with four countries on our original
list (Greece, Iceland, Portugal and Switzerland) being excluded because of
an insufficient response from local experts. In retrospect, more effort should
have been made to get sufficient data to include these missing cases, as well
as to include the various other competitive democracies for which we made
no effort at all (e.g., India, Israel, Japan, Malta, etc.). In addition, we also
mainly contented ourselves with too few experts. With the exceptions of
Germany, Sweden, Italy, and the UK, none of the country scores was based
on a double-figure response, and in some cases we received just four or five
 153
responses. These numbers are actually very small, and given that the range of
scores per party in each country was sometimes quite extensive, and that not
all relevant parties were scored by all respondents, results were not perhaps
as authoritative as would be desired or as they have been subsequently taken
to be.1 Both of these problems could have been remedied by a greater effort
on our part. That said, and in our defense, it must be pointed out that the
sheer novelty of the exercise meant that we inevitably lacked experience, and
many of these problems were simply not appreciated at the time. Moreover,
while networks of scholars to whom we could turn certainly existed in 1983,
contacts were not so easily established as is now the case. Finally, we were
also limited by financial considerations, in that almost the entire cost of the
exercise was borne by the day-to-day services of the European University
Institute.2
Three other shortcomings can also be noted. First, the ‘questionnaire’ (it
was really just one question) which went out to the various experts did not
include a pre-defined list of parties. Rather, respondents were simply asked
to locate ‘the parties in their system’, or some such formula. This was a
mistake since, as noted above, a number of respondents missed out certain
parties, either because the ‘parties concerned were too small, or perhaps
even because they were too difficult to locate. Either way, for the sake of
authoritativeness and comprehensiveness, it would have been better had all
respondents attempted to locate all relevant parties, and this could have been
effected had we included our own definitive list. Second, being inexperienced
in these matters, we decided that a ‘10-point scale’ – the obvious measure to
use – should run from 0 to 10. In fact, and as has been subsequently pointed
out to us ad nauseum, it should have run from 1 to 10. Thus, for example,
comparability with the subsequent Huber-Inglehart scale requires one or other
of the data sets to be adjusted (see also below). Third, in order to facilitate
subsequent replication, we should have included a copy of the ‘questionnaire’
as an appendix to the article in the EJPR. We might even suggest that the then
journal editors, Mogens Pedersen and Derek Urwin, who were good enough
to promise to accept the piece almost sight unseen, should themselves have
insisted that a copy be published with the article.3
Having the questionnaire to hand might also have encouraged us to repeat
the study at a later stage. In retrospect, it would have proved very useful
to have updated the information at the end of the 1980s, although we were
discouraged from doing this by the knowledge that Laver and Hunt were
preparing a much more extensive expert survey at the time. And while we
were actually making plans to repeat the exercise in 1994 – more or less a
decade later – we discovered at the last moment that Huber and Inglehart had
just conducted a much more broadly-based equivalent. All of this is to the
154
good, and whether on a formally coordinated basis or not, it seems eminently
desirable to continue to have regular updates in the future. As the subsequent
citation of our own 1984 article confirms, these left-right placings clearly
constitute an invaluable resource across quite a wide variety of fields within
the discipline. At the same time, however, for the reasons cited above, as well
as because this was simply a snapshot survey, these data should perhaps be
cited more cautiously than is presently the case, and they should certainly
not be automatically accorded a validity which extends both long before,
and long after, their actual application. As noted elsewhere (Mair 1996: 325–
328), comparative political scientists have a tendency to fetishize whatever
few indicators happen to be available, and these particular data have certainly
not escaped that fate.
This opportunity to reflect once again on our original article has also afford-
ed us the chance to compare briefly the patterns in 1984 with those which
are now provided for the same set of countries in the 1995 data of Huber
and Inglehart. Summaries of the two sets of figures are reported in Table 1,
with the slightly different scales having been standardized to afford a direct
comparison. Two measures are reported in the table. The first is the degree
of polarization, which measures the distance between the most left-located
and most right-located parties in each country in each survey, and which
harks back to the original concerns of Sartori (1976); the second is the degree
of core divergence, which measures the distance between the two principal
protagonists (simply, the two biggest parties) in each system, and which taps
into many of the contemporary debates concerning the apparently centripetal
tendencies which exist within established party competition.4 Both measures
are reported as a proportion of the distance involved in each scale as a whole,
with values ranging from a minimum of 0 (the two relevant parties occupy
precisely the same position) to a maximum of 1 (the two relevant parties are
at completely opposite ends of the spectrum), and the Table also reports the
difference which was recorded in these measures between the two survey
points.
Two general observations can immediately be made about these data. In
the first place, as is readily apparent from the differences in the distances
separating the leftmost and rightmost parties, the extent of polarization has
more often increased rather than decreased with time. Indeed, of the seven-
teen countries for which comparable data are available, there are eleven cases
which record an increased distance between the extremes, and only five (Bel-
gium, Denmark, the Netherlands, Spain, and the UK) in which this distance
has been reduced. An additional country, Austria, has remained unchanged
over time. Moreover, in three of the countries (Canada, Germany, and New
Zealand) the increase in polarization is particularly pronounced. Second, in
 155
a
Table 1. Polarization and core divergence in left-right distances, 1984 and 1995

