1. Introduction

In 1951, a British study concluded that the problem of poverty in the welfare state
was close to being eliminated. It found that in 1950 only Wi% of the population of
York lived under conditions which could be described as poverty.1

In 1965 another equally thorough British study appeared, but with remarkably different conclusions. According to this study, poverty was not only a persisting problem of great size, but the proportion of the population living in poverty was rising.₂ In 1953, the study concluded, 7.8% of the British population lived in poverty and by 1960 the proportion was as high as 14.2%.

How could two comparable studies reach such fundamentally different conclusions?
The answer, of course, lies in different definitions of the concept of 'poverty'.

The first study, undertaken by Seebohm Rowntree and G. R. Lavers, employed a moralistic definition of 'poverty'. A group of 'experts' decided what commodities, and in what quantities, should be regarded as a necessary minimum for families of various sizes. On this basis, it was easy to calculate poverty lines and find out how many families had incomes which were insufficient to purchase the goods included in the list of necessities.

Peter Townsend, co-author of the second study referred to above, criticized this approach
by pointing out that there was no way of putting together an objective list of
necessities:

If clothing, money for travel to work and newspapers are considered to be neces

saries' in the conventional sense, why not tea, handkerchiefs, laundry, contraceptives,
cosmetics, hairdressing and shaving, and life insurance payments?3

In the second study, entitled 'The Poor and the Poorest', a different approach was used to define 'poverty lines'. Here it was defined as the basic sum payable by the National Assistance Board (NAB) with a 40% addition for extra payments that the Board could make.

This technique is different from the one used by Rowntree and Lavers in two ways. Firstly, the definition of 'poverty' is not left to the researcher. It is assumed that the rules governing NAB payments imply a political definition of 'poverty'. Secondly, poverty is not seen as a static or mainly static phenomenon, but as a quality dependent on the culture of the society in question. Poverty is not simply a situation in which basic needs are lacking, but an inability to live according to prevailing norms in society. In The Affluent Society, Professor Galbraith puts it this way:

. . . people are poverty-stricken when their income, even if it is adequate for survival,
falls markedly below that of the community. Then they cannot have what
the larger community regards as the minimum necessary for decency, and they

cannot wholly escape, therefore, the judgement of the larger community that they
are indecent.4

For our purpose, the important change that has taken place in the discussion of poverty is that basic needs are no longer regarded as a question simply of food, clothing and shelter, but as a question of being able to live with a minimum of comfort as defined by one's surroundings.⁵

As demonstrated by these two studies, this change produced a dramatic 'rediscovery'
of poverty. This rediscovery, however, was not based simply on a redefinition of concepts,

there was also a 'real' rediscovery in the sense that existing elementary want and
deprivation was brought to the forefront. In the United States, this came about first of
all through Michael Harrington's dramatic book, The Other America.0

Paralleling the rediscovery of poverty is another international trend which should be noted in this introduction. It is often labeled 'the social indicators movement', and has led to a broadening of the concept of 'welfare'. It originated through dissatisfaction with measuring welfare in economic terms only. Markets are not perfect and money is only one of several resources necessary for one to act rationally in markets. In social indicators, 'welfare' is defined as a complex concept with several dimensions that cannot easily be put together in a simple index. Economic indicators capture but a few of the relevant dimensions and therefore do not enable a wide evaluation of progress or decline.7

These two trends - the rediscovery of poverty and the broadening of the concept of 'welfare' - have had an enormous impact on the scientific community. Economists have demonstrated a growing interest in micro-economics with heavy stress on welfare problems. Studies of distribution - not between factors but between individuals or groups - have come to the forefront. Sociologists have established a new and now rapidly growing discipline of the sociology of poverty.

