Resumé

This report deals with the influence of various organizational resources upon
methods of participation. We found it useful to subdivide the concept of political
participation into four groups: Input participation, decision participation, consumer
participation, and respondent participation. Using data from one city and
one sector we cannot suggest any general conclusions. Our findings show these
various methods of participation to be highly intercorrelated and that the various
methods of participation - along with organizational resources as income and
staff, the contextual variable, and preoccupation with the local system level - influenced
each method of participation. This means that we have to take the positions
of organizations in the political system into consideration if we want to
study their participation in the political processes. The resources stemning from
positions in the political system are perhaps more influential upon the method of
participation than the traditional organizational resource variables.

The process of consuming public finances is not without political effects. Consumer participation is related both to decision participation and to input participation. Our study also indicates that it is wrong to assume closed models with no feed-back loops when we study relations betwe3n interest organizations and public authorities.

1. Voluntary Organizations as Political Actors

From the beginning, election studies have been concerned with why some individuals participate in political processes and others not. It may be fruitful to use some of the ideas developed in election studies to describe why some organizations participate and others not. Organizations will participate in political processes if they perceive it advantageous to their aims. Some will participate on the local level, others at the national level. To appreciate this variation, we should discuss to what extent the actual level is able to realize the organization's aims by the means available at that system level, that is, those arenas the level controls and the extent to which control is autonomous at this level. Generally, we believe that organizations will concentrate their efforts at the system level which

controls those means that are necessary for the organization's goal attainment. The picture we can construct of the localization of the means is not a static one, and of course an organization can be preoccupied with what is happening on more than one system level.

In Norway, especially the large economic organizations participate at the national level. These are also called 'hard-value' organizations (haird values are economic values). Local studies indicate that it is 'soft value' organizations that concentrate at the local level. The main purpose of this article is to discuss organizational resources in relation to different forms of participation in local political processes.¹ The normal way of analyzing political activity is to treat it as a unidimensional concept. Here we shall propose a division of the concept of political participation into four components as a framework for discussing organizational resources. All the organizations included in this study operate within the social sector and all must be said to be 'soft value' organizations. They all have a local site; even so, they may not be interested in politics at this level. Consequently we must consider their preoccupation with what is going on at the local level when we discuss their 'behaviour'.

The normal way of discussing participation is to see it as an activity requiring resources. One expects organizations with resources of employed staff and money to participate more and at more stages in the decision-making process than those that lack these resources. A British study (Newton 1976, 1) sums up how some resources can influence political activity in this way: 'The evidence supports the suggestion that the membership size is an important determinant of income which in turn determines the ability to employ full-time, salaried administrators, and that membership size, income and staff are all closely related to rates of political activity'. The dependent variable, 'political activity', was defined in terms of'the number of issues it (the organization) had been involved in and thus to produce a simple measure of frequency of political activity' (Newton 1976, 41).

One may ask whether it is fruitful and possible to subdivide the concept of political activity, or more especially, political participation. If we discuss political participation in relation to a political system, we can immediately identify 3 main categories: One is linked to the input processes, Input Participation, one is linked to the formal and authoritative decision processes, Decision Participation, and one to the consumption processes, Consumer Participation. In all these methods of participation it is also possible to have a Respondent Participation, that is, a response to demands from authorities.

Input Participation. Organizations are active in the sense that they put forward proposals on the input side of the political system. Here we find claims for changes in rules, for a larger share, and so on. This is the way usually suggested to understand how 'interest' organizations further their own interests. The concept of pressure group activity' covers an activity which is part of Input Participation. But the latter could also merely be ideas put forward by one actor. All kinds of activity may influence the 'input' stream, but here we narrow the concept to refer only to activity that is directed deliberately towards public authorities. In this pilot study we have defined this concept operationally to be: 'The number of proposals an organization claims to have put forward to any local authority in the social sector last year'.²

Decision participation is a continuous method of participation, by holding a formal position in an authoritative decision-making body. In this study we have operationally defined this concept to be: 'If an organization claims to have one or more representatives on any board linked to the local political system'. What we have witnessed at the national level is that the larger organizations have increasingly become decision-participants. We do not know whether this has led to an increase or decrease in other ways of participation. But how this participation might influence the other methods of participation is a question we shall discuss in this article.

