In 1972 the authors of a Swedish textbook on Nordic politics accused Nordic political scientists of neglecting the unique possibilities to make comparisons among the Nordic national political systems. The Nordic countries were, according to the authors, rewarding objects for comparisons: on the one hand they were similar enough to encourage comparisons from a methodological of view, on the other hand they were dissimilar to an extent that guaranteed variation and thus provided the basis for meaningful comparisons (Lindblad et al. 1972, 8). Yet comparative works that analysed certain political aspects in two or more Nordic countries were, the authors stated, "extremely rare", and the authors did not know of any major work in Nordic political science that dealt with an essential aspect of politics in all four Nordic countries on a comparative basis (Lindblad et al. 1972, 8).

This description of the state of affairs in the early 1970s was fairly
accurate. The authors did, however, overlook an excellent pioneering work
in the field of comparative Nordic politics, namely an early study by Goran

Resumé

In the beginning of the 1970s comparative: politics was still a neglected genre in Nordic political science. During that decade, however, comparative research started to make headway. An examination of articles published in five leading Nordic political science journals during the 1980s indicates that more than one-tenth of the content was devoted to comparative research. The examination suggests that the Nordic approach to the small-N problem in comparative research is variegated: binary comparisons constitute just over one-quarter of the volume, whereas the share of studies exploiting three to five cases is slightly larger, and the share of studies discussing more than five cases is slightly larger still. However, in the selection of cases for study, a more clear-cut strategy and pattern is revealed. It is clear from the journal data that Nordic comparativists escape ethnocentrism by turning to another ethnocentrism - i.e. they are predominantly preoccupied with the Nordic countries. While this emphasis on Nordic comparisons certainly can be defended on methodological grounds, doubts can be raised about the overall methodological justification of the Nordic orientation. The risks involved are not always recognized. What is known as Galton's problem merits special attention in a Nordic context. The nature of the problem is that empirical relationships may be the result of learning, which means that the cases studied lack independence. Since it is evident that a lot of diffusion occurs in the Nordic countries, the challenges posed by Galton's problem should receive much more attention in Nordic research than is presently the case.

von Bonsdorff (1954) regarding liberal parties in the Nordic countries. The textbook description was also to some extent misleading, as it neglected the thrust towards comparative political analysis that was characteristic of Norwegian social science at that time, a thrust that was primarily inspired by Stein Rokkan, who had acquired a comparative orientation during his participation in several international, comparative and multidisciplinary research projects (Kuhnle 1986, 52-55). In fact, the comparative aspect of political analysis was institutionalized in Norway during the 19605, as Rokkan was appointed to the chair of sociology at the University of Bergen in 1966, his foremost task and field of responsibility being the promotion of comparative politics (Kuhnle 1986, 50). By and large, however, the textbook authors were right in stressing the lack of rigorous comparative research in Nordic political science, and in calling for large scale efforts in this field.

When a third and updated edition of the same textbook appeared in 1984, the evaluation of the field of comparative studies was somewhat moderated, although the authors still argued that comparative studies of Nordic politics were rare (Lindblad et al. 1984, 8). Some moderation was certainly justified. Much more work than before was done in the field of comparative politics in the 1970s and the 1980s. For instance, several important books by Nordic scholars dealing with Nordic politics on a comparative basis were already published during the second half of the 1970s (Anckar 1989, 2-5).

It is the aim of this article to elaborate these observations and to explore further the present state of the art. However, a first and preliminary task must be to establish within the context of this paper the distinctiveness of comparative politics as a field of study. Giovanni Sartori has stated as a fact "that a field called comparative politics is densely populated by noncomparativists, scholars who have no interest, no notion, no training in comparing" (1991, 243). Although this article does not try to specify the extent to which this verdict is justified in the Nordic context, it is still necessary to draw here a line between comparativists and non-comparativists: must know, before we can proceed, how we shall approach and understand the field of comparative politics. The next section of the article is therefore devoted to some clarifications.

Some Clarifications

In his presidential address to the American Political Science Association in 1966, Gabriel Almond declared that "it makes no sense to speak of a comparative politics in political science, since if it is a science, it goes without saying that it is comparative in its approach" (1966, 878). "Comparison",

Almond argued, "whether it: be in the experiment, in the analysis of the results of quantitative surveys, or in the observation of process and behavior in different contexts in the real world, is the very essence of the scientific method" (1966, 878). In a similar manner, Stein Kuhnle and Stein Rokkan, in a presentation of Norwegian political science, have emphasized that all political research proceeds by comparison: "Whatever the method, any systematic treatment of politics; must resort to comparative analysis" (1978, 152). There is much to be said in favour of this view, as it can hardly be denied that the scientific method itself assumes comparison. For instance, the very basic scientific task of defining concepts presupposes that comparisonsare out: the thing which is defined must be compared to things from which it differs (Merkl 1970, 4). However, the view that comparative politics is just another term for political science is not very instructive or useful. Furthermore, this view is not in agreement with normal usage and normal practice::, which frequently singles out comparative politics as a special subfield of political science. In a recent presentation of the European Consortium for Political Research, for instance, Kenneth Newton notes that the emphasis; in this organization from the very start "has been not just on political science or on teaching and research, but on comparative approaches to the subject" (1991, 449).