Polarization Core divergence

(distance between the (distance between the two
leftmost and rightmost major protagonists
parties)
Country 1984 1995 Diff. Major parties 1984 1995 Diff.

Australia 0.43 0.55 +0.12 ALP-Lib 0.40 0.41 +0.01

Austria 0.57 0.57 0.00 SPÖ-ÖVP 0.25 0.15 0.10
Belgium 0.76 0.64b 0.12 PS/SP-CV/PC 0.31 0.17 0.14
Canada 0.42 0.61 + 0.19 Lib-Con 0.11 0.22 +0.11
Denmark 0.72 0.71 0.01 SD-KF 0.32 0.34 +0.02
Finland 0.49 0.65 + 0.16 SD-KOK 0.38 0.30 0.08
France 0.76 0.77c + 0.01 Soc-Gaul 0.51 0.38 0.13
Germany 0.59 0.78 + 0.19 SPD-CDU 0.31 0.26 0.05
Ireland 0.45 0.56 + 0.11 FF-FG 0.05 0.12 +0.07
Italy 0.78 0.83 + 0.05 PCI/PDS/DC 0.35 0.38 +0.03
Netherlands 0.78 0.77 0.01 PvdA-CDA 0.28 0.21 0.07
New Zealand 0.20 0.38 + 0.18 Lab-Nat 0.20 0.15 0.05
Norway 0.75 0.79 + 0.04 AP-H 0.43 0.39 0.04
Spain 0.85 0.65 0.20 PSOE-AP/PP 0.44 0.35 0.09
Sweden 0.59 0.65 + 0.06 SD-MSP 0.44 0.42 0.02
UKd 0.50 0.33 0.17 Lab-Con 0.50 0.33 0.17
USA 0.18 0.27 + 0.09 Dem-Rep 0.18 0.27 +0.09
a
Distances are measured by taking the distances between the parties concerned as a pro-
portion of the number of points on the scale (11 in 1984 and 10 in 1995), thus creating a
standardized scale running from a minimum of 0 (the two relevant parties occupy precisely
the same position) to a maximum of 1 (the two relevant parties are at completely oppo-
site ends of the spectrum). Polarization is measured by taking the distance between the
leftmost and rightmost parties, while Core divergence is measured by taking the distance
between the two principal party protagonists as listed in the table.
b
Because the extreme right Vlaams Blok was not listed in the 1995 survey, its score has
been estimated as the same as that recorded in 1984.
c
The French score for 1995 does not take account of Lutte Ouvrière or Lutte Communiste,
neither of which were included in the 1984 data.
d
Refers to Britain only, thus excluding Ulster Unionists and Irish Nationalists.
Sources: Castles & Mair (1984); Huber & Inglehart (1995).

a majority of the countries, and notwithstanding this generalized increase in

polarization, we also witness evidence of a greater convergence between the
two principal parties, with the distance in this case being reduced in all but
six cases (Australia, Canada, Denmark, Ireland, Italy, and the UK). In other
words, and remaining within these simple categoric distinctions, we see at
least a plurality of cases (Finland, France, Germany, New Zealand, Norway,
156
and Sweden) in which increased polarization in the party system as a whole
(a growth in the distance between the extreme parties) is accompanied by
increased convergence at the core (a decline in the distance between the two
principal protagonists).
Other things being equal, this might have suggested a possible relationship
between the drawing together of the centre of the system, on the one hand,
and the mobilization of more extreme oppositions at the edges, on the other,
a dynamic which has already been hypothesized by Ignazi (1992). In prac-
tice, however, the two trends do not always coincide. In a small handful of
countries, for example (Australia, Canada, Ireland, and Italy), polarization
has actually been accompanied by greater divergence at the core, and these
systems might be regarded as more generally centrifugal. In three other cases,
Belgium, the Netherlands and Spain, core convergence has been accompa-
nied by a more or less pronounced depolarization, and these systems might
be regarded as more generally centripetal. Indeed, more generally, and taking
account of the levels of changes as well as their direction in all the rele-
vant countries, we actually witness quite a strong positive correlation (0.512)
between changes in polarization and changes in divergence,5 which suggests
that there is in fact a similar dynamic that is impacting upon both the core of
the system and the extremes: the more the system stretches, as it were, the
more likely it is that the two core parties will draw further apart; the more
the system contracts, on the other hand, the more likely it is that the two core
parties will increasingly nestle up against one another. Or so the ‘experts’
would lead us to believe.