Consumer Participation is the leasit political. Easton (1957) defines the concept of political activity as 'the activity that aims at influencing the authoritative distribution of goods in the society'. According to Easton's definition, participation in consuming goods which are present as a result of authoritative decisions is not political participation. But more recently, political scientists have asked how policy-outputs influence inputs (e.g. Ranney 1968). Here we shall not try to review this literature, but only suggest that just as it is natural to ask how outputs influence inputs, it is also natural to ask how participation in consuming values 'created' directly as a result of political decisions affects input participation. In the social sector many voluntary organizations themselves, or in cooperation with the authorities, run social institutions. This is mainly possible because of public money both from the national and the local level. This sometimes symbiotic relationship has been described as follows (Slack 1968): 'Voluntary effort as a whole was described in 1959 as having become an integral part of the health and welfare services. It became an integral part by virtue on the one hand of being first in the field, thereby gaining public acclaim, financial support, specialist knowledge and organizational

experience and sometimes, it must be added, vested interest and unwillingness to give up what statutory bodies in due course becaine better equipped to do. In addition, over recent years, there has been increasing support, both moral and financial, by government, departments and local authorities in respect of the statutory social services. That is to say they have frequently become agents for central or local authorities' (Slack 1968, 152-3). The same tendencies can also be seen in Norway: voluntary organizations run social institutions largely funded by public money. This could be said to have two main aspects. One is the dependence on public finances and the other is the professional concern with running social institutions. The former aspect is included in our consumer participant concept, the latter we have named the contextual variable. We have defined consumer participation operationally to be: 'Whether the organizations have received money from local authorities or not.'

To be a respondent participant you have first to be asked. The process of responding says something about the relations between authorities and organizations which we do not see in the other methods. Part of this concept is the 'remiss' institution, where the largest organizations are asked to give their views on important questions before the government puts forward a proposal to the national assembly: also relevant are informal 'hearings'. While reminiscent of decision participation, this kind of participation is less formal and so less binding on both the authorities and the organizations. It may be defined as follows: 'an organization responsively participates if it responds to demands/questions/proposals from the official authorities'.

This way of participation is different from pressure group activity in the sense that the organization must be asked first. It would be possible and perhaps fruitful to classify public sectors according to whether the organizations and the authorities both act (take initiatives) and react (respond to initiatives), whether one acts whilst the other does not respond, or whether one only responds and does not act.³ Here, however, we shall be concerned mainly with the kind of factors that affect this way of participating. Here we have operationally defined the concept as: 'An organization shows respondent activity if it says that it has been contacted by the local authorities'. A weak point is that we have no data on whether they have answered or not. But the authorities would not contact an organization if past experience was that it did not answer; and and indeed in this participation the crucial point is to be contacted by the authorities.

2. 'Organizational' Variables

Research with organizations as units sihows correlations between organizational resources and political activity. (See Newton 1976, 61.) In this study we shall discuss how organizational variables might affect the different methods of participation. We have data on the following commonly accepted organizational resource variables: membership size, income, employed staff. Since we have reason to expect that these are highly correlated, we shall resist the temptation to incorporate them in a single model. In fact, we are not especially interested in making a composite variable either, since we believe that different resources could influence different methods of participation. We have also added one 'resource variable', namely the proportion of the organizational membership which is said to be active in the organization's work. An organization which lacks staff resources could perhaps compensate by having more active members.

Preoccupation with the actual system level. Organizations will concentrate their efforts towards that system level which is relevant to their aims. We asked the leaders which level was most important to their organization. It is not only a question of which level is of relevance to organizations, but also how important that level is for their goal attainment. We asked the leaders how preoccupied their organization was with what was going on at the local political level. We will use the answers to this question as an indicator of how important thai level is for the organization's goal attainment. Through these two variables we are able to decide the system level to which the organizations are most attached, and how preoccupied they are with what is happening at that level.

Contextual variable. It is of course relevant to ask in what context an organization operates. To find one variable that could contain all the different contexts is of course not possible. But in this connection we are interested in seeing the kind of effect active social work might have on methods of participation. As an indicator of active social work we asked how many persons were employed by the organizations in social work: 54% had nobody employed in that work, whilst the remainder had from 1 to 20 persons employed in institutions which were run solely by the organization or in cooperation with local authorities.