What, then, is "comparative politics"? By what criteria can this subfield be singled out as a subfield, what are the defining characteristics of "comparative Again, this question can be answered in differing ways. Arend Lijphart notes that a variety of meanings is attached to the terms "comparison" and "comparative method", and prefers to define the comparative as one of the basic methods of establishing general empirical to him the comparative method is "a method of discovering empirical relationships between variables" (1971, 683). In a later contribution expands this view. Rather circumstantially, he defines the comparative method as "the method of testing hypothesized empirical relationships among variables on the basis of the same logic that guides the statistical method, but in which the cases are selected in such a way as to maximize the variance of the independent variables and to minimize the variance of the control variables" (Lijphart 1975, 164). These definitions, then, do not equate the comparative method with comparative politics. Concerning the relationship of comparative politics as a substantive field and comparison as a method, Lijphart emphasizes that the two are clearly not coterminous: "In comparative politics, other methods can often also be employed, and the comparative method is also applicable in other fields and disciplines" (1971, 690). This article accepts this view: comparative method is one thing and comparative politics is another. The comparative method, therefore, cannot be used to define comparative politics.

It seems only natural to state that comparative politics is not necessarily

about cross-national or region-specific research and foreign political

systems. The comparativist compares, and the things he chooses to compare depend on the research problem. He may dwell on intra-nation divisions as well as inter-nation divisions. In fact, it is not an unusual argument that comparability can be enhanced by focusing on intra-nation instead of internation intra-unit comparisons offer the advantage that interunit can be held constant (Lijphart 1971, 689; 1975, 167-168). The everyday approach to the task of delimiting comparative political research, however, is less sophisticated. The ordinary view is quite simple: comparative politics is about cross-national issues. Sartori has noted that the comparative politics field in the United States defines itself as studying "other countries" (1991, 243), and Peter Merkl has observed that in customary within political science "comparative politics" refers chiefly to the study of foreign political systems (1970, 4). Richard Rose has made a similar observation that the term "comparative politics" refers in everyday usage to the study of foreign countries and that familiar political science usage excludes within-nation comparisons from the field of comparative politics (1991, 446-447). Of course, this preoccupation with "other" and "foreign" countries does not necessarily exclude the comparativist's "own" country from consideration; the minimal requirement, however, is that this "own" country is contrasted with one or several "other" or "foreign" countries.

This article departs from this rather uncomplicated and admittedly rather artificial view. It deals with comparative political research in the Nordic countries, and understands this term to denote research by Nordic scholars that aims at the comparison of one or several countries with one or several other countries. This operational definition, which does not pay regard to works that contain only occasional and scattered references to conditions in "other" countries, comes close to the one used by Edward Page, who, in his treatment of comparative politics in British political science, takes as his subject "works of scholars which present empirical evidence of some kind in an attempt to compare systematically and explicitly political phenomena in more than one country" (1990, 439). However, whereas Page excludes from his discussion works focusing on Eastern Europe and the newly developed and the less developed nations (1990, 439), this article does not discriminate against any areas, regions or nations. Also, whereas Page excludes pure theories without empirical evidence (1990, 439), this article acknowledges theoretical and methodological pieces of research, as long as it is clear that their main concern is with comparative politics as defined here.

This article does not subscribe to the "whole-nation bias" (Rokkan 1970,
49): it deals with comparisons of subnational units in different nations as
well as with comparisons of nations. Comparing local party organizations

in, say, Sweden and Denmark, or comparing output performances of communes in, say, Finland and Norway, are activities that are clearly within the conception of comparative politics that will be used here. Concerning the debate about whether or not case studies should be included in the realm of comparative politics, this article takes the view that case studies must be excluded. Of course this view is by no means uncontroversial. In their valuable book on strategies in comparative politics, Mattei Dogan and Dominique Pelassy note that it may look paradoxical to include the case study among the strategies open to the comparativist (1984, 107). However, they add, many comparativists make use of this strategy (1984, 107). It is also obvious that sorae case studies are offshoots of the comparativemethod, others are not. On the one hånd, studies limited to one country often emphasize the uniqueness of the situation and are difficult to integrate into the comparative discipline (Dogan & Pelassy 1984, 107). On the other hånd, the study of individual cases, especially studies of so-called deviant cases (cf. Lijphart 1971, 692-693; Hakovirta 1976, 17-19),may at generalization aod theoretical reformulation. The point to be made here is that the inclusion of case studies in the realm of this article, which makes use of a set of empirical data, would require the classification of a vast amount of studies as being either comparative or non-comparative in terms of design and ambition. Since this task is insurmountable, case studies are excluded.

How Much Comparative Research?

It is of course impossible to state precisely how much comparative research is needed for the genre to be sufficiently represented in a totality of political research. It all really depends on the criteria and yardsticks employed. While some would argue that a certain volume suggests that the genre is clearly underrepresented, others would argue that the same volume stands for an overrating of the status of this field. Assessments of comparative research must therefore build on comparisons: entities like countries or time periods can be compared with each other in terms of the extent to which they have promoted comparative research. This strategy will be used here.

The strategy is implemented through an examination of five leading Nordic journals in the field of political science. Four of these journals represent national political science communities, namely Politica (Denmark), Politiikka (Finland), Norsk Statsvitenskapelig Tidsskrift (Norway)and Tidskrift (Sweden). The fifth journal, ScandinavianPolitical is a joint journal of the national political science associations in all five Nordic countries, and is published by the Nordic

Political Science Association. The examination does not cover all materials that have appeared in the journals. Only main articles and research reports are included. Other materials, such as editorial notes and comments, book reviews, reports of panel discussions, in memoriams, etc., are omitted. The time span of the examination covers the years 1980-89, and all main articles and research notes that have appeared in the journals during this period have been classified as being comparative or not, the defining criteria being those explained in the previous section. The classification unit of this content analysis is pages, whereas the context unit is separate articles. Norsk Statsvitenskapelig Tidsskrift has only been published since 1985, and the contents of this journal have left only marginal imprints on the quantitative basis for the observations and interpretations that are given in this article.

