Inequality and Health

The Institute for the Research of the Kibbutz and the Cooperative Idea

Effects of Intra-Community Inequality in Kibbutzim on Their Members' Well-Being and Health

Uriel Leviatan

1. 2009                                                 No.   90

Effects of Intra-Community Inequality in Kibbutzim on Their Members' Well-Being and Health

Uriel Leviatan

Abstract

This study tests hypotheses about the relationships of social structures with the role of local leadership in sustaining them, the ways that social structures influence levels of intra-community socioeconomic inequality, and how both contribute to expressions of social capital, well-being and health. Samples from 32 kibbutz communities (n=700) were broken down into four groups – one group of 11 kibbutzim that are still traditional in their reliance on the principle of "qualitative equality", and three other groups based on the length of time since their adoption of "Differential Salary Arrangements" (DSA): 1-2 years, 3-4 years, and 5-6 years. Analyses are performed with the kibbutz as the unit of analysis. The major instrument of study was a questionnaire, and the variables are based on aggregated measures across all respondents from each kibbutz. A model based on suggestions made by Wilkinson and others was tested, with criticisms of the model also incorporated. The basic relationships tested and supported by the data in the current study are the following: local leadership is responsible for sustaining the social structure of remuneration (traditional or according to position at work or in the community); this structure of remuneration determines the level of socioeconomic inequality (more years since the adoption of DSA leads to more inequality), but is also directly detrimental for some of the physical (or material) expressions of social capital. Inequality and DSA determine the level of social capital both in its psychosocial and physical expressions, which in turn affects indicators of aggregated levels of well-being and health. Expressions of social capital also serve as mediating variables between DSA and socioeconomic inequality on the one hand, and well-being and health on the other. The discussion section emphasizes the major unique aspects and findings of this study: that effects of social structures come into display very shortly after their implementation; that the model is upheld with a sample of very small communities; that the model is supported even though intra- and inter-kibbutz levels of inequality are small compared to the outside society; and the central role of leadership.

Keywords: SES inequality; Psychosocial Social Capital; Physical Social Capital; Health; Wellbeing; Kibbutz Communities; Structural Social Differences; Differential Salaries
Introduction

This study responds to and tests certain (methodological as well as content) aspects of the criticism raised against the theoretical model proposed by Wilkinson and others that sees intra-community levels of socioeconomic inequality as related to aggregated expressions of health and well-being. I aim to test for (a) the existence of sociological structures that are responsible for socioeconomic inequality; (b) the role of leadership in shaping such social structures; (c) how such social structures directly affect physical/material expressions of social capital; and (d) the mediating role of physical/material expressions of social capital in addition to those psychosocial expressions.

The site of this study is kibbutz communities in Israel. It compares groups of kibbutzim that differ in their level of socioeconomic inequality as a result of change, which has taken place over a number of years, from communities based on the principle of "qualitative equality" into communities where the overriding principle of distribution is "equity". This differentiation and the resulting inequality among members are used to relate to expressions of social capital and eventually to indicators of well-being and health. The units of analysis are kibbutz communities, and most variables are aggregated responses of members sampled from each of the 32 kibbutzim that took part in the study.

Theory and background research

Over the past 15 years, a rich research literature has developed around the thesis that the degree of socioeconomic inequality within a social entity (such as countries, states, counties, metropolitan areas, provinces, neighborhoods) negatively determines average levels of well-being, morbidity, mortality, and life expectancy for the population of that social entity. Moreover, research has shown that levels of income inequality within a societal level are associated with mortality over and above the impact of average income level (Robert, 1999).

Studies on this relationship have been published frequently since Wilkinson's (1992) paper on the topic – either as reviews in their own right or as part of a debate about the validity of the purported relationship (e.g. Wilkinson, 1996; Kennedy et al., 1996; Lynch and Kaplan, 1997; Judge et al., 1998; Robert, 1999; Kawachi, 1999; Kawachi, et al., 1999; Marmot, 1999; Robert and House, 2000; Lynch et al., 2000; Mellor & Milyo, 2001; Marmot and Wilkinson, 2001; Schnittker and McLeod, 2005; Wilkinson and Pickett, 2006). Wilkinson & Pickett's (2006) contribution is particularly significant because it reviews all studies on the topic (to the time of their writing – 155 papers that included 168 analyses). They concluded that 87 studies wholly supported the thesis, 44 partially supported it, and 37 were unsupportive.

The general model that was initially suggested, put forward a causal flow that included expressions of social capital in its social psychological sense (as defined by Colman (1988) and Putnam (1993; 2000)) as mediating variables. The basic thesis regarding these relationships holds that the level of socioeconomic inequality within a social entity causes relative deprivation among those at the lower strata. Specifically, higher levels of socioeconomic inequality degrade the level of social capital. This is expressed in lower participation in community, public, and political spheres; in reduced trust and solidarity among members of the community; in lower levels of trust in leadership and the political process; in weaker social support offered within the community; and so on. In turn, these reduced levels of social capital affect risk behaviors and risk states that are manifested in higher levels of morbidity and mortality, and shorter life expectancy.

In addition, this theoretical model was suggested as being most appropriate for developed, industrial societies that are at the higher end of prosperity, while economic level matters much more for societies at lower levels of prosperity.

However, while most (though not all) studies support the relationship of economic inequality with health, several aspects of the role of other factors and the general causal relationship among the variables have been debated. Research to date has left important questions unanswered. The current study of Israeli kibbutz communities aims to answer some of these questions and to relate to some of the critiques raised by previous studies.