Interest groups - cause groups. It is common to distinguish between 'Interest groups' and 'Cause groups', or between organizations that involve people in their occupational capacity and those that do not. According to this distinction all organizations included in this report are 'cause' organizations; none involve people in their occupational capacity. But in

the social sector we could distinguish between organizations that work mainly for people other than their members, and those that work mainly for their members. This grouping is somewhat similar to that between interest organizations and cause organizations. The former can be said to work directly and primarily for the members' benefits, whilst the cause organizations might well work for others. The leaders were asked about the objectives of the organization, to say whether the members of the organization joined mainly to take care of their own interests or to take care ofothers' interests. 64% of the organizations were classified working mainly for their own members' interests or both kinds, whilst 36% were classified as working mainly for others' interests. The former group we call interest organizations and the latter cause organizations. During the collection of data we became aware of how different the interest organizations were from the cause organizations. The former seemed to be less preoccupied with the local political level. Handicap organizations (organizing people with common physical or psychical handicaps) are almost all of this kind. And obviously their aims cannot depend too much on local authorities for their fulfilment.

Table 1 shows how the categorization of the organizations influences
system preoccupation. It is dangerous to make any inferences on how
various organizational characteristics affect participation and relations to

authorities without explicitly controlling for preoccupation with what is happening at the actual level and thereby indirectly for how dependent an organization is upon public decisions in order to reach its goals. Let us have a closer look at some differences between organizations working for their members and organizations which say they mainly work for others.

Table 2 shows a marked contrast between organizations working mainly for their members ('lnterest Organizations') and organizations working for others ('Cause Organizations). Two out of three cause organizations have persons employed in social work compared to only one out often of the interest organizations. The cause organizations in this local setting work for others by running social institutions. We must also see if they differ with respect to organizational resources.

Table 3 shows how interest organizations differ from cause organizations with respect to organizational resources. Cause organizations have higher income and more members than the interest organizations. All these findings underline the necessity of controlling for system level preoccupation, activity in social work, and organizational resources if we wish to discuss the effects of working for the members' interests or for others in the social sector.

Satisfaction with the relations to the authorities. One further approach is to argue that dissatisfaction creates activity. We asked the organizational leaders how satisfied they were with the relations they had with the local social authorities.

In Table 4 we have simply dichotomised between those who answered
very pleased and those who answered pleased and bad relations. This
variable could affect the style of participation.

3. Hypotheses

Although we have very few cases on which to apply all these variables, we
shall try to offer some assumptions about how these independent variables
will operate in the social sector of one specific city.

3.1 What Could Influence 'Input Participation'?

We expect preoccupation with what is going on at the local level to be positively correlated with input participation. Organizations will concentrate their efforts towards the system level which is seen as relevant to their aims. And further, input participation will increase with increasing preoccupation with what is happening at this level.

HI. Preoccupation with the local level will be positively correlated with
input participation at that level.

Organizational resources and participation

In the sense that input participation corresponds to the concept of political activity, we have at least two schools of thought. One opinion is that smaller groups are likely to be more politically active than large ones. 'In a variety of institutions, public and private, national and local, "action taking" groups and sub-groups tended to be much smaller than "non-action taking" groups and subgroups.' (Olson 1974, 53). Olson supports this claim by arguing that smaller groups are more effective: 'For the small, privileged group can expect that its collective needs will probably be met one way or another, and the fairly small (or intermediate) group has a fair chance that voluntary action will solve its collective problems, but the large, latent group cannot act in accordance with its common interests so long as the members of the group are free to further their individual interests.' (Olson 1974, 53: See also James 1951) Alternatively, it can be argued that larger groups will be more active, 'because the greater the number of members, the greater the resources of voluntary labour, the greater the likelihood of public authorities taking action which affects the interest of some members, and the greater the likelihood of organizations having a critical mass of members who will want to react'. (Newton 1976, 43).

To test these two hypotheses it is necessary to hold constant the degree to which one organization is dependent upon public decisions at the given level. If we do this, and then compare organizations which are equal with respect to system dependence, but have unequal size, we could say:

(1) Organizational resources as employed staff could reduce the costs
linked to participation,

(2) Leaders in larger organizations are probably more dependent on taking
home points' for public appearance because of their distance from

the members. This will inspire leaders of larger organizations to polirical
action. Input participation by putting forward proposals is often
visible action.

