Institute for Mass Communication Research,
University of Oslo

1. Introduction

In earlier papers and articles an outline has been given of phase movements in the development of national press systems.¹ Basically the idea is that the national press expands in terms of consumption to about its present level in a limited period seen in relation to the history of the system as such. Historical data, going back more than 120 years, for both Great Britain and the United States clearly manifest these phase movements in circulation. Wadsworth² demonstrates it by trends in aggregate circulation in the United Kingdom, while DeFleur³ has depicted an accumulated diffusion curve for the number of newspapers per household in the USA, which portrays the S-shaped curve typical for phase movements. In Scandinavia this development has been most fully analysed for Denmark, and this case clearly conforms to the others.⁴ Concomitant shifts in technology and market structure may easily be traced for various countries: capacity of production steadily improving and market structure moving from a centralized press through widespread local competition towards a state of local monopolies. To this process of concentration is sometimes added a drive towards national dominance and/or coordination in the production of newspapers.⁵ The long-term movements - in consumption and competition - are not exactly synchronized, but they nevertheless have an interdependent character. Likewise the expansion phase may occur at slightly different moments in the development of national political systems. Since the press as a system is thought to be more responsive to social conflicts in this period, these variations may in turn prove decisive for the ensuing stages of development in the political press.

The purpose of this paper is to bring this line of reasoning one step further and

---

* This is a revised version of a paper presented at the European Consortium for Political Research, Strasbourg 1974, Workshop on The Political Role of Mass Media. The report is also part of a project on the Development of the Norwegian Press, supported by the Norwegian Research Council for Science and the Humanities. I also wish to acknowledge the valuable assistance of my colleague Pål E. Lorentzen.

* This is a revised version of a paper presented at the European Consortium for Political Research, Strasbourg 1974, Workshop on The Political Role of Mass Media. The report is also part of a project on the Development of the Norwegian Press, supported by the Norwegian Research Council for Science and the Humanities. I also wish to acknowledge the valuable assistance of my colleague Pål E. Lorentzen.

to demonstrate some consequences of it. First we shall add some details to the description of phase movements, and secondly we will present some evidence that political conflicts which were instrumental hi forming the political press during the expansion phase have been able to survive to a greater extent in the audience market than in the electorate.

The fundamentals of market concentration as they are applied in this article are well known to students of industrial organization. The level of concentration in an industry evolves mainly from two factors: the size of an optimal plant and the size of the market.⁶ If we assume plant capacity to be fully employed and constant, an expanding market will mean a less concentrated industry, implying a more even size of firms and/or an increasing number of firms.⁷ Empirical tests of these propositions over periods from 10 to 16 years have led to ambiguous conclusions ⁸ In the long run, however, the two factors are clearly interrelated to each other: as the market expands, more efficient technology will be developed.⁹ The two factors may also be seen as mutually compensating in their impact on market concentration; a limited capacity for production may well make for a high level of concentration, provided there is limited consumption.¹⁰ Figure 1 depicts the most typical interconnections between the three variables:

The newspaper industry differs from other industries mainly hi the variability of the journalistic product, which reflects the variety of demands in the audience market. From this point of view the national press is rather a collection of diverse local markets than an integrated industry operating on a national market. In journalism and distribution the press may be compared to retailing, while in production the economy of scale typical for large industrial complexes is a more valid analogy. In both of these industries, however, the basic laws of market concentration have been shown to apply.

2. Social and Geographic Aspects of Expansion

In addition to the temporal movements hi consumption which the phase model mainly emphasizes, the spread of newspapers may also be characterized along two other dimensions: the geographic and the social. On both dimensions the expansion moves from the centre towards the periphery. The rate and extent of

expansion on these dimensions varies somewhat, however, creating certain interesting
phase-lags.