International development in this field has made its mark in Scandinavia as well. There has been a shift from welfare state optimism to fairly strong criticism and skepticism. The notion that poverty must be related to the culture of each society has produced the belief that pensions do not guarantee general affluence. Studies on income distribution show that in spite of economic growth and serious attempts at redistribution, inequality prevails all but unchanged. The search for poverty has demonstrated that fairly substantial groups - in particular in new urban residential areas and among the old - are unable to meet everyday needs even with apparently reasonable incomes. Researchers and politicians alike have come to stress the need for applied problem oriented research. In both economics and sociology there has been a movement in the direction of studies where the aim is to improve the ability to attack specific social problems rather than to bring about an understanding of economic and social mechanisms in an abstract sense. Political authorities allocate increasing funds for contract research in politically defined problem areas.

In this context, a number of large and ambitious research projects have been initiated and in the following I will present some notes of comparison on three of them - the Swedish Level of Living Survey, the Comparative Study of Scandinavian Societies undertaken in Finland and the Norwegian Level of Living Study, hereafter referred to as the Swedish, the Finnish and the Norwegian studies.B

2. Initiatives

The Swedish study was started in 1967 and ended late in 1971. It was undertaken as one of three projects under the Low Income Committee which was set up by the Government in 1965 to study the problem of low income in Swedish society and 'to consider the type of information necessary to give the low income discussion an adequate basis'.

The other two projects undertaken by the Committee were purely economic: one concerned the distribution of income, the other the distribution of purchasing power. The function of the Level of Living Survey was to broaden the basis for the discussion on low income by including a number of social elements.

The point to be made in this paragraph is that the study was started purely as a result
of a political initiative. The Low Income Committee was an ad hoc government
agency, one of whose tasks it was to undertake a survey of the kind under discussion

here. It was not clear from the political initiative what this survey would in fact be like, and it turned out that the result was not easily digestible by all members of the political establishment. The actual shaping of the project was left to the researchers but the whole process started outside the scientific community.

The Finnish study is quite different in this, as in several other respects. It is a research project in the conventional sense. By this I do not mean that it is less interesting or of less importance or that it is more old fashioned, I only want to stress that it was initiated and undertaken by researchers who did not have the direct links to political institutions that are characteristic of the Swedish study and even more so of the Norwegian one. This difference implies a number of consequences in the organization of the studies and also, I believe, in their theoretical foundation.

In Norway, the Level of Living Study was started in 1972 and is planned to be completed by the end of 1975. Like the Swedish study, it is purely a political initiative. It started with a letter to the Government from the National Federation of Trade Unions asking for a low income and level of living study to be undertaken in Norway. Several members of Parliament had also submitted similar proposals. Obviously the Swedish study was a source of inspiration for Norwegian politicians. It is interesting to note that at the time the Swedish study was disestablished for political reasons, the Norwegian study was launched through a political initiative.

Unlike what happened in Sweden, the Norwegian Government not only initiated the study, it to a fairly large extent framed its content and approach. Again, the Swedish study is the background. The Swedes had undertaken a pioneer conceptualization of the term 'level of living' and their concepts were more or less copied in the commission the Norwegian Government issued for the Level of Living Study.

Thus we have one study with no direct links to political institutions, a second study
initiated politically, and a third one with a political initiative and fairly precise directives
attached to the initiative.

3. Purposes

Because of the different types of initiatives behind the studies, they have fairly different explicit purposes. In the Finnish study, the purpose is to do research and produce knowledge. The other two studies have broader purposes related to the fact that they are not just research projects, but are also Government appointed committees. Again it should be stressed that this is not necessarily a difference in quality to the advantage of the Swedish and Norwegian studies. The fundamental background and purpose of all three projects are applied research within the field of welfare. In terms of practical results, the links with the political community might very well turn out to be a disadvantage.

In addition to the purely scientific purposes of the Swedish and Norwegian studies, they were given explicit purposes with regard to each country's system of national statistics and to practical policy making in the welfare sector. For both studies, an important reason for their starting was some dissatisfaction with the state of statistical information. In the two countries, wage equality had long been a high priority goal. Still, existing statistics did not make reliable evaluation of the development in income distribution possible, nor did they enable one to make statements on the size of low income groups in these welfare states. Sten Johansson - responsible for the Swedish study - has stated that the most important practical result of their study would probably be a reconsideration of national statistics in the direction of reproducing - more or less - the kind of statistical reviews of the Swedish society that were produced through the study. The directives for the Norwegian study include as one of its purposes to create a basis upon which relevant statistics on social indicators might be organized in the future.