The first aspect that is debated is represented by those who argue that we should look "upstream" for larger and more sociological social and structural aspects of society if we wish to explain the causes of social inequality (House, 2001). This view does not see socioeconomic inequality as the "beginning of the causal chain". Instead, it sees it as representing structural and ideological aspects of society (such as the ideology of neo-liberalism, or the result of social class structures) that lead to socioeconomic inequality on the one hand and affect other social and economic structural aspects of society on the other (such as level of individual income or investment in health services), all of which lead to deteriorated social capital and reduced health and well-being (e.g. House, 2001; Coburn, 2000; Muntaner & Lynch, 1999). However, if central social structures are seen as "the beginning of the causal chain", this begs the question as to what determines them. I suggest that it is social entities' political leadership that is responsible for changes to and/or the preservation of central social structures.

To test for this alternative interpretation, all of its components should be incorporated within the design of the study. Moreover, the study should be longitudinal, and should allow for the interpretation of causal relationships and associations among variables when others are held constant.

This gives rise to the current study's first research question: do ideology and the basic social structures in society that emanate from it determine levels of socioeconomic inequality in that society while at the same time expressing social capital of both kinds (but particularly of the physical kind)? Is a given society's political and social leadership responsible for the ruling ideologies and the social structures that emanate from them?

The second aspect under debate is whether the causal flow runs from socioeconomic inequality, through degraded expressions of psychosocial social capital (PSSC), and then to higher average levels of morbidity and mortality, lower levels of well-being, and shorter life expectancy. Alternatively, maybe socioeconomic inequality operates through another major pathway: that is, it adversely affects levels of material resource investment by the society under study for the sake of fulfilling its individual members' needs (a process that its proponents term neo-materialism). The society thus invests less in upgrading and sustaining aspects of life that contribute to health, well-being, and survival (such as education, health services, job creation, ecology conservation, public transportation, and the like). A summary of this debate is captured in two papers: Lynch et al. (2000) for the latter approach and Marmot & Wilkinson in a response (2001) for the former. Since then, Kawachi et al. (2002; 2004) have suggested that the debate is somewhat artificial. They argue that the two approaches could be reconciled, given that socioeconomic inequality probably leads both to lower investments in material aspects such as health, education, ecology, job creation, level of the welfare system (as suggested by the neo-material approach; e.g. Kaplan et al., 1996; Lynch et al., 2000) and also to PSSC, as suggested by Marmot and Wilkinson (2001) and others. Kawachi (2004) notes that there is no sound research that can help us to choose one approach over the other, and that both seem to explain the disparity in average levels of health among social entities with differing levels of socioeconomic inequality. In the current study I refer to both aspects as two different kinds or dimensions of social capital.

Following Colman (1988) and Putnam (1993, 2000), PSSC has been instated as a characteristic of relationships within a social entity. It differs from human capital, which expresses individual characteristics such as level of education and job training, and from financial capital, which expresses the amounts of money individuals or organizations possess. As another expression of social capital (let us call it "physical social capital", or PHSC), I propose considering the decisions and investments made by social entities – via their leaderships or other institutions of decision-making – regarding their priorities for spending or investing economic resources for the benefit of their members – such as whether or not to invest in health services, education, the environment, or jobs creation. Since similar social entities with similar levels of financial resources might have totally different priorities for using them, and since those priorities are decided on the basis of political and social rather than economic considerations and benefit members of the social entity differentially, these priorities express certain characteristics of the social entity.

Specifically, the view of PHSC as resulting from economic inequality within a society is translated as follows: at high levels of socioeconomic inequality, society – or rather, its leadership – invests little in developing PHSC. Low investment in this kind of social capital does not affect all members of society to the same degree. It is likely that members at low socioeconomic levels will be more affected, because members of society who enjoy a high socioeconomic status also possess more private resources. Therefore, they can invest privately and compensate for the lack of public investment. For instance, they can afford private health services or quality private schooling for their children, private security arrangements against crime in their neighborhoods, private higher education, ecological solutions within their areas of residence, and so forth. These compensatory private investments are of course not available to the weaker members of society. However, not everything can be bought by private money; for instance, clean air or good major highways cannot be bought even by the very rich. As a result, the well-being, and perhaps the health of the entire society would, on average, be affected by socioeconomic inequality.

Empirical investigation is required to determine the relative importance of each of the two dimensions of social capital as mediating variables between socioeconomic inequality and expressions of health and well-being. There is a need, therefore, for a study that would include both the psychosocial and physical kinds of social capital as mediating variables.

Thus, the second group of research questions for the current study: are the two dimensions or kinds of social capital (PHSC and PSSC) affected by level of socioeconomic inequality? Do they act as mediating variables between socioeconomic inequality and expressions of health and well-being? Do they contribute independently and together to explain variance in health and well-being? What is their relative importance in that contribution?

The third aspect that the current study relates to, involves criticism of the validity of the very relationship between the degree of socioeconomic inequality within a social entity and expressions of health, mortality, and well-being. The second aspect deals with the proposed pathways – the mediating variables between inequality and health – suggested by the theory. The third aspect is twofold. First, it asks whether socioeconomic inequality should be viewed as the beginning of the causal chain or rather as just one expression of many more central societal and economic structures. Second, it deals with the question as to whether or not these structures translate ideologies that, on one hand lead to economic inequality, but on the other also lead to other societal structures, all of which are detrimental to health and well-being. I shortly summarize the main points of these debates.