Publication activity naturally comes earlier to the populous urban centres. In less developed countries there is a sharp difference in consumption of mass media, as well as in knowledge of the political system, between urban and rural districts.¹¹ This difference may even be traced to a limited extent in highly developed media systems.¹² As consumption rises in the centre it triggers off publishing activity in more remote areas until a limit is found.

Historically the spread of publishing activity hi developed media systems deescalated from larger to smaller markets. Variations within this pattern may be explained by differences in average income of the population.¹³ Consequently, towns that were first established as publishing towns have also been most stable, while greater turnover has been observed among later additions to newspapertowns ¹⁴ While thus becoming part of the ongoing urbanization from the mid nineteenth century, the spread of publishing has clearly been overcome by more recent increases in the number of urban areas. This is true for Scandinavia as well as the United States, and probably elsewhere.¹⁵

This pattern of diffusion may yield different structures in the national press, depending somewhat on transportation facilities, and other infrastructures. Owing to their topography, Norway and Switzerland have a widely dispersed local press, while the press in more densly populated and flat countries, like Denmark and the United Kingdom, is more centralized.

Social differences in newspaper-reading also level out while consumption in general rises. As the audience-market expands towards the middle and working classes, political social differences make it increasingly difficult to balance the diverse interests of readers within one journalistic product alone, especially when conflicts arise.

This idea was subjected to a limited test for Norway in 1885 at an early stage of party formation as well as in the expansion of the Norwegian press. The age distribution of the party press in 1885 showed the older newspapers to be proportionally more conservative, reflecting dominant attitudes of the early newspaper-reading public, while the younger newspapers had more liberal inclinations. The balance between conservative and liberal newspapers in the local press was also influenced by the electoral majority. While the conservative press was more numerous, despite the fact that the party became a loser in the elections, liberal newspapers tended to appear earlier and in proportionally greater numbers in towns where the liberal party had a clear majority.¹⁶

Territorial and socio-political differentiations follow each other in the national press system. The limit to market segmentation is set by the least accessible social groups, not sufficient in size or economic ability to form a viable newspaper public. Territorial differentiation seems more penetrating, since the demand for local news may overrule social differences. Territorial differentiation, then, both starts earlier and lasts longer. Newspapers established in the last two decades are small, parochial, and mainly apolitical. Today newspapers of the geographic periphery are predominantly uncommitted politically.¹⁷

3. Temporal Stratification in the Press

In addition to geographic and political differentiation, a temporal stratification may easily be identified in the press. With but one exception, today all the larger papers in terms of circulation were started well before the turn of the century. As indicated by Table I, 61 per cent of present local market-leaders were started before 1900. We also learn from this Table that the size of local markets corresponds to the period of establishment for present market-leaders: the larger the local market the earlier establishments tend to be. This of course also reflects the process of geographic differentiation mentioned above. Although some newspapers in less fortunate market positions may compare in age to market-leaders, they tend, on the average, to be younger.

A temporal stratification arises from the widely different economic terms meeting new entries as the audience market expands and later consolidates. Through the expansion phase a rapid rise in newspaper reading makes for a short period of introduction before new entries make a profit. As the market approaches saturation, this period tends to be stretched, leaving new entries to appeal to readers of more established and well-funded papers. Figure 2 illustrates what I have in mind.

In Figure 2 we have made the following two assumptions: a) that the minimum profitable circulation (x) remains constant, and b) that new entries will find their readers in the expanding market only, leaving the already utilized part of the market to earlier starters. The Figure then demonstrates that the condition of entry during the expansion phase allows for altogether four of the five establishments occurring on the market. When the market saturates the period of introduction (it) will grow sufficiently to prevent new entries.

Under more realistic conditions x will become a variable which tends to increase, providing less room for new entries than we assume. To some extent new entries will also take their readers from the audience of more established newspapers. Consequently, early starters may also succumb, allowing for turnover throughout all of the three phases.