The total number of pages exceeds 13 000. Of these, fully 1600 are coded as representing comparative politics. The exact percentage is 12. The figure is in fact rather high, at least when compared with proportions reported by Page, who has observed that articles in the field of comparative politics represent around five percent of all articles published in Political Studies and British Journal of Political Science (1990, 445). Richard Rose has argued that the scholarship of professionals in a big country tends to be confined by the richness of materials there, whereas scholars in smaller countries tend to be more aware that other countries exist and more inclined to see their countries as some among many in the field of comparative politics (1990, 583). The overall validity of this generalization may be questioned, but the argument appears to be supported by the findings reported here. Even so, a research volume which barely exceeds 10 percent is hardly impressive.

The relationship between the content of the journals on the one hand and the emphasis on comparative politics in the Nordic political science communities on the other hand is of course open to interpretation. The fact that a certain journal devotes a certain number of pages to comparative research does not imply a precise and one-to-one representation of the emphasis at large in this field. A great part of the totality of comparative research is indeed published elsewhere than in the journals in question ~ in books and monographs, in other journals, in research reports, etc. All the same, the journals are all general political science journals which do not discriminate against certain fields of research or certain orientations. They are all likely to treat contributions in the field of comparative research with the same generosity or the same deliberation as they treat contributions in other fields of political science. They reflect currents within the political science cultures from which they have emerged; they can therefore be expected to reflect with some accuracy the interest that various fields of political research arouse in the Nordic countries. With these general

comments in mind, we may consider Table 1, which reports the volume of
comparative research in the journals during the 1980s as well as three subperiodswithin
decade.

At least three observations merit attention:

(1) Comparative research does not seem to have advanced during the latest decade. In fact, three journals out of four devoted more space to comparative politics during the first subperiod than during the last. On the other hand, neither is comparative research declining. There is, in short, no clear trend in the data. It has been stated that comparative politics as a field of study has in general attracted growing interest in terms of publications, and study programmes (Lane & Ersson 1990, 61). This observation is probably true with regard to the discipline as a whole. However, the validity of the statement might be questioned in a Nordic context. A lot of comparative research has been and is going on in the Nordic countries, but the volume of research publication does not appear to be increasing. The data in Table 1 rather suggest that a saturation point may have been reached.

(2) Concerning differences between countries, Finland is relatively strong in comparative research whereas Sweden is weak. The Finnish journal has contained a higher percentage of comparative research than the journals of the other countries, and over the years Finnish scholars have produced a respectable amount of comparative research, a great part of which is unknown for language reasons even to the other Nordic political science communities (Anckar 1985). It is not a difficult task to point at individual and impressive comparative works by Swedish scholars, but the overall picture is clearly more dismal than in the Finnish case. A review of political research in Sweden during the first half of the 1960s (Molin 1966) did not even mention the comparative genre, and when, some years later, Olof

Ruin published a bibliographical survey of political science research in Sweden, he regretted that attempts to compare political conditions in several countries were to a great extent lacking in Swedish research (Ruin 1969, 174). About one decade later, in another survey, Ruin noted that a reversal seemed to be in progress in this respect (1978, 179); this change, however, has not materialized to any noticeable extent. Recently another prominent Swedish political scientist has in faet characterized comparative research in his country as still being neglected, meagre and non-cumulative (Hadenius 1989).

The reason for this state of affairs is not clear. One important factor is probably the emergence in the 1960s and the 1970s of sector research, i.e. research which is motivated by the needs of special social sectors such as, for instance, housing and labour, and which is financed by the sector in question. The spread and impact of sector research is not equally characteristic all Nordic countries. There is perhaps a movement in this direction everywhere, but the Swedish scientific community has become dependent on sector research and sector funding to an extent which is clearly higher than that in the other countries. An evaluation of Swedish political science points at a relevant outcome: research on politics in Sweden is formed by political and administrative priorities to a higher extent than in many other countries (Eliassen & Pedersen 1984, 85). Such priorities are likely to work against comparative research. True, sectoral funding and similar arrangements need not in themselves always form obstacles to comparative research. In some instances they may even involve promises for the initiation and funding of such research (Anckar 1991, 248-249). As a rule, however, they probably foster tendencies towards maintaining and strengthening the ethnocentrism of political science. They stem from the perceived need to bring political science closer to society and societal problems, and they therefore probably imply, in practice, an estrangement from topics that go beyond a near-sighted occupation with problems of the specific country in question.

(3) Scandinavian Political Studies is clearly superior to the individual country journals. During the 1980s the journal has devoted more than onefifthof space to comparative research. During the last subperiod the share of comparative politics in fact exceeds one-quarter. This is by any reasonable standard a quite impressive achievement for a general political science journal. One possible explanation is that the journal, because of its affiliation with the Nordic Political Science Association, has a special obligation to look after and improve the publishing of comparative politics. It may be the case that to some extent the special profile of the journal serves to invite available comparative contributions by Nordic scholars and thereby makes inroads in the offering of comparative pieces of research to

the other journals. There is, however, no reason to believe that this factor
to any noticeable degree distorts the proportions reported by the individual
country journals.