Regarding the first aspect of debate, various critics (e.g. Judge et al., 1998; Catalano, 1998; Plat, 1998; Mellor & Milyo, 2001) have challenged the work carried out by Wilkinson and others on a number of methodological grounds. These have included sample size and selection (e.g. too small in studies comparing developed countries); a lack of clarity and uniformity in defining the social units for analysis; the use of GNP to compare countries rather than measures that represent net income per capita; no control for potentially confounding variables that might affect health (such as people's prosperity, the generosity of the welfare system, characteristics of social environment, level of healthcare expenditure, culture, history, persistent effects of epidemics and diseases, differences in lifestyle, health care markets, political systems); too much heterogeneity among the units of comparison in terms of size, economic structure, and years when measures of inequality and health were taken; no longitudinal designs.)

Some of these criticisms may be valid. Even though statistically significant findings that have been registered after failing to control for potentially confounding variables may sometimes be seen as adding support to a hypothesized relationship (unless an alternative explanation is suggested), it is methodologically sounder to control for them. However, some of the variables that critical researchers refer to as confounding are in effect viewed as mediating variables by the proponents of the theoretical model they criticize (e.g., the generosity of the welfare system is seen as a result of the level of inequality, since in less equal societies there would be less investment in the welfare system, as explained in the next section).

Judge et al. (1998) summarize by suggesting three important lessons: first, it is essential to use the most consistent measures of equalized disposable income that are available; second, serious attempts should be made to adjust for possible confounding variables; and third, the largest possible number of countries should be included. The current study endeavors to take these points into account.

This leads us to the current study's third research question: does the level of socioeconomic inequality translate into lower average levels of health and well-being?

These three research questions are investigated in a study recently conducted in kibbutz communities. Before describing it, though, I first present a short background of the characteristics of kibbutz communities that are relevant to this study.

Contemporary kibbutz communities as the research site

The current state of Israeli kibbutz communities (plural in Hebrew: kibbutzim) offers a unique opportunity to study the research questions summarized above. The following sections present the background that explains why this is so.

Principles of equality. Until the late 1980s, all kibbutzim were very similar to one another, particularly in terms of the socioeconomic equality among the members of each community. I refer to these kibbutzim – communities that followed principles of equality called "qualitative equality" – as the "traditional" kibbutzim. Individual members expected their communities to take care of their unique personal needs and expectations (within the resource capabilities of their kibbutz and in keeping with the norms of a relatively modest lifestyle), and the community expected all of its members to contribute all of their personal resources to the community. This principle of "qualitative equality" among members was unique in that individuals did not judge its successful fulfillment by comparing themselves to others, but rather in terms of their own needs and expectations. Therefore, while a society that exercises this principle of equality as its major distributive principle may appear to have extensive inequality among its members (because members have differing needs), there is actually a very high level of equality since members are able to satisfy their needs and expectations to similar levels.

"Qualitative equality" is different from other views of equality, such as "mechanical or arithmetic equality" (as first suggested by Aristotle), where everyone is treated the same way regardless of individuals' unique differences. With mechanical equality, people feel that they are treated fairly if they receive similar compensation to significant comparable others.

"Qualitative equality" also differs from the principle of "equity" (Homans, 1961; Adams, 1965). Here, people expect the ratio of the compensation they receive for their contribution (however that is measured) to be similarly applied as to significant comparable others.

The dominant mode of resource distribution in traditional kibbutzim was mostly based on the principle of "qualitative equality" (according to needs), as expressed in resources allocated to domains such as education (both for children and adults); job training; leisure pursuits; health services and medication; taking care of dependents; home maintenance; transportation needs; consumption of most food items; culture; and even free-for-use personal budgets (based on family structure and size). Where the criteria to define differential needs were too complex, or where need fulfillment required excessive resources, consumption items were allocated following the principle of "mechanical equality". This applied to domains such as housing, vacations, and clothing (from the 1970s on). It is also worth noting that, until the 1980s, not only were the budgets allocated for private consumption equal within a given kibbutz community but also across communities from same federation of kibbutzim. Recommended budgets per domain were sent from the federations' headquarters to each kibbutz, and surpluses of the richer kibbutzim were partially taxed by the federations. This money was used to help out poorer kibbutzim as well as for more general purposes – political and educational. However, while the principle of "mechanical equality" was deployed in some instances, the "equity" principle (based on social or occupational status) was not applied to any domain or resource distribution.

In addition, kibbutzim adhered to the values of the community's total and unconditional responsibility for each member and his or her dependents, full solidarity among members, a strong emphasis on democracy, mutual ownership of all property and means of production, and partnership in all domains of community life (Leviatan et al., 1998).

Expressions of social capital. Traditional kibbutzim translated the values and principles of "qualitative equality", solidarity, partnership, and democracy into social and structural arrangements. These social and structural arrangements were adjusted to create both kinds of "social capital" (Leviatan et al., 1981; Cohen-Mansfield & Leviatan, 1992; Leviatan, 1999) as resources conducive to the enhancement of kibbutz members' health and well-being.

For example, kibbutzim expressed the physical aspect of "social capital" through their commitment to create appropriate jobs for members at every age and for as long as they were willing to continue working (some kibbutz members have continued working even past the age of 90). Jobs were created that took into account the particular limitations of elderly members, and work arrangements were adapted to changing capabilities, for instance by reducing the number of working hours or the number of working days (Leviatan, 1983). Similarly, jobs and branches of production and services were either newly created or went through major changes to accommodate young members who had graduated from institutions of higher learning so that their newly acquired skills, interests, and knowledge could be expressed in their work. Resources were invested in the creation of appropriate possibilities for leisure activities; the material standard of living of the aged was kept equal to that of younger members, and their special financial needs were shouldered by their kibbutz. Kibbutz health institutions saw themselves as responsible for members' health, and therefore institutionalized preventive medicine, ambulatory institutions, and medical check-ups – all under the community's financial responsibility.