Another indication of temporal stratification comes from a project on the historical statistics of the Norwegian press. We have collected data with 5-year intervals on the number of newspapers, places of publication, etc., since 1862. On the basis of these data four sub-periods were identified: 1) early expansion until 1880; 2) a major expansion period from 1880 to 1915; 3) a period of transition when the structure of the press was stable, but turnover high, stretching from 1915 to 1939; and finally 4) a period of consolidation and market concentration from 1950 to the present. Close-downs compared to new establishments made up 42 per cent through the expansion phase 1880-1915, rising to 95 per cent hi the period of transition, and reaching 209 per cent in the last 25 years.¹⁸ Judging from the turnover rates, it seems fair to expect that the chances of survival from sub-period to sub-period vary with the time of entry: i.e., the earlier a newspaper is established the longer it will tend to be sustained in the market (proposition one).

To demonstrate this proposition on temporal stratification we singled out places of publication which existed both hi 1865 and 1965 and arranged all newspapers published according to their period of establishment. The rate of decline for each age-group of newspapers was then isochronized to the varying starting points indicated in Table II below.

The regularity which the turnover rates suggested comes through clearly in this Table: all age-groups are reduced in size over time, irrespective of their period of establishment and independently of the year from which the reduction is computed. Proportionally, however, the younger age-groups must take the highest reduction. It took a hundred years to reduce newspapers publishing hi 1865 to one-third of their original size. For later starters this rate of reduction took an increasingly shorter period of time: i.e., 25 years for newspapers started between 1921 and 1939.

The reason why older newspapers stand a better chance of survival may be derived from their fortunate basis relative to the expansion phase: they have greater potentialities to draw on in the audience market. In general they can appeal to new readers and are not forced to convert readers with established loyalties to their competitors. This ability may be decisive in the vulnerable early period of operation for new papers. Looking closely at Table II we discover that

the mortality rate is highest in the first period computed for all age-groups, anc
that this 'initial reduction' increases for the younger age-groups.

One result of this process is demonstrated in Table 111, which shows the age distribution of the press for different years 1900-65. Clearly, the proportion which the older newspapers represent tend to stabilize over time, leaving less room for upstarts.

4. Traces of Phase Lags

The linearity of our data so far gives few indications of phase movements. According
to the model, the oldest newspapers, started well before the expansion
phase, should have a higher mortality rate than newspapers operating in the early

period of market growth. In a longitudinal perspective it is really the middleaged newspapers which should have shown most vitality. There may be several reasons why this pattern has not shown up in our Tables so far, one of them being that our material does not include many of these oldest newspapers. There is, however, one possibility to stretch the initial phase in our material by distinguishing the smaller towns of the periphery from the more populous cities hi the centre.

As was discussed in Section 2 publishing activity moved from the centre towards the periphery. A parallel trend in a number of cities with competing newspapers was also found, although not to the same extent, as may be expected.¹⁹ This suggests that both expansion phase and market saturation come earlier in the larger cities than hi the smaller towns. These differences will also make for a difference in the mortality rates for newspapers in larger and smaller places of publication, since survival is made relative to the expansion phase.

We do not have sufficient data on circulation to depict this process more precisely for the pre-war period. Using the weight of newspaper volumes as a rough indicator for economic ability or size more generally, we have been able to confirm part of the process outlined. Clustering of volume weights was read as a sign of small differences in circulation, while the spread of weights was interpreted as a skewed size distribution typical for market concentration. From this it was clear that concentration started earlier in the larger markets.²⁰

Taking account of these differences, we shall also expect the temporal stratification to vary somewhat between larger and smaller places of publication: For comparable periods of time, earlier starters among newspapers will survive to a greater extent in larger markets than in smaller ones. Inversely: late starters among newspapers will survive to a greater extent in smaller markets than in larger ones (proposition two).

To demonstrate this proposition we further divided our material into two parts: newspapers publishing in markets with a population either above or below 10,000 residents centrally located through the whole period 1865-1965. Newspapers publishing in towns where the population has moved from less than 10,000 to more than that were left out of the comparison. The computation in Tables IV and V follows the same method used for Table 111.