The Small-N Problem

The principal problem facing comparative research has been stated by Lijphart in a much quoted passage as being a condition of many variables, small number of cases (1971, 685). In the same context Lijphart also discusses means of minimizing these problems, and he puts forward four suggestions: (1) increase the number of cases as much as possible; (2) reduce the property-space of the analysis; (3) focus the comparative analysis on comparable cases; and (4) focus the comparative analysis on the key variables (1971,686-690). All these strategies will not be explicitly discussed here, but the first and the third merit special attention. This is due to Lijphart's own observation (1971, 687; 1975, 163) that these recommendations in fact incompatible. The incompatibility stems from the fact that a search for comparable cases, i.e. cases where many variables are controlled, usually will cause the number of cases subject to analysis to decrease. This observation is highly relevant for the purpose of this article, as the Nordic countries are, on the one hand, few in number, and, on the other hand., similar to an extent which offers the possibility of establishing controls and thus makes them comparable cases. To the extent that Nordic comparativists focus on comparisons of Nordic countries, they then make a choice in favour of the third recommendation rather than the first.

In Table 2 some basic information is provided on the way the small-N
problem has been approached and handled in the comparative articles

included in the four Nordic journals. The calculations operate with three categories: (a) binary comparisons (Dogan & Pelassy 1984,112-116), where N = 2; (b) comparisons that encompass three to five cases (countries); and (c) other comparative studies, based on a number of cases exceeding five.

Two conclusions seem warranted on the basis of the materials in this table. The first is that Nordic political science makes use of various methods of dealing with the small-N problem: there is not one dominating school. The share of binary comparisons is just over one-quarter (26 percent), whereas the share of studies exploiting three to five cases is slightly larger (32 percent). Studies that discuss more than five cases attain the largest share (42 percent). Thus, the Nordic approach to the small-N problem is variegated and thereby really not Nordic at all. In his study of British political science and comparative politics, Page reports that of 132 articles published in some leading European journals, 40 compared two countries, whereas 32 compared 3 to 4 countries, 25 compared 5 to 10 countries and 35 compared more than 10 countries (1990, 447). This profile is thus very similar to the Nordic profile.

The other conclusion is that Finland stands out as a deviant case, which puts a clear emphasis on comparisons of many cases. Two-thirds of the comparative materials in the Finnish journal belong in this category, and during two sub-periods out of three, the share in fact exceeds 70 percent. One perhaps extreme example in the Finnish research literature is a recent study of the process of democratization in 147 states (Vanhanen 1990).

Further discussion of these conclusions cannot be pursued on the basis of the data provided in Table 2 alone. Since the selection of countries for comparison may be based on very different criteria and assumptions (Teune 1990, 43-45), the mapping of the number of cases studied only is clearly insufficient. An estimate of the theoretical justifications for selecting countries a more refined approach, and we therefore turn to an analysis of the concrete country patterns that are embedded in the three categories.

In their influential work on The Logic of Comparative Social Inquiry (1970) Adam Przeworski and Henry Teune contrast two different conceptionsof comparative method. One, regarded by the authors as the "predominant view among social scientists" (1970, 32), is the most similar systems design. The implication of this design is the search for comparable cases: systems as similar as possible on attributes other than the one that requires explanation are selected for comparison. The strategy sets out to neutralize certain differences in order to permit a better analysis of others; the aim is to achieve a large measure of control (Dogan & Pelassy 1984, 118). The other view of the comparative method, introduced by Przeworski and Teune, is the most different systems design, which has been characterizedas "genuine innovation", which "represents one of the most

important proposals to be found in the literature on comparative analysis" (Meckstroth 1975, 136). This design aims at the elimination of irrelevant system factors and the formulation of statements that are valid, regardless of the systems within which observations are made (Przeworski & Teune 1970, 353). The search is for systems "that differ as much as possible and yet do not differ on the phenomenon under investigation" (Sartori 1991, 250); contrasting comparisons "try to find analogies in contrasting political systems" (Dogan & Pelassy 1984, 127).

Which research methodology, then, is preferred by Nordic comparativists? kinds of systems are selected for comparison in Nordic comparative research? Do Nordic scholars lean towards the most similar systems design or are they predominantly adherents of the most different systems design? Is there a pattern? The following presentation of country patterns in the journals at least hints at some answers. Four points can be made.

(1) It has been said that one can greatly improve insights in one country by immersing oneself in at least one other country (Daalder 1987, 16). To compare is to escape from ethnocentrism (Dogan & Pelassy 1984, 5-12). It is very clear from the journal data that Nordic comparativists escape ethnocentrism by turning to another ethnocentrism: they are predominantly preoccupied with the Nordic countries. To a high extent indeed, comparisons Nordic scholars are Nordic comparisons. This is evident from the information provided in Table 3. One-quarter of the volume devoted to binary comparisons involves comparisons between two Nordic countries, and half the volume is basedon comparisons that involve one Nordic country, the other country then being, in most cases, either the UK or the USA.

The Nordic predominance achieves still more impressive proportions in
studies that involve three to five countries. In this category an astonishing

share of more than 90 percent is devoted to the Nordic countries only. A characteristic pattern in this category is that if one Nordic country is dropped (N = 4), that country is Iceland; if two are dropped (N = 3), they are Finland and Iceland. While quite understandable in terms of data accessibility, this strategy appears unsound in terms of methodology. A search for differences in roughly similar contexts is hampered when and if Finland, which often stands out as a Nordic deviant case (see, for example, Berglund & Lindstrom 1978, 19-20), is left out. Finally, comparisons of more than five countries involve in more than half of the cases one or usually several Nordic countries. The population in these cases is as a rule a variant of a West European configuration, i.e. the Nordic ethnocentrism takes the form of a West European ethnocentrism. In short, comparative research in the Nordic countries is characterized by a lopsidedness, the implication of which is that research focuses on areas that are "domestic" rather than "foreign".