Along with its physical form, social capital was also constructed in the social sphere. This was expressed in an emphasis on the social integration of all members and the offer of social support by community institutions to those without family-based support, such as single or widowed members (Leviatan, 1988; 1999). The creation of "social capital" was exemplified in the intensive purposeful integration of all members – both young and old – in the civic life of their kibbutzim; in on-the-job training of young people for social and civic leadership positions in the community; and in an effort to involve all members in the community's cultural and social life.

Expressions of health and well-being. As a result of these social arrangements, the populations of the traditional kibbutzim demonstrated high levels of physical health and exceptional levels of well-being and longevity. Moreover, their members were a paragon of successful aging. There were many indicators of this phenomenon. For instance, death rates of members aged over 50 years were much lower than those for the Jewish population in Israel to the ratio of 1:2 or 2:3. Life expectancy of the kibbutz population was thus three to four years greater than that of Israel's Jewish population at large, and was also longer than in most other populations in industrial societies (Leviatan and Cohen, 1985; Leviatan et al., 1986; Leviatan, 1999). Illustrative data are shown in Table 1.

Moreover, kibbutz populations, and particularly their aging population, also enjoyed very positive levels of physical health and well-being in comparison with other populations. This was expressed in indices such as satisfaction with specific life domains, with kibbutz life, and with life in general (Tannebaum et al., 1974; Leviatan et al., 1981; Carmel et al; 1995; 1996; Leviatan, 1999). Research has supported the thesis that social arrangements positively impact on well-being. For instance, older members of kibbutzim with differing levels of supportive social arrangements, which were assumed to contribute to health and well-being, did indeed display the expected differences in relation to these outcomes (Leviatan et al., 1981; Leviatan, 1999). Similar findings were reported when kibbutzim were compared with other populations in Israel and abroad (e.g., Leviatan, 1988), and when the relative strength of personal characteristics and social arrangements were contrasted as predictors of members' well-being and health, with social arrangements found to be much more important (Leviatan et al., 1981).

Table 1: Life expectancy (LE) at birth and at age 50 for the permanent kibbutz population and Israeli Jews in 1977, 1984, and 1995 (by gender)^a

		LE at birth		LE at age 50
Gender	Year	Kibbutz	Israeli Jews	Kibbutz	Israeli Jews
Males	1977	74.4	71.9	28.3	25.7
	1984	76.7	73.5	29.6	26.5
	1995	78.1	75.9	30.8	28.3
Females	1977	79.0	75.4	31.0	28.0
	1984	81.3	77.1	33.4	29.2
	1995	82.5	79.8	33.8	31.2

^a Sources: Leviatan (2003); data for the Jewish population in Israel are taken from Statistical Abstracts of Israel, Central Bureau of Statistics (1979; 1986; 1997).

Changes started in the 1980s. This somewhat "rosy" picture of traditional kibbutz communities started to change in the mid-1980s following a major economic crisis in Israel as a whole (description and analysis of that economic crisis is beyond the scope of the current paper; see, for instance, Leviatan et al., 1998). Following the economic crisis, many kibbutzim looked for solutions externally, seeming to view the inferiority of the kibbutz system as a matter of fact. The economic crisis also revealed an undercurrent of ideological crises (e.g. Rosner & Getz; 1994; Leviatan et al., 1998), expressed in a wish by many members to abandon their uniqueness and become "like everyone else". In particular, they were called upon by some of their leaders to abandon the principle of qualitative inequality (Weber, 1992; Harel, 1993). As a result, many of the proposed solutions to the economic problems suggested restructuring kibbutz businesses based on the principles of the market economy and neo-liberal and individualistic ideology. Thus, kibbutzim began to create boards of directors to control industrial and other domains (instead of the kibbutz's general assembly). They also started phasing out managerial rotation among members, abandoning direct democracy in kibbutz businesses, and rejecting "qualitative equality" by privatizing public budgets. Ultimately, contribution and remuneration were linked. These ideas came to be known in kibbutz lingo as the "changes". They were first introduced in industry (as the major source of income production and the domain in most contact with the outside world) before being implemented in other production domains such as farming and tourism, then in community services, and finally in the social domain of community life.

The privatization of public budgets and Differential Salary Arrangements (DSA). The most significant structural changes in community structures, and those that are most relevant to our study, deal with the rejection of the principle of "qualitative equality". This principle was abandoned by transformed kibbutzim who adopted the "changes" in favor of the "privatization of public budgets", and, later on, "differential salary arrangements" (DSA) (in which remunerations are based on the individual's professional, social or managerial positions at work or in the community).

The "privatization of public budgets" means that a public budget, such as the food budget, or the budget for health services (that are not covered by national health insurance), which had previously been distributed according to individual needs (within the capabilities of the kibbutz), has been "privatized". In such kibbutzim, each member now receives a food or health budget that is equal to that of other members, regardless of the individual's unique needs or wishes. Many kibbutzim have similarly "privatized" public budgets for other consumption domains (such as vacations, home maintenance, extracurricular education for children and adults, higher education, health services, and more).

The principle of the privatization of public budgets negates the view of each individual as unique, and does away with the principle of "qualitative equality"". This also relieves kibbutz institutions of their responsibility to treat each individual member "according to his or her needs", and undermines the notion of unconditional solidarity among members. Thus, the distributive principle expressed through this process of privatization might be properly labeled "mechanical equality".

The second major topic of structural change involves the differential salaries that members receive from their kibbutz based on their position at work or in the community (known as Differential Salary Arrangements, or DSA). The distributive principle operating here is the "equity" principle of remuneration, which runs counter to the principles previously exercised by kibbutzim, by which personal or family consumption budgets were calculated to ensure a similar standard of living for all members, while taking into account the family's or the individual's unique needs ("qualitative equality").