Table IV confirms proposition one even more clearly than Table 111, which might have been expected. If we then carefully compare Table IV and Table V we shall see our second proposition confirmed also: smaller places of publication deviate from larger ones mainly by being one step behind. Newspapers started in or before 1865 are more drastically reduced in number in smaller markets, while the younger papers, especially those started between 1866 and 1900, stand a better chance of survival here.

For example, it took 100 years to reduce papers publishing in 1865 to 42 per cent of their original size on larger markets, while in smaller markets the same reduction was completed in 75 years. For newspapers started between 1866 and 1900 we will find the opposite trend: they are reduced faster in the larger markets

(to 48 per cent in 40 years) than in smaller ones (to 43 per cent in 65 years). This overall pattern even comes through in the data for smaller markets alone. From Table V we can see that it is the 'middle-aged' newspapers, started 1866-1900, which most stubbornly carry on in the smaller markets.

5. The Timing of Electoral Support

The time of entry relative to market expansion is evidently not the only factor to influence the life expectancy of newspapers. The size and stability of the target group which a newspaper appeals to might be seen as an even more decisive factor.

In a study of newspaper competition in eleven Swedish towns Lars Furhoff found some support for the hypothesis that newspapers founded first eventually became dominant in circulation.²¹ He adds, however, two important modifications: the first modification is that 1920 is a year from which a more automatic market concentration gains momentum, i.e., that papers which had won market leadership at this time would steadily improve their position in the audience market and drive out weaker competitors. Not by any accident, the 1920s might also be seen as the end of the first major expansion of the newspaper industry. Another modification which Furhoff points out is that electoral support is instrumental in winning a favourable position in the audience market prior to 1920.

Furhoff had to rely on circumstantial data. Niels Thomsen by contrast was able to reconstruct complete tune series of circulation data back to the 1870s.²² Thomsen can then demonstrate for Denmark what Furhoff speculated for Sweden: the correlation between the size distribution in the party press and the comparable distribution of valid votes steadily improves towards the 19205, but is then gradually weakened. This relationship is also reflected in the succession of establishments moving from conservative towards liberal/radical and labour papers. But the pace of the pendulum was greater in the middle than at the ends of the swing. In general the liberal press was able to capture a lion's share of the emerging middle class in politics as well as a fair share of the working class. The upcoming labour press was never able to reach more than two-thirds of labour voters. Throughout this period, previous market leaders lost ground where they had been locked into a political stance out of tune with the electorate. As the pendulum swung back, close-downs tended to come in a reverse order hitting the labour press hardest.

Such observations suggest that all newspapers surviving to the present must have had strong support in the local electorate for some time. On the other hand, taking into account what was found on temporal stratification, time of entry and electoral support may be viewed as partially compensating factors. The next problem to consider then concerns the timing of the two factors relative to each other: which is most important and at what time and in what type of market in the development of the national press.

As we have already learned (p. 162), 'infant mortality' is high even among

early starters. It is in this vulnerable early period of operation that we should expect electoral support to be most decisive for later success, irrespective of time of entry of newspapers, but more decisive as the market gradually becomes saturated (proposition three).

In Table VI we have taken all dailies publishing in 1965 and classified them according to their period of establishment. Each sub-period indicated covers six general elections. Thus the basis for computation in each cell is the number of newspapers operating multiplied by six. The numbers express instances where newspapers operating in the sub-period were committed to the dominant party in their community in per cent of all possible instances. As will be seen, most of our expectations are confirmed.

For the oldest newspapers strong electoral support was more widespread in their early period of operation and has since been on the decline, most typically after about 1920. Younger newspapers demonstrate the opposite trend: their electoral support has been on the rise. The disadvantage of a late start is evidently compensated for to a certain degree by a strong electoral support to prevent closedowns.