An additional purpose of these two studies is to produce a foundation of information upon which practical political initiatives could be made for improving welfare. This was a consideration which strongly influenced the Swedish study and also - though somewhat more indirectly - the Norwegian one.

In deciding what social indicators to include, the Swedish study employed a pragmatic approach. The criterion of 'manipulability' was a point of departure and since the relevant manipulating organizations were judged to be Government agencies and trade unions, the study was organized in an attempt to produce practical information for the existing administrative structure. It provided, so to speak, one indicator for each ministry. Johansson and his colleagues were aware of the static bias in this approach and partly for that reason included a review of political resources where they asked the question of to what extent the Swedish people considered themselves able to influence the outcome of political processes. The result of this review was rather discouraging compared to the classic ideals of democracy and in a sense underlined the status-quo-like approach of the entire survey.

I have previously indicated that the directives given for the Norwegian study were, let me say 'inspired' in order not to exaggerate, by the conceptualization and experiences of the survey undertaken in Sweden. A brief description of these directives should therefore give an idea of the subject matter of both studies.

The directives were prepared by a committee with members from several ministries
and the Federation of Trade Unions. They indicate both what is to be understood by
'level of living' and what the main approach of the study should be.

Level of living is defined as a number of components in regard to which an individual
may have a favorable or unfavorable position:

1. Income

6. Health and nutrition
7. Childhood quality
8. Family life

2. Employment

3. Working conditions
4. Housing

9. Leisure
10. Political participation

5. Education

The approach is specified in six points where the last one is on a somewhat different
dimension from the other five.

Firstly, the situation in Norwegian society should be described for each component,
and in presenting these descriptions, differences should be stressed rather than averages.

Secondly, to the extent possible, the study must explain the existence of these differences.
Because of the nature of the study, explanations which point to possible political
actions are of course of special interest.

Thirdly, a total perspective should be employed so that, to the maximum possible extent, it demonstrates how an individual's position in one component is likely to be correlated with his position in other components. Of special interest in this context is the extent to which level of living is described when income is known.

Fourthly, special attention should be given to groups of the population with an extraordinarily poor level of living. The problem of low income was an important point of departure for initiating the study, and remains of special concern. The questions are: What is the extent of Norway's poverty problem? What does it imply in terms of the other components to be poor?

Fifthly, geographical and regional variation must not be left out. Norway is a
sparsely populated country and there is fair agreement that regional equality of opportunity
should be maximized.

Finally, on a somewhat different dimension, the study should include the distributive
effects of public policy and public enterprises.

Since the administrative structure of the Norwegian Government is fairly similar to

that of Sweden, the static bias of 'one indicator for each ministry' is preserved in the directives. As will be indicated later, however, the frame of reference adopted includes an attempt at counterbalancing this status quo approach by emphasizing social and economic structures and not only administrative structures. The directives also underline the 'total perspective' more than in the case of the Swedish study. The problem of the distributive effects of public policy was not a part of the Swedish survey.

The Finnish study is very different from those discussed above, which are being undertaken on a purely national basis. In being comparative and including the four Nordic countries of Denmark, Finland, Norway and Sweden, the Finnish study raises an important principle question. In a comparative study, one always has some standard of comparison in how the situation in one country is seen in relation to that in other countries. In studies which include only one nation and, in addition, with very few opportunities for historical comparison, there is virtually no standard against which empirical results might be compared. When, for example, the Swedish study found that 9.1% of the adult population had experienced unemployment or underemployment in some form during the year 1966, was this alarming, normal, acceptable or encouraging? One could say that the question is not important in relation to the study, since evaluation of the results might be left to politicians or the general public, or since it is only a baseline piece of information which becomes comprehensible only when statistics using the same definitions are compiled at some later point in time. Another possibility would be to establish 'desired states' against which results could be compared. This, however, is fairly difficult both theoretically and because of the link to the political community. Therefore, the actual evaluation is in fact left to someone else. In the Swedish study, the intention was that the Low Income Committee should sum up its results and suggest political initiatives, but probably because of their unexpectedly controversial nature, responsibility for summing up was eventually taken out of the hands of the Committee and left to another group under stronger political control.