(3) Smaller groups often have a more open and democratic way of working. This represents a cost with respect to initiatives related to input participation. Larger organizations have to a greater extent delegated authority to their leaders to take initiatives. This will reduce the costs linked to initiatives related to input participation.

(4) To the extent that input participation is linked with dissatisfaction with present output from the system, we would probably find small groups more active. Because smaller groups are poorer and weaker, they will probably have a higher propensity to be dissatisfied than rich and wealthy groups.

We do not have data here on whether input participation is of a protest nature or not. But if our expectations are not supported, the degree of dissatisfaction in the input streams could represent an alternative hypothesis to the assumptions made here.

H2: 'When we control for system level preoccupation, staff and income,
we expect membership size to be positively correlated with input
participation.'

Contextual variables

Normally, to be actively involved with social work in cooperation with the authorities would lead to organizational experience and thereby reduce the costs linked to making proposals. On the other hand, running socml institutions indicates good contacts with the authorities and this could lead to passivity in relation to input participation: the cases are dealt with during daily contact and there is no need to put up proposals. But if they would like to do so, it would not cost much because of their good contacts with the authorities. It could also be agreed that active social work in running social institutions would direct organizational attention away from politics. But since this social work is heavily dependent upon political decisions for money, we do not think it would.

It is of course dangerous to go too far with such a discussion with only 28 cases for analysis. However, within this limitation, it might be profitable to see how far this variable is positively correlated with input participation, especially because this kind of work is closely related to public financing.

H3: We expect activity in running social institutions to be positively
correlated with input participation.

Interest groups/Cause groups

As we already have seen, the organizations in our sample which are classified as interest groups (working for their own members) are much less concerned with what is going on at a local level than the cause organizations. For this reason we expect interest groups in this sample to be negatively correlated with input participation.

H4: In the local social sector interest groups are negatively correlated
with input participation.

If we control for system preoccupation, and other factors which are correlated with input participation, we would expect 'working for own interest' to be positively correlated with input participation. It is perhaps more stimulating to work for your own interests than for others.

Methods of participation as independent variables

To have representatives in decision making bodies would indeed reduce
the costs linked to making proposals.

H5: We assume organizations which are represented in decision making
bodies to show a higher inputpanicipation than those which are not.

The relationships with the authorities

Until now we have argued that input participation would be influenced by organizational resources. But another way to view the problem could be to see input participation as a result of dissatisfaction with what the system creates. As organizational theory on organizational problem-solving finds dissatisfaction with the present situation as a point of departure, we could assume that these organizations submitted proposals to the authorities especially when they were not satisfied with what the system produced. If this is the case we would assume:

H6: When we control for organizational resources and system preoccupations
we expect less satisfactory relationships with the authorities
to be positively correlated with input participation.

3.2 What Could Influence 'Consumer Participation'?

We assume preoccupation with what is going on at this level to be positively
correlated with consumer participation. Of the organizational resources,

we believe the organization's income to be positively correlated with consumer participation. For organizations, high income also means high expenses. And high expenses are often financed by public funds. Accordingly, we must control for the contextual variable.

H7: When we control for system preoccupation and running social institutions,
we expect organizational income to be positively correlated
with consumer participation.

If it is the case that to work for other people in the social sector means
running social institutions, interest groups will be negatively correlated
with consumer participation.

H8: When we control for system preoccupation and running social institutions,
we expect interest organizations to be negatively correlated
with consumer participation.

We believe that it could be favourable for an organization to have representatives
in the decision making bodies if it wishes to 'consumer participate'.

H9: Having representatives in decision making bodies is positively correlated
with consumer participation.

3.3 What Could Influence Decision Participation?

When we control for system level preoccupation, what kind of organizational resources could be positively correlated with having representatives in the decision making bodies at local level? First, membership size could be positively correlated with decision participation. In politics numbers count, and direct representation of an organization which represents many people would be more politically significant than that of one with few people. At least this would normally be the case, but sometimes small organizations possess a strategic position in the political system and could be selected to join decision bodies for this reason.

H10: Membership size is positively correlated with decision participation.

Another variable that we assume is correlated with decision participation
is the contextual variable of being involved in running social institutions.

These organizations would have specialized knowledge and contact with
what was happening in the sector. For this reason we assume:

Hll: Running social institutions is positively correlated with decisionparticipation.