The ability to survive and the possibility of market dominance merge in the long run. Under conditions of market concentration the last surviving paper will also be a winner in terms of circulation. But even before this stage the factors instrumental for survival will tend to differentiate local newspapers according to

circulation. It will only be a matter of degree. In other words: market leaders compared to newspapers in less fortunate positions will in general have a) a more favourable time of entry, and b) more widespread electoral support (proposition four).

Since circulation data are not available to the extent required here, we have settled for a reconstruction instead. We assume that present market leaders have won their position gradually and not by circumstance. The dailies publishing in 1965 were then dichotomized according to market position and period of establishment. Electoral support was computed for Table VII in the same way as was done in Table VI.

The pattern which emerges from Table VII lends credence, if not firm support, to the latter part of proposition four above. Among the older newspapers, market leaders compared to others have been able to capitalize on more widespread electoral support if we consider all sub-periods. The main difference between market leaders and papers in less fortunate positions is that electoral support for the latter has declined more drastically. Before 1920, however, the situation is less clear.

For newspapers established after 1900 the picture is much more in line with
what we have expected. Market leaders compared to others have had more widespread
electoral support, and their support has been more clearly on the rise.

The overall difference between older and younger newspapers which comes
through in Table VII, irrespective of market position, lends support to proposition
three.

6. On the Dynamics of Expansion: A Summary

A short survey of the ground covered so far may be appropriate at this point:

The national press expands from the geographic and social centre towards the periphery. As the market expands, newspapers tend to appeal to the most available among distinct groups of the public. The ensuing market segmentation stops short of the smaller communities, where the press tends to be apolitical or uncommitted. Segmentation also leaves out minority groups in competing markets, being too small to form a viable readership. Sometimes such shortcomings may be compensated for by newspapers having a nation-wide distribution.

On balance, this process increases the capability of the press to reflect a given configuration of conflict and interest in society, producing a close interrelationship between party strength in the electorate and the press. However, structural movements in the newspaper industry have dynamics of their own which eventually

overrule changing party preferences in the audience. Some details of these
dynamics evolving from two phenomena called temporal stratification and the
timing of electoral support have been identified in this paper.

Temporal stratification reflects the varying capability of newspapers to gradually aggregate support at the margin of target groups defined by their party commitments. This capacity varies with the tune of entry into the industry relative to the expansion phase of the audience market. It will give early starters an advantage provided they were initially able to select the optimal target group in the long run. This of course can only be determined from hindsight.

In describing the impact of electoral support the term 'timing' reflects the shifting character of this factor. Strong electoral support will always be important to some newspapers at some tune, but not to all at the same time. Electoral support seems most important in the early period of operation. This period of introduction, before new establishments can make a profit, tends to stretch as the market expands, which increases the demand for stability and strength hi electoral support. Over time, early starts have been successful in aggregating sufficient support outside their given market segment to become less dependent on fluctuations in the electorate. These papers by their own success prevent new establishments simply because the audience market has its limits. As the market saturates and the industry becomes closed, this mechanism will tend to lock the press as a system into traditional positions, reflecting more the political configuration some decades ago than the actual contemporary situation.

7. The Party Press 1885-1985

The futurology in this section is based on the mechanics of newspaper close-downs as they have occurred for the last twenty years. With few exceptions closedowns occur in competitive markets and hit newspapers with the smallest circulation first.

More exactly our assumptions have been the following four: a) all present places of publication will be operative throughout the period, b) new papers will not be started, c) all party commitments will remain unaltered, and finally d) newspapers with the smallest circulation in the most competitive markets will always go first. On this basis our more detailed projections are:

By 1975: There will be no towns with three newspapers locally published, except for Oslo and
Bergen where local newspapers will be reduced by one in each city.