In the case of the Norwegian study, the directives clearly state that it must not itself
suggest political consequences of its findings, but merely present them in a way that
allows specific initiatives to be discussed practically.

In a more pragmatic perspective, the problem of non-comparability is not all that serious. There are in fact standards of comparison even though they are not explicit. As demonstrated by the Swedish study, the welfare state reality is in fact so different from the broadly accepted ideals that, with little or no qualification, it is recognized as 'problematic.

It is true that studies of the kind we are discussing here could be framed in a way that emphasized the bright sides of Scandinavian welfare. This, however, is not the purpose. As indicated in the introduction, lack of welfare should be defined in terms of the culture of the society in question. At least in principle, equality is part of the value culture of Scandinavian societies. It is therefore 'correct' to stress the inequality found in these societies more or less independently of whether or not it is larger or smaller than what one might find in other societies. The directives of the Norwegian study are fairly explicit on this point. In describing the situation on each indicator, deviation from equality is to be stressed.

The Finnish study is different simply because it compares nations. The 'absolute welfare level' of each country will be overshadowed by the relative level as compared to other Scandinavian countries. From the point of view of political results, the approach might be advantageous to the country which in comparison comes out at the bottom, but disadvantageous to the country which comes out at the top. This is a general problem of comparative research. It presupposes a large degree of cultural equality between countries. Since welfare is related to the culture of a country, the ranking which comes out of a simple comparison might be different from the ranking which would result

from studies where the empirical level of welfare in each country was related, not to that of another country, but to the value culture of the country itself. For example, if wage equality were a stronger element in Swedish than in Danish culture, the difference between ideal and reality might be larger in Sweden than in Denmark even if income inequality was in fact smaller.

In this case, comparisons are made between fairly 'equal' countries, but it is not difficult to find examples where this is not the case. Criticism of poverty in 'affluent societies' is often met by a reference to far greater poverty in underdeveloped nations. In my opinion, this is an irrelevant and trivial fact. The point here is that this bias is found to some extent in all comparative research, even projects like the Finnish study where only very 'comparable' nations are included.

4. Theories

The three studies differ in the type of theory they have based their work on and in the extent of theorizing. The Finnish study has worked out the most elaborate theory, the other two studies simply frameworks for organizing analysis and sorting out elements in their vast subject matter. This difference is probably due to the fact that the Finnish study is purely an academic enterprise and therefore in a sense needs a thorough theoretical foundation and justification, whereas the Swedish and Norwegian studies are defined by outsiders as 'interesting'. In principle their job is simple in that all they have to do is supply information to political elites who have stated that the information is necessary for easily acceptable reasons.

The theory of the Finnish study is based on the concept of 'need satisfaction', and
that of the Swedish one on the concept of 'individual resources', whereas the Norwegian
perspective includes both individual resources and social structures.

Professor Erik Allardt is responsible for the Finnish study and with his colleagues has developed a very exciting and stimulating theory.⁹ The concept of 'individual resources', they state, is analytically practical but theoretically insufficient if the point of departure is need satisfaction. In making a distinction between intrapersonal and interpersonal needs, they find that resources cover only the intrapersonal part. They describe what a person or group has, but when relying only on individual resources, the researcher overlooks the fact that people have needs not immediately related to what they have, but to their position in relation to or in interaction with other people, i.e. interpersonal needs. On this basis, three types of commonly accepted needs are listed: 'having', 'loving', and 'being'. It is indicated that the needs form a hierarchy where 'having' is the most basic.

'Having' is a fairly simple analytical case and covers those resources the individual controls or has access to. 'Loving' is manifested in interaction with other people where some kind of solidarity (not necessarily in the political sense) is the central value. 'Being' covers the individual's relation to other people. The central value is self actualization, a value which is manifested in degree of insubstitutability.