Notes

1. The same criticism might also be levelled at the survey by Huber & Inglehart (1995),
where many of the multitude of party scores are based on just a handful of respondents.
2. The major exception was the input of the Open University’s Social Sciences cartographer,
John Hunt, whom we failed to acknowledge at the time. Figure 1 of the original article,
which spatially summarized the position of more than a 100 parties in 17 countries,
involved a major commitment of time and skill in an era when there were effectively no
computers to assist such endeavours.
3. By now, in fact, neither of us can actually lay our hands on a copy of the original
questionnaire, and we would delighted to be sent a copy if any of the present readers
might still happen to have one to hand.
4. While there are sometimes differences between the two surveys regarding which partic-
ular parties happen to be located at the extremes, the two core parties are regarded as
being the same in both surveys.
5. It should also be noted that neither the UK nor the USA are considered here in that the
principal protagonists are also the leftmost and rightmost parties, and hence no differential
dynamic can be discerned.
 157
References
Barricelli, L.A. (1985). Considerations on a time-democratic electoral system, European Jour-
nal of Political Research 13(4): 379–386.
Cameron, D.R. (1982). On the limits of the public economy, Annals, AAPSSS, 459.
Castles, F.G., ed. (1982). The impact of parties. London: Sage.
Castles, F.G. (1994a). On religion and public policy: Does Catholicism make a difference?,
European Journal of Political Research 25(1): 19–40.
Castles, F.G. (1994b). On religion and public policy: The case for covariance. Reply to Göran
Therborn, European Journal of Political Research 26(1): 111–115.
Castles, F.G. & Mair, P. (1984). Left-Right political scales: Some ‘expert’ judgements, Euro-
pean Journal of Political Research 12(1): 73–88.
Castles, F.G. & McKinlay, R.D. (1979). Does politics matter? Public welfare commitment in
advanced democratic states, European Journal of Political Research 7(2): 169–186.
Hewitt, C. (1978). The effect of political democracy and social democracy on equality in
industrial societies: A cross-national comparison, American Sociological Review 42(3):
450–464.
Huber, I. & Inglehart, R. (1995). Expert interpretations of party space and party locations in
42 societies, Party Politics 1(1): 73–111.
Ignazi, P. (1992). The silent counter-revolution: Hypotheses on the emergence of extreme
right-wing parties in Europe, European Journal of Political Research 22(1): 3–34.
Inglehart, R. & Klingemann, H.D. (1976). Party identification, ideological preference and the
left-right dimension among Western mass publics. In I. Budge, I. Crewe & D. Fadie (eds.),
Party identification and beyond (pp. 243–273). New York: John Wiley.
Laver, M. & Hunt, B. (1992). Policy and party competition. New York. Routledge.
Laver, M. & Schofield, N. (1990). Multiparty government: The politics of coalition in Europe.
Oxford: Oxford University Press.
Lipset, S.M. (1960). Political man. London: Mercury Books.
Mair, P. (1996). Comparative politics: An overview. In R. Goodin & H.D. Klingemann (eds.),
A new handbook of political science (pp. 309–355). Oxford: Oxford University Press.
Morgan, M-J. (1976). The modelling of governmental coalition formations: A policy-based
approach with interval measurement. University of Michigan, PhD thesis.
Pedersen, M.N., Damgaard, E., & Nannestad Olsen, P. (1971). Party distances in the Danish
Folketing, 1945–1968, Scandinavian Political Studies 6/71: 87–106.
Sartori, G. (1976). Parties and party systems: A framework for analysis. Cambridge: Cambridge
University Press.
Sjöblom, G. (1977). The cumulation problem in political science: An essay on research strate-
gies, European Journal of Political Research 5(1): 1–32.
Swaan, A. de (1973). Coalition theories and cabinet formation. Amsterdam: Elsevier.
Weber, R.P. (1982). The long-term dynamics of societal problem-solving: A content-analysis
of British speeches from the Throne, 1689–1972, European Journal of Political Research
10(4): 387–405.