(2) The heavy emphasis on Nordic comparisons can of course be easily defended on methodological grounds. Since the Nordic countries are more or less similar in a great many respects, they stand out as good examples of comparable cases that fit, very neatly, central requirements of the most similar systems design. Therefore, the emphasis on Nordic countries, one could argue, represents an emphasis on this design: there is a pattern. It is not the intention here to refute the overall validity of this approach.

It is, however, difficult to escape the impression that the methodological justification in many cases plays a minor role, and is, in fact, not adduced as a motive for the selection of Nordic countries. Quite often, and especially in cases when 3-5 countries are compared, the selection is not defended or justified at all. The selection is rather an outcome of an unspecified ambition to compare all or most Nordic countries, to cover in one crush the Nordic totality. In many instances the selection of Nordic countries for study may quite simply be a matter of availability. If so, the research site is not selected in order to control and manipulate specified variables (cf. Holt & Turner 1970, 12-13), and the preference for the most similar systems design is then really no preference at all, but rather a sort of unreflecting inclination. The question of what comparison is for (Sartori 1991, 244), is not addressed in these cases.

(3) When studying more than five cases and thus widening the geographical scope outside the Nordic countries, Nordic comparativists more often than not stop at the boundaries of Western Europe. Of course, this favouring of the European context to the disadvantage of a global context or other regional contexts builds upon a tradition, as it links up with Stein Rokkan's preoccupation with Europe (Allardt 1981; Daalder 1987, 13-17). The

predominant concern is also with many West European countries rather
than with few: the vast majority of the studies in this category deal with 10
to 12 countries or more.

Although the "large nation bias" in comparative European politics (Urwin & Eliassen 1975, 87-88) can still be found in Nordic textbooks on comparative politics (e.g. Helander 1984), it must be acknowledged that the general research literature (e.g. Lane & Ersson 1987) as well as the journal articles display a preoccupation with all, rather than a few, European Again, however, the selection of countries for analysis is merely presented or declared more often than explained and justified. It must appear that the choice expresses a preference for analogies that are present in a rather well-defined region: "in spite of their diversity, the countries of Western Europe display substantial similarities" (Dogan & Pelassy 1984, 121). One key to the selection probably lies in the word "democracies". By focusing on political homogeneity, a relative similarity is incorporated in the analysis.

A few additional words need to be said here about the Finnish case. As mentioned previously, the Finnish journal appears strong in the morethan-five-cases Finnish comparativists, in other words, are more inclined than their Scandinavian colleagues to transgress the Nordic horizon. An interpretation that stresses scientific implications and, for instance, identifies a preference for a most different systems design, would be premature, however. A stronger trend towards holonational analysis is not to be found in the Finnish comparative literature any more than in the other Nordic countries (Sigelrnan & Gadbois 1983,282). The Finnish profile in this category is, even according to the material investigated here, similar to the general Scandinavian profile. The inclination in Finland to avoid purely Nordic comparisons probably reflects lingering feelings of a linguistically isolation and estrangement from Nordic solidarity among Finnish political scientists. It is perhaps symptomatic in this respect that representatives of the Swedish-speaking population in Finland have played a conspicuously active role when it comes to Finnish contributions to the comparative study of Nordic politics (e.g. Helenius 1975; Isaksson 1989; Karvonen 1981).

(4) Some decades ago a critic of traditional comparative politics alleged that many scholars regarded the study of politics outside the nations of the West as a waste of time, since non-Western political patterns were neither natural nor durable (Macridis 1955, 10-11). Although the motives for disregarding non-Western systems at present are other than this, it appears a fact that comparative research in the Nordic countries is still relatively narrow and restricted in terms of the geographical scope of the field. The Nordic literature is by no means void of studies of non-European countries

and regions, but these studies are seldom designed in terms of systematic
comparisons. Comparative research in the Nordic countries, in other words,
is certainly not cosmopolitan in scope.

The narrowness of scope may in fact be even more pronounced than in many other comparative politics communities. A content analysis of articles published in Comparative Politics and Comparative Political Studies for the years 1968-81 reveals that 18 percent of the articles dealt with comparisons of more than five countries (Sigelman & Gadbois 1983, 282-283). The percentage is exactly the same as in this review. In the light of these figures, the fit between Nordic comparative research and comparative research elsewhere is perfect. However, since articles dealing only with one country are included in the content analysis by Sigelman and Gadbois, but excluded from this review, the figures are not comparable. An elimination of the onecountry indicates that almost half (46 percent) of the contributions to Comparative Politics and Comparative Political Studies are about six or more nations, whereas the volume of corresponding contributions to the Nordic journals, is, as indicated, less than 20 percent.

In conclusion, more than 80 percent of the comparative materials in the Nordic journals deal with the Nordic countries, either as such or as parts of the West European community. This clearly indicates a search - be it conscious, half-conscious or unconscious - for similarities in context. The emphasis is on most similar systems, not most different systems. The emphasis may not always, or not even as a rule, be the result of methodological but it is there. The following concluding section of this article will therefore argue for the view that there is ample room in Nordic comparative research for an expansion of the geographical field of analysis and for an inclusion of cross-cultural topics and problems.