Since these are commonly accepted values, the level of development, or welfare, of nations may be described by measuring their manifestations. According to the theory, they are not manifested on one dimension, but on three: central tendency describing the general level of welfare, dispersion describing inequality, and correlation describing clustering of inequality. This, of course, immediately brings the project into some very tricky problems of operationalization, but that is a technical problem and does not deprive the theory of its fascination.

The Swedish study has a narrower frame of reference relying exclusively on the concept
of 'individual resources',10 i.e. the 'having' element in Allardt's model. The study
rejected an approach based on need-satisfaction because of practical, theoretical and

philosophical reasons, deciding instead to define the concept of 'level of living' in a list of resources or components. It also rejected the idea of putting all the components into one overall concept or index. This was done because a piecemeal approach was seen as more politically relevant, i.e. more instrumental as an indicator of concrete political initiatives.

For each component it was decided to compare the actual situation with 'the good
life' instead of 'evil conditions'. In fact, in the main, comparisons were made only implicitly.

As mentioned previously, the overall criterion for choosing components was manipulability
by the existing political institutions. The result was a list of components equivalent
to the one previously quoted for the Norwegian study.

Although the Norwegian study is strongly influenced by the Swedish experience, a somewhat broader frame of reference was drawn up.¹¹ Also here, the notion of 'need satisfaction' was rejected, but the framework does not rely exclusively on individual resources. Drawing on work by James S. Coleman,¹² it distinguishes between individual resources on the one hand and properties of arenas for social action on the other. This results in a dynamic model where resources at one point in time may be invested in different arenas. The outcome at the next point in time is a new set of resources which in turn may be invested in arenas. Resources are wealth, education, family or social background, etc., and arenas, e.g. the labor market, the educational system, or the housing market.

There are several reasons for the distinction between resources and arenas. Firstly, the criterion of manipulability. Individual resources are not generally manipulable directly. They are produced and distributed in markets or other types of arenas, and from a political point of view the alternatives for action generally go in the direction of changing or influencing what in this perspective would be called 'arena properties'. Secondly, the value of a resource is not a simple function of the size of the resource in question. Good education may be an advantage in itself but it is a real resource in the sense that it may be invested to produce new resources or values only if the labor market where the person happens to live offers relevant jobs for highly trained people. This is important in a country like Norway, where there are very substantial regional differences. Thirdly, the directives for the study included a paragraph on the distributive effects of public policy, a subject which clearly cannot be studied simply in a perspective of individual resources.

In this study, then, 'level of living' has a twofold definition; a function of individual
resources and the structure of the social and economic environment.

5. Data

All three studies are data extensive. Empirical descriptions are of value in themselves.
In the case of the Swedish and Norwegian studies, plain statistical results are a part of
the expected outcome.

The Level of Living Survey in Sweden mainly analyzed data collected through a survey interview constructed by the study itself. Roughly 6000 individuals were interviewed, the sample being representative of the entire adult population. It was a subsample from a large income survey and so provided data on economic resources without actually including many questions on that aspect in the questionnaire.

The analysis was completed in 1972, resulting in no less than 16 different reports.

The Finnish study relies exclusively on a cross-national survey in four countries with
a sample of some 1000 individuals in each.

The Norwegian study has a number of different data sources, among them a survey
13 Scandinavian Political Studies

of its own with a 4000 member sample which is representative of the adult population. It is a subsample of the 1973 Household Consumption Survey and Income Survey of the Central Bureau of Statistics. Household data, income data and consumption data are thus provided for the sample. In addition, other data sources will be analyzed. This includes the 1967 and 1970 Income Surveys giving the income component of the study a diachronic perspective. It also includes a Life History Survey undertaken by the Norwegian Institute of Applied Social Research. This extremely exciting surveyl3 also enables the study to introduce a dynamic perspective into elements other than income. The survey includes samples of three cohorts of men (born in 1921, 1931 and 1941) interviewed about their careers from the age of 14. It includes information on education, occupation, residence, family, and health, and in addition some information on childhood and situation at time of interviewing. Results of IQ tests of some of the respondents at time of entering military service are included. *

A Time Budget Survey completed by the Central Bureau of Statistics should also be
mentioned.