The adoption of differential salaries usually followed the privatization of major public budgets and represented another step away from earlier definitions of equality in kibbutz society. Even the "mechanical" principle of equality is given up, and the ruling principle becomes "equity" (Homans, 1961; Adams, 1965).

DSAs are the strongest manifestation of the ideological changes undergone by kibbutzim, as they reflect the outright abandonment of communal values of equality and solidarity for values of individualism, market principles, and neo-liberal ideology. The clearest expression of the new social structure can of course be found in what initiated it in the first place, namely, a desire to implement the principle of "equity" in remuneration. However, the ideologies behind this principle should also turn kibbutzim into communities with much less solidarity, reduced concern for individual members, and less willingness to invest in, and give expression to, communal life.

It has been observed that these major changes in the social structure of kibbutzim were led by the local leadership in each kibbutz (Rosner & Getz, 1994). Indeed, this is also the group that gains most (economically) from the shift to DSA. A survey of salaries in kibbutzim that had adopted DSA put the range of salaries between 4,000 NIS per month to 40,000 NIS, a ratio of 1:10 (Cohen, 2006). (The newspaper that reported the survey also mentioned that a few top managers working in businesses outside the kibbutz were being paid salaries of more than 100,000 NIS per month). It is true that differences in net income are smaller as kibbutzim apply progressive community taxing on salaries, but even then the ratio remains between 1: 3 or 1:5 (since each kibbutz autonomously decides its internal taxation rates). It has also been observed that the ratio of highest to lowest salaries positively correlates with the number of years since the kibbutz adopted DSA, and that the level of communal taxation of the top salaries decreases with time. This is because those in leadership positions are also the big earners, and so it is in their individual interest to reduce levels of communal taxation. This takes time, however, for two main reasons: first, the decision to move to DSA was always taken (as required by law) by a majority of at least two thirds of members (who were promised lower differentiation in salaries); and second, at least during the first years of DSA, members in leadership positions are those who had been committed to equality in the past. Their usual excuse for changing their ideologies was that the "old ideology, while desirable, is no more feasible" (Harel, 1990). It is hardly surprising that it should take time for the deeply embedded commitment to values of equality and solidarity to be erased.

These changes began to be implemented in the early 1990s. At first, only a few kibbutzim made the transition, but then the number of kibbutz communities adopting DSA began to grow very quickly. Table 2 shows the percentages of kibbutzim that had privatized various domains and those that had adopted differential salaries as their system of remuneration between 1990 and 2003.

Table 2 shows the rapid increase in the percentage of kibbutzim that privatized important public budgets (food, enrichment classes for children, higher education and health services). In addition, it records a steep increase in the percentage of kibbutzim that award differential remuneration for work positions. DSA had been implemented in 57% of all kibbutzim in 2003 when our study was initiated, and in about 71% of all kibbutzim by the end of 2007. Since no kibbutz has ever reversed the "changes", the end result is that a federation of homogeneous communities became, in effect, a federation of two dissimilar kinds of communities divided by the principle of equality they adopt: the "traditional" kibbutz with its commitment to "qualitative equality", and the "differential" kibbutz with its adoption of "equity" as its definition of equality. These two differing groups of kibbutzim serve as the basis for the research described below.

Table 2: Percentage of kibbutzim reporting the adoption of structural privatization or differential salaries (1990-2003)^a

DOMAIN // YEAR (19.. or 20..)	'90	'91	'92	'93	'94	'95	'96	'97	'98	'99	'00	'01	'02	'03
* Privatization of food budget	3	NA	6	7	16	25	38	48	60	64	69	72	80	85
*Privatization of enrichment studies for children	4	5	9	10	12	14	13	19	26	28	32	41	45	50
* Privatization of higher studies (part or whole)	1	3	3	4	7	7	10	8	11	15	21	28	48	53
* Privatization of health budget (part or whole)	NA	NA	NA	NA	NA	NA	NA	6	9	9	21	32	49	57
* Differential salary (part or whole)	NA	NA	NA	NA	1	2	6	10	16	12	31	43	50	57

^a Source: Getz (1998-2004) Annual surveys.

The Current Study

The unique advantages of kibbutzim for the current study. As mentioned above, research of kibbutz populations has found a positive relationship between adherence to the kibbutzim's unique principles of equality, solidarity, partnership, and democracy with health and well-being. However, a crucial test for this relationship was missing: what would happen if these communities were to give up their unique values? Would their levels of health and well-being deteriorate? Questions such as these have only recently become amenable to investigation because of the previous lack of variety between kibbutzim in this regard. This deficiency in past studies has now been "remedied" due to the structural transformations undergone by many kibbutzim. The current study exploits this opportunity.

The setting of contemporary kibbutz communities is of special importance to the research questions posed here due to several specific characteristics:

First, kibbutz communities are small (with between 100-1000 members) and decide autonomously (by a majority vote) on almost everything to do within the community and its individual members (within the confines of the state's laws). They also autonomously prioritize the usage of their economic resources. Second, kibbutz communities have very similar cultural backgrounds, histories, and political, geographical, and economic environments. Until very recently, members of different kibbutz communities also shared a very similar life style and experience. Third, the detailed records kept in the kibbutz federations' headquarters make it relatively easy to compare individual kibbutzim based on their size, economic performance, and other variables needed to control for possible confounding statistical relationships. Fourth, kibbutzim can clearly be defined as prosperous developed societies in terms of their material standard of living and life style.

The move to DSA and the ensuing socioeconomic inequality among members of kibbutzim is a very recent event. Therefore, if we find that DSA and inequality are related to social capital and health, this would indicate how little time is needed for changing levels of inequality to impact on society and its members' lives.