Pitfalls of Ethnocentrism

The concentration in Nordic comparative research on the Nordic countries has several disadvantages and shortcomings in its wake. For one thing, the concentration shuts out an abundance of events and phenomena that are of immediate interest to political science and political scientists. By the end of the 1980s the nation states of the world counted 165 (Derbyshire & Derbyshire 1989, 6-8). It goes without saying that a research policy that advocates an orientation towards only four or five of these units represents a narrow view of the challenges political science has to confront. It is, for instance, a rather restricted range of questions concerning democracy that can be answered through an examination of the northern and West European manifestations of democracy. It has been predicted that the focus in comparative research on conditions for democracy and democratic

performance will give way to a broader perspective on various types of regimes (Lane & Ersson 1990, 75). To the extent that Nordic researchers persist in their predilection for their own environments, they will be unable to contribute to this broadening process.

One possible objection to this kind of reasoning is that a division of labour in research is important and desirable, and that it is only natural that Nordic researchers take the responsibility for the study of their own countries, which otherwise, perhaps, would not receive enough scholarly attention. Several counter-arguments can, however, be given. One transfers an argument in favour of comparative research into an argument against Nordic ethnocentrism: in the same way as one can improve insights in one country by immersing oneself in other countries, one can improve insights in Nordic politics by immersing oneself in other regions and geographies. Also, there is no intrinsic scientific value in the study of the Nordic countries. Of course, political authorities, civil servants and perhaps even the public are interested in and well served by information about the way these societies function, but this should be of little concern to comparativists, should, instead, regard the countries as a few cases out of many, which can be selected for study or not, the decision being dependent on the extent to which a selection of Nordic cases can be expected to promote scientific goals and aspirations.

Finally, regarding the principle of a division of labour, two things need to be said. One is that political science does not thrive in totalitarian and authoritarian regimes (e.g. Berndtson 1992, 50-51), which, if the principle is applied, will therefore largely remain outside the reach of the discipline. Nordic comparativists should rather counteract than submit to this state of affairs. The other thing that needs to be said is that a division of labour is in poor agreement with any ambition to steer the efforts of the discipline in a nomothetic direction. The lack of cumulativeness in political science is to a great extent a consequence of the fact that political scientists are not convinced about the need to extend generalized knowledge (e.g. Sjoblom 1977, 5-6). In efforts to educate political scientists in this respect, ethnocentrism certainly of no great help.

A limited orientation may also serve to make way for research strategies which are less fruitful. The study of similar countries may lead to a decline of theoretical ambitions: the farther apart and the more sharply contrasting the compared countries are, the greater the need to rise on the scale of abstraction (Dogan & Pelassy 1984, 27). Of course, many would argue that political science is already overpopulated with highly abstract inventions and models which are void of content and therefore cannot guide empirical research. This may be true, but the absence in the political science journals studied here of methodological arguments in favour of the choice of Nordic countries suggests that the preoccupation with the familiar Nordic societies

may subdue even the fashioning of less pretentious generalizations. Furthermore,an sloppiness in the selection of cases for comparison may overlook the risk that the tempting similarities are, in faet, too farreaching,as exist not only with regard to control variables, but also with regard to operative variables (Lijphart 1975, 163-164). The risk is also obvious that easy access to similarity dictates the choice of research problems. If comparative research on the Nordic countries is conducted because the Nordic countries are easy to compare, then research strategies are implemented in reverse order: method dictates problems whereas problems should dictate rnethod.

In addition, Nordic cornparativists need to be aware of and capable of handling a specific problem, which poses an intricate and difficult challenge namely, that the logic of the comparative method and, indeed, of scientific research, builds upon a belief that the greater the ability of a proposition to survive repeated tests, the greater the truth that is assigned to the proposition. However, this will hold only if the cases that are tested are independent. If they are not, no additional confirmation of the proposition is obtained by studying several cases. Existing support for a proposition does not become more convincing if it is counted twice or several times. Interdependent cases should not be assumed to be independent, causal relationships should not be confused with historical diffusion: correlations may merely reflect incidences of borrowing or migration. For instance, if one finds a relationship between two variables X and Y, this relationship may only reflect historical diffusion from Z (Moul 1974, 148-149; Karvonen 1981, 26-27; Karvonen 1991, 78-79).

This is what is referred to as Gallon's problem, which has been saiid to be "one of the most important problems of scientific method which face social scientists" (Naroll 1965, 434). The problem was raised by its originator the end of the last century in reaction to the introduction of the cross-cultural survey method (Naroll 1965, 428-429; Scheuch 1990, 28). Although important, the problem is largely neglected among comparativists in political science, including, certainly, Nordic comparativists. There is a striking difference in this respect between political science and some other disciplines. With anthropologists, it has been said, the difficulty has often been the reverse: they have had to be reminded that some traits might to some extent have some functional linkage (Naroll 1965, 434).

The lack of attention among political scientists to Galton's problem is rather astounding. Ignoring the problem implies adherence to an assumptionof closed system (Moul 1974, 149), and it is commonplace that social and political systems are open systems. True, the degree of openness may vary: there are many social, economic and political barriers to the spreading of ideas, customs and goods (Moul 1974, 150). This, however, does not contradict the fact that openness is the rule, that social systems interact

with each other and with their environments, and that they by so doing are
influenced by as well as influence each other.