The reason for so much work in data collection and such reliance on the survey technique is not just that one needs data - a vast amount of information on social indicators is available in the systems of national statistics - but these data tend to be based on administrative procedures in that they indicate what the activities of existing institutions are, e.g. number of patients in hospitals, number of students in schools, or number of criminal cases in courts. These statistical indicators demonstrate not what health care is necessary, not what education is demanded or the real level of crime, but what services a given apparatus can offer. This tends to overestimate the real level of welfare as well as to distract attention from necessary administrative reforms.

6. Organization

The Finnish study is organized as an independent research group under the directorship
of Professor Erik Allardt. It is not a direct part of any university institute and the
members are full-time researchers with no teaching duties.

In Sweden, Professor Sten Johansson was responsible for the study discussed here. As mentioned previously, the study was part of the work initiated by the Low Income Committee and the reports prepared by Johansson and his colleagues was submitted to this Committee. The Committee itself was a one man show, but it was advised by a group of three experts, one representing the National Federation of Trade Unions, one representing the Federation of Government Employees and the third member representing the Federation of Employers. The study had a secretariat to arrange and organize administrative tasks.

In Norway a group of three researchers under the directorship of Professor Tor Rødseth is responsible for the study, and a Consultative Committee with members from various ministries and labor market organizations was appointed to assist and advise. This Committee meets two or three times a year and there is communication between individual members and the study between meetings.

The study has established a secretariat of its own, consisting of a secretary, a programmer,
three research assistants and a project manager. A few researchers and a
number of students are attached to the study on a contract basis.

7. Problems of Research Policy

The Finnish study, being purely academic, has the advantage of being free of political
considerations. The project will, of course, consider the political consequences of its

work and there will always be a question of influencing the possibility of desired
policy effects, but it is perfectly free to discuss and recommend initiatives.

The fact that the other projects are obviously limited in this respect is illustrated by
the faith of the Swedish Low Income Committee and by the directives of the Norwegian
study which state that it must not itself consider political initiatives.

However, I do not believe this is a limitation that is felt to be a great problem while research is being undertaken. It is a question of accepting the terms once and for all. It is also far from obvious whether the political importance of a project depends on its ability to present specific proposals or not. In the Norwegian study there is a feeling that close contact with the political community through the Consultative Committee is in itself of great value to the research. The members of the Committee influence the project in a positive way by pointing out what problems on the basis of their experience should be considered most important. It also guarantees the study some attention and thus assures that the results are known by members of the political community.

There are still a number of practical and principal problems attached to projects of
this kind. In this discussion I shall not refer to the individual studies but mainly draw
on my own experience from the Norwegian study.

The most simple aspect of the most simple problem has to do with the size of the
study in the technical sense. A lot of people are involved and a lot of data will be
analyzed.

The fact that the studies have established ad hoc organizations of their own has the advantage of freeing participants from other duties they would have had if they worked at university institutes, and of bringing together a group of (young) enthusiastic people. But because of a lack of experience they are bound to have a few disheartening experiences during an initial 'learning period'.

The projects are large not only in the technical sense but also with regard to the field to be covered. In each aspect of the studies, substantial research has already been done, so this raises the problem of avoiding unnecessary repetition of previously completed work. Even in small scientific communities such as those in Scandinavia, this is not simple. In the Norwegian study we started by trying to map the existing state of knowledge, but it soon became apparent that this was not a job that could be completed. We simply had to learn to live in some frustration.

The size of the task creates problems of a more principal and serious nature. The very idea of these mammoth projects is to some extent self-contradictory. All the objectives cannot possibly be fulfilled. In the Norwegian study, for example, one objective is that insight is given into the situation in the population for each of the components listed in the directives. On the other hand, a total perspective is to be preserved so that clustering of advantage or disadvantage is made apparent. It is very difficult to achieve this. The question of clustering is theoretically complicated and needs a more sophisticated model than the one actually developed. Different preferences will have to be allowed for so that the dispersion on welfare indicators explained by different preferences is isolated from the remaining dispersion.