Input participation and decision participation

Input participation over a longer period could result in decision participation. By being active and constructive in input participation, an organization could show that it deserves a voice in the relevant decision making body.

Hl2: Input participation and decision participation are positively correlated.

Interest groups and decision participation

The interest groups in our sample were less preoccupied with what is going
on at local level. This indicates that they are negatively correlated with
decision participation.

Hl3: Interest organizations are negatively correlated with decision participation.

Consumer participation and decision participation

If you consume public goods in a particular field, that would indicate your
interest in what is happening in that field. We assume consumer participation
to influence decision participation.

Hl4: Consumer participation and decision participation are positively
correlated.

We have already suggested that decision participation could be positively
correlated with consumer participation. This might, then, be a two-way
relationship:

Consumption influences the chances of being a decision participant,
and being a decision participant influences the chances of
becoming a consumer participant.

We shall see later if we can trace any such tendencies.

3.4 What Could Influence a Respondent Participation?

For the authorities time is a scarce resource. They therefore only contact
organizations if the latter are 'powerful' and if the contact can be made
without too much loss of time.

Organizational resources and respondent participation

Of the organizational resources, we believe that to have employed staff is of importance. Employed staff would drastically reduce the costs for the authorities to make a contact. This is also so because membership size and staff are positively correlated. The Pearson's r between these two variables is 0.61 (s = 0,001).

Hl5: Having employed staffin the organizations and being contacted by
the authorities are positively correlated.

Social work and respondent participation

To be involved in social work would increase the chances that the authorities
would have reason to make a contact, and would probably reduce the
costs linked to making the contact for the authorities.

Hl6: To be involved in running social institutions and respondent participation
are positively correlated.

Decision participation and respondent participation

To have representatives in the decision-making bodies would make it easier for the authorities to make contact, but it would probably be less necessary for them to do so because of the continuous contact in the decision-making body. For this reason we do not expect to find any significant correlation between decision participation and respondent participation.

Hl7: Respondent participation and decision participation are not significantly
correlated.

Interest groups and respondent participation

Because of the lesser preoccupation of 'interest groups' with what is happening at local level and because of their lack of staff, we expect interest organizations and respondent participation to be negatively correlated.

H18: Interest groups are negatively correlated with respondent participation.

4. Analysis

We will begin the analysis by testing the assumptions we have made. Then we will discuss models that could explain some of the variances in the dependent variables. Let us start the test of our assumptions by first showing some simple correlations.

Let us discuss Table 5 step by step in relation to our assumptions before
we trace more general patterns.

Input participation

Table 5 supports our assumption with regard to a positive correlation between system occupation and input participation, (r = 0,59). We expected further membership size to be positively correlated with input participation. Table 5 does not support this assumption. A partial correlation between membership size and input participation, controlling for the other organizational resource factors and system preoccupation, shows no connection between membership size and input participation. Table 5

supports our assumption of a positive correlation between running social institutions and input participation. The hypothesis still holds when we control for the other factors which are positively correlated with input participation. The assumption of a negative correlation between interest organizations and input participation is supported, as is that of a positive correlation between decision participation and input participation. The assumption is still supported when the other influential variables are controlled in a partial correlation. We assumed that those less satisfied with their relations with the authorities would be more active in input participation than those who were satisfied. Table 5 shows this is not the case. On the contrary, it shows a significant positive correlation between being very pleased with the relations with the authorities and input participation. The process of problem solving described in organizational theory literature does not seem to be similar to the process of input participation. Rather, it could be the case that previous success in putting forward proposals creates good feelings towards the authorities and inspires further input activity. There is also a negative correlation between the proportion of members being active and input participation, but this correlation disappears when controlling for the other intervening variables. We shall later consider these variables in a regression model. At present we can conclude that neither of the organizational resource factors are correlated with input participation. The latter seems rather to be a function of the contextual variables: system preoccupation, good relations with the authorities, and decision participation.

Consumer participation

As we expected, organizational income is positively correlated with consumer participation. This is so even if we control for other intervening variables in a partial correlation. The organizations which are created mainly to protect their own members' interests are negatively correlated with consumer participation in the local sector. The assumption made about a positive correlation between consumer participation and decision participation is also supported by Table 5. As we might have expected, we also have a positive correlation between involvement in running social institutions and consumer participation. This is not surprising since this involvement is heavily financed from public funds. Secondly, Table 5 also shows that it is possible to be a consumer participant without being too concerned with what is going on at the local political level. Consumer participation seems to be a function of organizational income (expenses), involvement in running social institutions, and decision participation.