In addition, research with kibbutz communities raises the question of community size. What size of social entity should one refer to when asking whether socioeconomic inequality influences levels of social capital (of any kind), health and well-being? While various studies have offered answers, they are not satisfactory. Studies have been conducted at different social levels: countries, states (in the US), provinces, counties, metropolises, communities and neighborhoods (Wilkinson & Pickett, 2006). The general view is that the social entity in question should not be too big (so that it does not include several economies and cultures; e.g. Judge et al., 1998), though writers have also warned against the units being too small (Wilkinson, 1997; Wilkinson & Pickett, 2006; Robert, 1999). Wilkinson and Pickett, for instance, argue that the relationship between inequality and health is weakest in small social entities. Their explanation is that small communities are not in a position to enact social capital, particularly of the physical kind. Also, they may not have enough relevant comparative others to affect relative deprivation among the "have nots".

Nonetheless, this study relates to very small communities: the smallest ever studied in this domain, to the best of my knowledge (adult population of 100-1000). Why should I expect that such small communities will provide findings to my research questions? What matters, I suggest, is not the size of the social unit per se, but rather its position in several dimensions, the most important of which are: the social unit's perceived and actual ability to autonomously decide about investment in PHSC; its perceived and actual ability to autonomously develop opportunities and institutions for PSSC; members' closeness to one another and their mutual knowledge of each other such that they serve as significant points of comparison for each other; and finally, individuals carrying out most, and their most important, roles within the boundaries of the same social unit, implying that individuals cannot (psychologically) seek comparisons with member in other social units. The more of these criteria that a social entity meets, the more likely it is to display the relationship between level of socioeconomic inequality and health, as well as the other relations referred to in the former sections. Kibbutzim are probably communities that display these characteristics in the strongest possible way.

Hypotheses

I shall now rephrase the research questions posed in the introduction as hypotheses to be tested in the current kibbutz setting.

First research question: Do ideology and the basic social structures in society that emanate from it determine levels of socioeconomic inequality in that society while at the same time expressing social capital of both kinds (but particularly the physical kind)? Is a given society's leadership responsible for social structural changes?

Hypothesis 1: (a) Communities' leadership led changes to abandon "qualitative equality" as a guiding principle.

(b) Kibbutzim that decided to adopt DSA will show higher levels of socioeconomic inequality among their members; the more time that has passed since the kibbutz adopted DSA, the higher the level of socioeconomic inequality.

(c) The number of years since the adoption of DSA directly affects aspects of social capital, particularly of the physical kind, which involve the commitment of resources by the community (and not only via the intervention of socioeconomic inequality).

Second research question: Are the two dimensions of social capital (PHSC and PSSC) affected by level of socioeconomic inequality? Do they act as mediating variables between socioeconomic inequality and expressions of health and well-being? Do they contribute independently and together to explain variance in health and well-being? And what is their relative importance in that contribution?

Hypothesis 2: The level of socioeconomic inequality resulting from the adoption of DSA by kibbutz communities is negatively associated with:

(a) Expressions of PHSC in those communities;

(b) Expressions of PSSC in those communities.

And:

(c) Associations (a) and (b) are independent of one another;

(d) Expressions of social capital of both kinds independently and cumulatively affect levels of health and well-being;

(e) Expressions of social capital of both kinds act as mediating variables between level of inequality and expressions of health and well-being.

Third research question: Is greater socioeconomic inequality translated into lower average levels of health and well-being?

Hypothesis 3: The level of socioeconomic inequality caused by the adoption of DSA among kibbutz communities is negatively associated with average expressions of the physical health and well-being of members in these communities.

Chart 1 presents a summary of the hypotheses depicted in a causal flow.

Chart 1:Hypothesized causal flow model for the study: Leadership ideology, longevity-into-DSA, socioeconomic inequality, social capital with health and well-being.

Number of years in differential salary arrangement

Indicators of well-being and physical health

Level of socio-economic inequality

Expressions of physical & psychosocial social capital

Leaders' ideology

H1c

inequality

H2e

H3

H1a

H1b

H2a;H2b;H2c

H2d

It is first suggested that community leadership plays an important role in introducing DSA arrangements into kibbutzim (since this is based on observations and has not been validated yet by research, the arrow is thin). Then the number of years since the adoption of DSA affects levels of socioeconomic inequality in kibbutz communities, both of which affect levels of physical and psychosocial social capital (although years since the adoption of DSA affect them less). In turn, the existence of social capital is a major contributor to positive indicators of physical health and well-being and, as a mediating group of variables, absorbs most or all of the influence of DSA and inequality on well-being and health.

Previous research. A small scale study conducted in 2002 with four kibbutzim and members aged 55 or older (Leviatan & Salm, 2007) showed strong support for some of the hypotheses. However, that study had major limitations: it studied only two kibbutzim of the traditional kind and two of the differential kind; the kibbutzim with DSA had only adopted it one or two years prior to data collection; the sample size was small (N=140); there were no measures of the physical kind of social capital; and it did not represent all age groups. The current study aims to overcome these limitations.

Methods

The findings reported here are based on a two-stage study (focused on health services in kibbutzim) conducted by researchers at the Institute for Social Research of the Kibbutz at the University of Haifa, Israel (Adar et al., 2005; Leviatan et al., 2006). The first stage (Adar et al., 2005), conducted in 2004, was a survey of local expert informants about the health service domain in 131 kibbutzim (out of about 265). Tests performed on the data to compare this sample's economic and demographic data with the central tendencies and distributions for the total kibbutz movement showed very good matching on age and sex distribution, size of membership population, economic situation, and the distribution into budget arrangements (differential or traditional). The informants were interviewed about the state of health services and the budgetary arrangements for various health services.