The aim of the studies is not simply to describe welfare but to explain inequality. Because so much territory is to be covered it is not possible to go into each sector in depth. This is not only difficult for practical reasons but the whole purpose of the studies is, so to speak, to give large overviews, an attempt which is not instrumental in relation to the objective of presenting explanations, since explaining implies depth.

The somewhat self-contradictory aspect of the studies is further illustrated in another non-complementary aspect of the Norwegian study. It is to give information on both the entire population and special groups, most notably low income groups. This is problematic because going into the situation of special underprivileged groups is very different from studying the population as a whole. The poor are different not only in

that they have less money, they also have a different culture. For a number of groups it would be disrespectful to treat them simply as 'problems'. Let me take an example. The Norwegian study was urged (by non-Lapps) to look into the situation of the Lappish minority, but this was rejected mainly on the grounds that the study did not have the necessary competence, i.e. knowledge of Lappish culture and way of life.

There are also problems of a more practical nature with the total description/explanation and groups/population dilemmas. Causal analysis and analysis of special groups require large samples. The Norwegian study has a sample of 4000 individuals in its own survey. For multivariate analysis this is small. The same is true for analyzing special groups. The study is directed at the underprivileged, who are, relatively speaking, often small groups. Emphasis on the entire population, then, to some extent undermines the study's ability to provide information about the underprivileged.

In concluding this overview I shall make a few remarks relevant to the Swedish and Norwegian studies, but not to the Finnish one, again because of its status as an independent academic project. These two projects are research on the underdog for the topdog. There is not too much to be said about this question. Again, for the researchers, it is a question of accepting the terms, an acceptance which implies considerable optimism with regard to the political establishment's willingness and ability to solve the problems of the underprivileged. The Swedish study attempted to attack this problem by including a survey of political resources. The publishing policy is probably of equal importance. To the extent possible, knowledge should be made available to groups needing it to improve their ability to solve their own problems or influence politicians to do so. In this respect, the researchers have a responsibility. They are the ones who actually pry into the private lives of citizens.

The Level of Living Study, Bergen

Stein Ringen

---

NOTES

1. S. Rowntree and G. R. Lavers. Poverty and the Welfare State. London, Longmans, 1951.

2. B. Abel-Smith and P. Townsend. The Poor and the Poorest. London, G. Bell and Sons Ltd., 1965.

3. P. Townsend. 'Measuring poverty', British Journal of Sociology, 1954: p. 131.

4. J. R. Galbraith. The Affluent Society. London, Penguin Books Ltd., 1962, p. 261.

5. For full discussions on the concept of 'poverty', see, e.g., K. Coats and Richard Silburn, Poverty: The Forgotten Englishmen. London, Penguin Books Ltd., 1970; or D. Jackson, Poverty. London, Macmillan, 1972.

6. M. Harrington. The Other America. London, Macmillan, 1962.

7. On social indicators, see, e.g. R. A. Bauer (ed.), Social Indicators. Cambridge, Mass., The M.I.T. Press, 1966.

8. A presentation of the Swedish study may be found in S. Johansson, 'The Level of Living Survey: A presentation', Acta Sociologica 16 (3), 1973. The theoretical background for the Finnish study is discussed in E. Allardt and H. Uusitalo, 'Dimensions of welfare in a comparative study of the Scandinavian societies', Scandinavian Political Studies, Vol. 7/72. Oslo, Universitetsforlaget, 1972. A description of the Norwegian study is presented in S. Ringen, 'An introduction to the Level of Living Study' (Mimeo), Bergen, 1973.

9. E. Allardt and H. Uusitalo, op. cit.

10. S. Johansson. Om levnadsnivåundersokningen. Stockholm, Allmanna Forlaget, 1971.

11. K. A. Mork. 'Draft of a theoretical total model for the study', (Mimeo), Bergen, 1973.

12. J. S. Coleman. Resources for Social Change. New York, Wiley & Sons, 1971.

13. N. R. Ramsøy. 'The Norwegian occupational life history study'. Oslo, Institute of Applied Social Research, 1973.