Decision participation

We assumed membership size to be positively correlated with decision participation. Table 5 shows this is not the case. If numbers count in politics, this does not seem to be the case in the local social sector with regard to having representatives in decision making bodies at the local social level. One reason for this could be that these arenas are not seen as 'political' - in the sense of a local political struggle on these questions. In this case more professional considerations could be of importance when representatives are chosen. If so, we would expect our next assumption to be supported: (Hll)

Running social institutions is positively correlated with decision
participation.

Table 5 supports that assumption. There is a positive correlation between
decision participation and the contextual variable.

As we have already seen, there is a positive correlation between input participation and decision participation. We asssumed interest organizations to be negatively correlated with decision participation, but Table 5 does not support this assumption. The main explanation for this is that the authorities in our city instituted a special board for the handicap organizations. Their relations with the authorities are channelled mainly through this board. If we exclude this board from consideration, r 'improves' to r = -0,24, but the level of significance is only 10%. As we already have seen, the assumption about a positive correlation between consumer participation and decision participation is supported.

Decision participation seems to be a function of contextual circumstances, consumer participation, and input participation. The contextual variable and consumer participation are highly correlated, but we shall discuss that when we use regression analysis.

5. Four Closed Systems

The analysis clearly shows that several of our variables are closely connected. This indicates models with feedback loops between the various variables. But let us assume that we have four different systems, one for each method of participation - and see if our tendencies are persistent before we put the pieces together again.

What seems to influence input participation?

There was a positive correlation between input participation and system preoccupation, running social institutions, cause organizations, decision participation, and satisfaction with local authorities. If we assume a closed system with these as independent variables, we can make a regression equation. In a regression analysis containing all these variables, only two turn out in the end to be significantly and positively correlated with the dependent variable system preoccupation and decision participation.

A regression equation with these variables gives:

Y = The number of proposals put forward = input participation.

X = particularly preoccupied with what is happening at local political level
(dummy)

Z = having representatives in a decision making body at local level
(dummy) = decision participation

The result of this regression analysis was:

When we standardize the beta coefficients we have the following model:

These results indicate that system preoccupation and having representatives in a local decision-making body are together able to explain 439?' of the variance in the dependent variable: system preoccupation is the more important.

The regression analysis 'throws out' the contextual variable, and the relations with the authorities as insignificant in this model. Our 'relations with the authorities' variable was not a good one: 83% answered that they were pleased with the relationship or very pleased with the relationship. The few cases that we have make us hesitant of drawing definite conclusions on these points.

What seems to influence consumer participation?

We found a positive correlation between consumer participation and organizational income. If we see organizational income as an indicator of organizational expenses, we have a positive correlation between organizational expenses in the social sector and consumer participation.

We also have a positive correlation between having representatives in
decision-making bodies and consumer participation. This gives us a regression
equation with:

Y = consumption of local public money (dummy)

X = organizational income (expenses). The incomes are grouped in
categories ranging from 0 to 12

Z = Having representatives in decision-making bodies within the
local political system (decision participation) (dummy)

Having a dummy variable as a dependent variable we are most interested
in the effects of the independent variables on the dependent one. The
result of this regression analysis was:

Level of significance: 0,05%

When we standardise the beta coefficients we have the following model:

These results indicate that the model is able to 'explain' 34% of the variance in the dependent variable. Of the two independent variables, the standardized beta coefficients show that organizational income is the more important.

What seems to influence decision participation?

We have found positive correlations between decision participation and
the contextual variable: running social institutions, input participation,
and consumer participation.

We will again assume a closed system with these variables as independent ones and make a regression equation in this instance. The contextual variable drops out as insignificant, and we are left with input participation and consumer participation as significant independent variables.

Y = having at least one representative in a local decision-making
body (dummy)

X = number of proposals put forward = input participation

Z = having received money from local authorities = consumer participation
(dummy)

The result of this regression analysis was:

When we standardize the beta coefficients we have the following model:

These results indicate that decision participation is influenced directly by consumer participation and indirectly by consumer participation through input participation, and that input participation influences decision participation. Of these, consumer participation is the most influential variable. The level of significance is not very good, but the results seem reasonable. However, the model 'explains' only 27% of the variance in the dependent variable.