Out of these 131 kibbutzim, we sampled 32 kibbutzim that represented various important criteria: (1) at least 80 members who are thirty years of age or older; (2) a distribution along the number of years since the adoption of DSA; (3) federation with which the kibbutz is affiliated; (4) economic situation. Our sample included 11 traditional kibbutzim, five kibbutzim that had adopted DSA 1-2 years prior to data collection, eight kibbutzim that had adopted DSA 3-4 years prior to data collection, and eight kibbutzim that made the shift 5-6 years prior to data collection. We had an average of 22 respondents per kibbutz, with a range of 18-33 respondents across all kibbutzim. Altogether, this stage comprised 700 respondents (44% response rate).

Instruments and variables of study. The major instrument for this study was a questionnaire sent by mail to all potential respondents. While questionnaires were answered by individuals, the current analysis is based on aggregated measures across all respondents in the sample from each kibbutz community. The relevant variables for the current report were as follows:

Demographic and biographic data: age; sex; family status; years of education; number of children in school; highest social office or managerial position held in the kibbutz in the five years prior to data collection; and self-reported monthly family salary (for "differential" kibbutzim) or family budget (for traditional kibbutzim) expressed in grouped categories of NIS.

Four measures indicative of socioeconomic inequality in kibbutzim: perceived level of disparities among members in the spheres of economic and material standard of living; satisfaction with the level of economic equality among members; perceived relative level of income of own family (a combined index of such questions about self, spouse, and family) compared to other members (an average perception of being relatively low indicates higher inequality); and the standard deviation (SD) of the former measure (the larger the SD the higher the perceived disparity in income distribution). Responses to each of these questions were aggregated across respondents from each kibbutz. These four items were later combined into an index of inequality as explained in the "Findings" section below.

Six measures of a priori assumed "physical social capital" (PHSC) that express investments of financial and other material resources by the kibbutz for the sake of its individual members: (1) mean number of years of education (signifies investment in higher education); (2) perception of the strength of the kibbutz's commitment to investing in creating appropriate jobs for members (two items: for members in general and for older members); (3) percentage of members reporting that they are unemployed; (4) satisfaction with kibbutz institutions' response to personal unique demands and needs; (5) average amount privately paid by respondents for medications (in NIS); (6) an index of satisfaction with health services on the kibbutz (16 items).Indices composed of two or more items are the result of factor analyses at the individual level of analysis and computed as means across the items in the ensuing factor. Means, SDs, and αs (where applicable) are shown in Table 6. Indices for the next two groups of variables were similarly calculated.

Eleven measures of a priori assumed "psychosocial social capital" (PSSC):(7) satisfaction with sense of feeling at home and belonging to kibbutz; (8) improvement or deterioration in the social situation of the kibbutz in comparison to two years earlier; (9) frequency of participation in social and cultural events on the kibbutz (7 items); (10) satisfaction with social relations on kibbutz ( 4 items); (11) satisfaction with the information one has about the kibbutz and one's level of influence in it (4 items); (12) social support on the kibbutz (2 items); (13) the frequency of organized social events on the kibbutz (4 items); (14) the frequency of meeting other members at work and in the dining hall (2 items); (15) trust in the kibbutz's leadership (2 items); (16) trust in other members of kibbutz; (17) decision-making on the kibbutz (1=a few people make all the decisions; 5=most members participate).

Three expressions of well-being: (18) satisfaction with kibbutz life in general; (19) an index of three items relating to "organizational commitment to kibbutz life" (assuming that members' commitment is an indicator of adjustment to their life, which is an important aspect of well-being): choosing kibbutz life again, recommending it to a young loved one, selecting kibbutz life over other forms of living in Israel; (20) satisfaction with life in general. There were also two indicators of physical health: (21) health symptoms – a response indicating how many, out of 16 psychosomatic symptoms (e.g. blood pressure, back pain, abdominal pains, insomnia, headaches) – that one "always" experiences (the direction of the final score was transformed such that better health was represented by higher scores); (22) subjective health – an index composed of answers to three questions: subjective evaluation of own health, current health vs. health five years ago, own health vs. health of members of same age and gender.

Most of these measures were obtained on a Likert-type scale of responses between 1-5, where "5" denotes the most positive response and "1" the least positive or most negative response. In this report, the direction of several items had to be reversed in order to fit with the general direction of items.

In addition, we allocated each kibbutz to one of four categories based on the number of years (at time of data collection) since the adoption of DSA: "traditional"; 1-2 years since the adoption of DSA; 3-4 years; and 5-6 years. Respondents also indicated the extent to which they would implement "qualitative equality" (each according to his needs, limited by kibbutz capabilities)" in their hypothetical "best place to live".

Analyses procedures

Groups of variables from same sphere of content and with similar standing in the causal flow model were submitted to factor analyses. Indices were derived by averaging the scores across variables that loaded at least .50 on the same factor. Aggregated data of respondents from same kibbutz formed the basis for analyses with the kibbutz as the unit of analysis.

Findings

Descriptive statistics. I first present the age and sex distribution within each group and across the whole sample.

Table 3: Number of kibbutzim per years since the adoption of DSA, number of respondents in each group; Mean (SD) age, and percent of women.