One obvious solution to Gallon's problem is therefore to choose cases for comparison which can be assumed to have influenced each other to a minor extent only. For instance, one may try to find cases that are at a sufficient distance from each other in terms of time or space (Manheim & Rich 1991, 221). Other factors should also be considered. The literature that attempts to explain diffusion emphasizes, on the whole, two factors (Karvonen 1981, 15—16). It is concerned, on the one hand, with contact patterns and networks and with communication. And it underlines, on the other hand, the importance of structural similarities for the penetration and application of innovations. One leading conclusion is that diffusion is promoted by similarities between societies: if two societies are similar, they are likely to influence each other through diffusion. The implication of this is that societies that are similar are risky objects for comparison: comparativists are attracted by similarity, but the similarity may hide pitfalls. Areas that are regions are especially suspect in this respect. It is indeed a plausible assumption that the social and cultural homogeneity that is characteristic of regions is a product of diffusion rather than a product of independent development (Moul 1974, 148).

The Nordic countries are close to each other, they are socially and culturally homogeneous, they form a region. The contacts between the countries are frequent, systematic and integrated; there are obvious structural between the countries. The similarities suggest that the decision-makers in the individual countries have to cope with similar problems and circumstances and act in rather similar contexts as they try to seek new solutions and views. They can therefore be expected to turn to the neighbouring countries for models and inspiration (Anckar 1991, 249). This, in fact, happens frequently. It has, for instance, been demonstrated empirical research that Finnish politics and legislation has a propensity to reflect impulses from Sweden, and that elements from Swedish have been consciously introduced in Finland (Karvonen 1981). There is a great deal of diffusion between the countries, so much so that if, in some case, diffusion cannot be found, this arouses curiosity and calls for research into the mechanisms that prevent diffusion (Isaksson 1989). Comparative research on the Nordic countries is therefore especially exposed to the threats; against the empirical validity of findings that are lodged by Galton's problem. Of course, the problem does not aim a deathblow against attempts to study the Nordic countries on a comparative basis: whereas the relevance of the problem is high in some types of approaches and research tasks, it is less so in others. The problem does, however, imply that comparativists need to approach the Nordic area with caution and in full consciousness of the risks that are involved in comparative in that area.

One conceivable alternative is to avoid the Nordic countries and seek objects for comparison from other geographic areas. There are, however, quite a few impediments to a reorientation among Nordic comparativists. Some impediments are purely technical and practical and therefore easy to overcome. Others are psychological, Nordic scholars being, like scholars are everywhere, products and members of their own societies and cultures, to which they are tied with strings that emerge from tradition, affinity and emotion. Still others are institutional, as there are special and influential units for the promotion of comparative research, such as the Nordic Political Science Association and the Joint Committee of the Nordic Social Science Research Councils, which cherish the goal to advance comparative research on as well as in the Nordic countries. Furthermore, those who pay for and make use of research are inclined to support and demand undertakings that deal with their own region and its internal problems.

All in all, therefore, the ethnocentric feature of Nordic political science is not likely to disappear. And, indeed, it is not to be desired that it disappear entirely. It is, however, desirable that the Nordic orientation is implemented in full consciousness of the dangers associated with the orientation, and it is also desirable that the orientation is supplemented by an increased interest in the comparative study of more remote and unfamiliar systems.

---

NOTES

1. This is a revised and abridged version of a paper prepared for a conference "Political Science in Scandinavia: Trends and Challenges", arranged by Scandinavian Political Studies in Oslo, 22-24 May, 1992. An abridged version in the Swedish language was delivered at the Annual Meeting of the Norwegian Political Science Association, Bergen, 28-29 October, L992. The paper in its finalized form has also been delivered at the 25th Annual Meeting of the Finnish Political Science Association, Helsingfors and Stockholm, 13-15 January, 1993.

REFERENCES

Allardt, E. 1981. "Stein Rokkari om politikens makroforeteelser", Statsvetenskaplig Tidskrift
84, 187-192.

Almond, G. 1966. "Political Theory and Political Science", The American Political Science
Review 60, 869-879.

Anckar, D. 1985. "Political Facts on Finland: An Overview", European Political Data
Newsletter No. 54, 5-15.

Anckar, D. 1989. "Jamforande nordisk forskning om politik och kommuner: inledancle iakttagelser", in Oulasvirta, L., ed., Kommunalforvaltningen och kommunalvalen i Norden. Tammerfors: Kommunalvetenskapliga Institutionen, Tammerfors Universitet. Publikationsserie

Anckar, D. 1991. "Nordic Political Science: Trends, Roles, Approaches", European Journal
of Political Research 20, 239-261.

Berglund, S. & Lindstrom, U. 1978. The Scandinavian Party System(s). Lund: Studentlitteratur.

Eterndtson, E. 1992. Politiikka tieteenå. Helsinki: Valtionhallinnon kehittåmiskeskus.

Bonsdorff, G. v. 1954. Studier rdrande den moderna liberalismen i de nordiska lånderna.
Ekenås: Ekenås Tryckeri Aktiebolag. Skrifter utgivna av Fahlbeckska stiftelsen, 40.

Daalder, H. 1987. "Countries in Comparative European Politics", European Journal of
Political Research 15, 3-21.

Derbyshire, J. D. & Derbyshire, 1.1989. Political Systems of the World. Edinburgh: Chambers.

Dogan, M. & Pelassy, D 1984. How to Compare Nations. Strategies in Comparative Politics.
Chatham, New Jersey: Chatham House Publishers.

Eliassen, K. & Pedersen, M. N. 1984. Svensk samhållsorganisation och forvaltning: En
inventering av svensk forskning. Uppsala: Humanistisk-samhallsvetenskapliga forskningsrådet.

Hadenius, A. 1989. "Forsuinmade omraden inom jåmforande politik", Politologen 18, 93-99.

Hakovirta, H. 1976. Puolueettomuus ja integraatiopolitiikka. Tampere: Acta Universitatis
Tamperensis, Ser. A Vol. 78.