What seems to influence respondent participation?

We found positive correlations between respondent participation (having been contacted by the authorities) and having employed staff, involvement in running social institutions, and being a cause organization. Because the cause organizations were so highly correlated with running social institutions we experienced multicolinearity in our first regression model. The cause organizations were so highly correlated with both the

contextual variable and the organizational resource variables, that a result with this factor would be hard to interpret. Hence we removed that variable from our regression equation. The remaining variables turned out to be significant in our regression equations.

Y = having been contacted by local authorities (dummy) = respondent
participation

X = number of persons engaged in social work for the organization
= contextual variable.

Z = having persons employed in staff positions in the organization
(dummy) = organizational resource variable.

Again we have assumed a closed system with these two variables operating
as independent ones.

The result of the regression equation was:

When we standardize the beta coefficients we have the following model:

The model shows that the variables having employed staff and the number of persons employed in social work for the organization together 'explain' 44% of the variance in the dependent variable. Having staff is the most influential variable, with both a strong direct effect on the dependent one, and an indirect effect through the running of social institutions.

6. From Closed Systems to Open and Dynamic Systems with Feedback Loops

We have seen that it is artificial to assume closed and unrelated systems
when in fact they are open and intercorrelated.

Our three decision variables which are linked directly to a model of a political system, input participation, decision participation, consumer participation turn out in this analysis to be intercorrelated in one dynamic model instead of three static and closed ones. And respondent participation seems to be different, and not so closely connected with the three other methods of participation.

A review of our regression models gives in fact this result:

We have already commented upon the respondent participation model. The multiple participation model shows that the main variables influencing the method of participation are the other methods of participation, system preoccupation, and organizational income. The two systems are probably connected; having employed staff and organizational income are highly correlated.

When we study the participation in political processes, it is not enough to allow for system preoccupation, organizational resources variables, and contextual variables. We must also take into account the organizations' various positions in the political system itself. The resources stemming from positions in the political system are influential variables operating on each method of participation. The consumption of public finances has political consequences, influencing both input participation and decision participation. Accordingly, we must not exclude the consumption of public goods when we study the political behaviour of organizations.

These results have several consequences. The beta coefficients were
created with the assumption of the absence of feed-back loops. The results

show that this is not the case. This means that the beta coefficients in our regression models should have been estimated under the assumption of a dynamic system, with feed-back loops. But this also implies that it is wrong and dangerous to assume closed and none feed-back models when we discuss connections between interest organizations and public authorities. Participation in one part of the political system influences and is influenced by participation in other parts of that same system.

We chose the dependent variables from a general model of a political system. These data give empirical support for the usefulness of having a general concept of a political system - a model which links processes together again.

---

NOTES

1. An earlier version of this article was presented at European Consortium for Political Research Workshop on Interest Group Strategy, Freie Universität, Berlin 1977. The data are basedon interviews with the chairmen of 28 social organizations and the administrative leaders within the social sector in a city in southern Norway. I would like to thank Karl Erik Brofoss and Tore Hansen for valuable comments. I am also indebted to the Department of Government, University of Essex, where the work on this project has been carried out. My thanks also to Mike We ston for linguistic advice. This project is financed by the Norwegian Research Council for Science and the Humanities.

2. The year in question was 1972.

3. I have tried to describe this way of thinking in Sol vang 1973.

REFERENCES

Easton, D. 1957. 'An Approach to the Analysis of Political Systems', World Politics 9.

James, J. 1951. 'A Preliminary Study of the Size Determinant in Small Group Interaction',
American Sociological Review 16, 474-7.

Newton, K. 1976. Second City Politics. Oxford: Clarendon.

Olson, M. 1974. The Logic of Collective Action. Cambridge, Mass., Harvard University
Press.

Ranney, A. (ed.) 1968. Political Science and Public Policy.Chicago, Markham.

Slack, K. M. 1968. 'The Machinery of Administration,' in A.V.S. Lockhead (ed), A Reader
in Social Administration. London, Constable.

Solvang, B. K. 1973. 'Interesseorganisasjonenes politiske aktivitet på nasjonal og lokalplanet,'
Kristiansand, ADH skrifter, nr. 3