By 1980: Towns with a population less than 20,000 have developed newspaper monopolies.
At other places of publication the situation remains unchanged except for Oslo, which
publishes five papers, and Bergen where the number will be reduced to two newspapers.
By 1985: All places of publication will have only one newspaper, except for the three largest
cities outside Oslo, where there still are two local papers, and the capital itself which
publishes four newspapers.

From this progression, simple arithmetic tells us that 42 newspapers will fold up by 1985, 38 of which belong to the party press. The gradual erosion of the party press is projected in Table VIII. Up to 1965 this Table is based on empirical data. For simplicity party commitments have been grouped into four major political movements and ordered vertically according to their time of entry (as a group) into the party press.

If we compare Table VIII and Table IX, the basic logic of this paper is brought out again. As the number of newspapers increases until about 1939, first the conservative and then the liberal bias are levelled out as the variety of the party press increases. However, a certain bias remains and it is in favour of the parties that first gained a foothold in the press. As the number of newspapers starts to decline, this bias is again emphasized in the direction of traditional parties. The agrarian press deviates slightly, and the reason is that some of the agrarian newspapers were started before the party.

Table VIII conceals an even more striking bias hi the local press. By region the party press is heavily dominated by the party first establishing a stronghold in the district, irrespective of the fact that their support may have declined later on. Thus the conservative press dominated the Eastern region, while the liberal press is virtually confined to the Western coast. In the North the party press is polarized

between conservative and labour. Not accidentally the first Labour representatives
were returned from this region. These regional contrasts, based more on
historical facts than on current events, will become even sharper in the future.23

8. Postscript

These national and regional imbalances between the electorate and the party press represent historical elements in the present structure of the Norwegian press, shaped by market conditions in the past and by later developments. Such elements are not fully comprehended from a given situation alone; only over time and from hindsight may the riddle be solved. By the logic of a simple market mechanism and the phase model the development of such historical elements was traced.

The temporal stratification found in the development of the Norwegian press fits nicely into a so-called sequential model as outlined by Sidney Verba.²⁴ Principal properties in such models are defined as dynamic explanations where at least some of the explanatory variables used belong to an earlier period and where the terminate state is explained by a number of initial conditions. Sequential models make the dependent variable hi one phase an independent variable through the next. In such directions some of the celebrated long-term effects in mass communication may also be found: by going back and attempting to explain how the weight of earlier choices and market conditions shape the alternatives presented to the readers of today. The choice of mass media is in some respects like the use of a language. Only by accepting much of what is given may a person use it as a means of communication. Changes come gradually and are not influenced by individual attitudes alone.

Rokkan and Torsvik²⁵ found strong interrelations between loyalty to the party press in the electorate and political activity. No doubt the party press is instrumental in maintaining such cores of stable and opinionated voters who may be vital to the party in the next two or three steps of opinion formation. Rokkan and Torsvik also found that the propensity to choose a newspaper according to party preferences among voters increased with the increasing availability of a relevant party paper hi the community. Such findings in turn point to factors we have classified by terms like 'temporal stratification' and 'timing of electoral support'.

---

NOTES

1. These models, alternately named 'transactional' and 'phase' models, have been described in, e.g., Svennik Høyer, A Transactional Approach to the Study of A National Press System, paper prepared for the IPSA Round Tableon 'Democracy and Information', Bucharest, September 1972; Svennik Høyer, 'En transaksjonsmodel for studiet av pressen', Tidsskrift for samfunnsforskning, Vol. 14, 1973, pp. 81-105 and in Svennik Høyer, Stig Hadenius and Lennart Weibull, 'The Politics and Economics of the Press: A Developmental Perspective', Sage Professional Papers in Contemporary Political Sociology, No. 06-009 (London and Beverly Hills, California: 1975).

---

2. A. P. Wadsworth, Newspaper Circulation 1800-1934, Manchester Statistical Society, 1955 (London, Crosly Lockwood, 1967).