Helander, V. 1984. Låntisten suurvahain poliittiset jårjestelmåt. Tampere: Finnpublishers.

Helenius, R. 1975. Akademikernas fackliga organisationsberedskap. Helsingfors: Skrifter
utgivna av Svenska Handelshogskolan, Ekonomi och Samhålle, Nr 24.

Holt, R. T & Turner, J. E. 1970. '4The Methodology of Comparative Research", in Holt, R.
T. & Turner, J. E., eds.. The Methodology of Comparative Research. New York: Free
Press.

Isaksson, G-E. 1989. Resultatids omlokalisering. Åbo: Åbo Akademis Forlag.

Karvonen, L. 1981. "Med vart våstra grannland som forebild". Åbo: Meddelanden från
Stiftelsens for Åbo Akademi forskningsinstitut, Nr. 62.

Karvonen, L. 1991. "'A Nation of Workers and Peasants': Ideology and Compromise in the
Interwar Years", in Karvonen, L. & Sundberg, J., eds., Social Democracy in Transition.
Aldershot: Dartmouth

Kuhnle, S. 1986. "Linjer i norsk statsvitenskap: Institusjonell differensiering og forskningsekspansjon",
Kuhnle, S., ed., Det politiske samfunn. Otta: Tano A.S.

Kuhnle, S. & Rokkan, S. 1978. "Political Research in Norway 1960-1975: An Overview", in
Scandinavian Political Studies, Volume 12/77. Oslo: Universitetsforlaget.

Lane, J-E. & Ersson, S. 1987. Politics and Society in Western Europe. London: Sage Publications.

Lane, J-E. & Ersson, S. 1990. "Comparative Politics: From Political Sociology to Comparative
Public Policy", in Lefiwich, A. ed. New Developments in Political Science. Aldershot:
Edward Elgar.

Lijphart, A. 1971. "Comparative Politics and the Comparative Method", American Political
Science Review 65, 682-693.

Lijphart, A. 1975. "The Comparable-Cases Strategy in Comparative Research", Comparative
Political Studies 8, 158-177.

Lindblad, 1., Wahlbåck, K. & Wiklund, C. 1972. Politik i Norden. En jåmforande oversikt.
Stockholm: Bokforlaget Aldus/Bonniers.

Lindblad, 1., Wahlbåck, K , Wiklund, C. & Stålvant, C-E. 1984. Politik i Norden. En
jåmforande oversikt. Stockholm: Liber Forlag.

Macridis, R. 1955. The Study of Comparative Government. New York: Random House.

Manheim, J. B. & Rich, R. C. 1991. Empirical Political Analysis. 3rd edn. New York:
Longman.

Meckstroth, T. W. 1975. " 'Most Different Systems' and 'Most Similar Systems'", Comparative
Political Studies 8, 132-157.

Merkl, P. 1970. Modern Comparative Politics. New York: Holt, Rinehart and Winston, Inc.
Molin, B. 1966. "Political Research in Scandinavia, 1960-65: Sweden", in Scandinavian
Political Studies Yearbook, Vol. 1. Helsinki: The Finnish Political Science Association.

Moul, W. B. 1974. "On Getting Nothing for Something. A Note on Causal Models of Political
Development", Comparative Political Studies 7, 139-164.

Naroll, R. 1965. "Galton's Problem: The Logic of CrossCultural Analysis", Social Research
32, 428-451.

Newton, K. 1991. "The European Consortium for Political Research", European Journal of
Political Research 20, 445-458.

Page, E. C. 1990. "British Political Science and Comparative Politics", Political Studies 38,
438-452.

Przeworski, A. & Teune, H. 1970. The Logic of Comparative Social Inquiry. New York: John
Wiley.

Rokkan, S. 1970. Citizens, Elections, Parties: Approaches to the Comparative Study of the
Processes of Development. Oslo: Universitetsforlaget.

Rose, R. 1990. "Institutionalizing Professional Political Science in Europe: A Dynamic
Model", European Journal of Political Research 18, 581-603.

Rose, R. 1991. "Comparing Forms of Comparative Analysis", Political Studies 39, 446-462.

Ruin, O. 1969. "Political Science in Sweden in the Post-War Period", in Scandinavian Political
Studies Yearbook, Vol. 4. Oslo: Universitetsforlaget.

Ruin, O. 1978. "Political Science Research in Sweden 1960-1975: An Overview", in Scandinavian
Studies Yearbook, Vol. 12. Oslo: Universitetsforlaget.

Sartori, G. 1991. "Comparing and Miscomparing", Journal of Theoretical Politics 3, 243-257.

Scheuch, E. K. 1990. "The Development of Comparative Research: Towards Causal Explanations",
Øyen, E., ed., Comparative Methodology. London: Sage Publications.

Sigelman, L. & Gadbois, G. H. 1983. "Contemporary Comparative Politics. An Inventory
and Assessment", Comparative Political Studies 16, 275-305.

Sjoblom, G. 1977. "The Cumulation Problem in Political Science", European Journal of
Political Research 5, 1-32.

Teune, H. 1990. "Comparing Countries: Lessons Learned", in Øyen, E. ed. Comparative
Methodology. London: Sage Publications.

Urwin, D. W. & Eliassen, K. 1975. "In Search of a Continent: The Quest of Comparative
European Politics", European Journal of Political Research 3, 85-113.

Vanhanen, T. 1990. The Process of Democratization. A Comparative Study of 147 States,
1980-88. New York: Taylor & Francis.