3. Melvin L. DeFleur, Theories of Mass Communication, Second Edition (New York: David McKay Company Inc., 1970).

4. Niels Thomsen, Dagbladkonkurrencen 1870-1970, (Politics, Journalism and Economics in the Structural Developmcnts of the Danish Press 1870-1970) (Copenhagen: G. E. C. Gads Forlag, 1972).

5. See: Høyer, Hadenius and Weibull, op.cit.

6. See, for instance, Frederic M. Scherer, Industrial Market Structure and Economic Performance mance (Chicago: Rand McNally & Co., 1970) and M. A. Utton, Industrial Concentration (Harmondsworth: Penguin Books, 1970).

7. See, e.g., R. L. Nelson, 'Market Growth, Company Diversification and Product Concentration', Journal of The American Statistical Association, Vol. 55, 1960, pp. 640-49; and W. G. Shepherd, 'Trends of Concentration in American Manufacturing Industries 1947-58', Review of Economics and Statistics, Vol. 66, 1964, pp. 200-212.

8. David R. Kamerschen, 'Market Growth and Industry Concentration', Journal of The American Statistical Association, Vol. 63, 1968, pp. 228-241.

9. Jacob Schmookler, Invention and Economic Growth (Cambridge, Mass.: Harvard University Press, 1966).

10. Frederic M. Scherer, op.cit. (note 6), pp. 72-103.

11. Wilbur Schramm, National Development and Mass Communication, (Stanford: Stanford University Press, 1964).

12. Egil Augedal, Fritidsbruk i Oslo og Ha, (The Use of Leisure in Oslo and A West Coast Rural Community), Oslo, Institute for Mass Communication Research, Report no. 34, 1974.

13. This pattern has been analysed, e.g., for Denmark in Thomsen, 1972, op.cit. (note 4), and for Norway in Svennik Høyer, Norsk presse mellom 1865 og 1965: Strukturutvikling og politiske mønstre, (The Norwegian Press 1865-1965: Structural Developments and Political Patterns), Oslo, Institute for Mass Communication Research, Report no. 32, 1975.

14. See Høyer, 1975, ibid.

15. For Scandinavia, see Høyer, 1975, ibid. and Thomsen, 1972, op.cit. For the United States see Ben H. Bagdikian, The Information Machines (New York: Harper & Row, 1971), p. 80.

16. Svennik Høyer, 'The Political Economy and the Norwegian Press', Scandinavian Political Studies, Vol. 3/1968 (Oslo: Universitetsforlaget, 1969). For a more elaborated version of these findings, see also Høyer, 1975, op.cit.

17. Høyer, 1975, op.cit.

18. Høyer, 1975, op.cit.

19. Høyer, 1975, op.cit.

20. Høyer, 1975, op.cit.

21. Lars Furhof f, Tidningsdoden på elva orter - forslag til ett politiskt perspektiv', (Newspaper Close-Downs in Eleven Communities. A Political Perspective), Statsvetenskaplig Tidskrift, p. 211 f f., 1960.

22. Thomsen, 1972, op.cit.

23. Stein Rokkan and Henry Valen, 'Regional Contrasts in Norwegian Politics', in Erik Allardt and Yrjo Littunen (eds.), Cleavages Ideologies and Party Systems: Contributions to Comparative Political Sociology (Helsinki: Westmarck Society, 1964), pp. 162-238.

24. Sidney Verba, 'Sequences and Development', in Leonard Binder et. al., Grises and Sequences in Political Development (Princeton: Princeton University Press, 1971).

25. Stein Rokkan and Per Torsvik, 'Der Wåhler, Der Leser und Die Parteipresse. Eine Analyse iiber politische Einstellung und Lesen von Zeitungen in Norwegen', Kolner Zeitschrift fiir Soziologie und Sozialpshychologie, 12. Jahrgang, Heft 2/1960. An English translation may be found in Stein Rokkan, Citizens, Elections, Parties (Oslo: Universitetsforlaget, 1970).