Type of kibbutz	Number of kibbutzim	Number of respondents	Mean age (SD)	Sex (% women)
Traditional	11	260	56.2 (13.48)	52.6
1-2 years as Differential	5	114	55.5 (13.69)	54.5
3-4 years as Differential	8	158	61.1 (13.55)	54.9
5-6 years as differential	8	168	58.2 (12.77)	58.5
	32	700	59.5 (13.66)	54.7
F test			3.00 (p<.05)	NS

Respondents from the four groups of kibbutzim do not differ in their gender composition. While the difference in age composition is statistically significant, it does not seem meaningful and no two groups differed from one another (in a post hoc analysis) at a statistically significant level. In addition, the four groups did not differ from each other in family status (in all, about 80% lived with a spouse); there was a statistically significant difference between the groups in terms of having school-aged children: 44% of members in traditional kibbutzim had at least one child at school; 42% of those 1-2 into DSA; only 36% in the 3-4 years group; and 35% in the 5-6 years group.

The 17 variables depicting aspects of social capital are strongly inter-correlated (median level of correlation coefficient is r=.58 and only 19 out of the 136 inter-correlations are not statistically significant). Nevertheless, I decided to carry out the analyses using the single items or indices rather than creating indices based on second level factor analyses at the aggregate level. There are several reasons for doing this. First, the individual expressions of social capital describe various valid perspectives of social capital and the strong associations among them are probably real and not spurious associations or such that stem from the methodology in use. Second, the number of units of analyses is only 32 (at the aggregate level), which is too small for valid factor analyses of 17 variables. Nevertheless, at a later stage of the analysis, when testing for the mediating role of social capital, I created one global index of social capital (GSC), which is the mean of 16 of the 17 variables (the variable indicating amount of money paid for medications was not included due to its extremely different shape of distribution and central tendency). The reliability level of this variable (GSC) is α=.95.

Testing the hypotheses

Leaders' ideology (Hypothesis 1a). First, I tested the observation that the leadership of traditional kibbutzim is more supportive of the value of "qualitative equality" than the leadership of the differential kibbutzim. This analysis was conducted at the individual level of analysis. All respondents were asked to indicate how desirable it would be for them to have "qualitative equality" (to each according to needs, limited by kibbutz capabilities)" in their hypothetical "best place to live" (they were to choose from different kinds of settlements in Israel or abroad). Responses ranged through five categories, two of which indicated "desirability", two "undesirability", and the middle category "undecided". Since another question asked about the highest social office and/or managerial positions respondents had held within the past five years, it was possible to break the sample into two subsamples: those who had held (or hold) a senior social position or top managerial position (for instance, members of the kibbutz secretariat, chairs of committees, managers of work branches or businesses) and those who did not. Table 4 shows the percentages of supporters of "qualitative equality" among central office holders and rank and file members in the four groups of kibbutzim.

Table 4: Percent of those expressing the desirability of "qualitative equality" in a hypothetical "best place to live". Two groups: central office holders during the last five years and non-office holders. Four groups of kibbutzim according to time since the adoption of DSA

	Holders of high social or managerial positions		Members with no high social or managerial positions
Years in DSA	Percent desire	N	Percent desire	N
Traditional	75.1	112	69.2	120
1-2 yrs.	35.3	44	44.5	54
3-4 Yrs.	46.2	52	53.9	63
5-6 Yrs.	36.2	69	47.9	70
Average/total	52.8	277	56.7	307

Two interesting findings are demonstrated in Table 4. It is clear that the vast majority of the leadership of the traditional kibbutzim desires "qualitative equality" to be implemented (75.1%), while less than half of the leadership in the differential kibbutzim expressed such a desire (35.3%-46.2%). Analysis of variance returned F=12.08 (p<.000). Each of the differential groups has a lower score at a statistically significant level from the traditional group, though the differences between them are not statistically significant. The second important finding is that while a very high percentage of the rank and file members of traditional kibbutzim express a desire for "qualitative equality"; this desire is even more strongly felt by their leadership. However, the opposite is true for the three groups among the differential kibbutzim, where rank and file members are more supportive of "qualitative equality" than their leadership (and analysis of variance returned F=.037 (p<.05)). It is also important to note that across all 32 kibbutzim there were no correlations between the position of the central leadership and that of the rank and file. These findings are at least a partial indication of the importance of local leadership in determining the direction of social change even when it does not have the full support of the rank and file members.

Differential salary arrangements and socioeconomic inequality (Hypothesis 1b). The next test investigates how longevity in kibbutzim with DSA affects perceptions of socioeconomic inequality among community members.

Four measures were employed to test for level of socioeconomic inequality in the kibbutzim: the perceived level of disparities between members in the sphere of economic and material standard of living; satisfaction with the level of economic equality among members; the perceived relative level of one's own family income compared to other members (an average perception of being relatively low indicates higher inequality); the standard deviation (SD) of the former measure (the larger the SD, the higher the perceived disparity in income distribution). Responses to each of these questions were aggregated across the respondents from each kibbutz. Table 5 shows the levels of Pearson correlation coefficients of each of these measures with the number of years since the kibbutz adopted DSA.

Table 5: Correlations between number of years since adoption of DSA and expressions of socioeconomic inequality (N=32 kibbutzim)

Expressions of inequality		Pearson correlation coefficient (r) with yrs. since adoption of DSA
Perceived level of disparities among members a		.54**
Satisfaction with degree of equality in kibbutz		-.58**
Mean estimated relative level of family income compared to other members		.-59**
SD of perceived family monthly income compared to other members in kibbutz		.45**

** p<.01; ^a Appendix 1 includes analysis of variance for the four variables comparing the means of the four groups of kibbutzim.

As shown by Table 5, the four measures of economic inequality correlate with the number of years since the kibbutz adopted DSA, but the correlations are not very high. In addition, the mean of the absolute level of inter-correlations among the four variables is .55. These two characteristics suggest that each of the four measures implies a different perspective on economic inequality. However, the four measures converge to one index (after changing the direction of two variables) with a reliability of .80 (Cronbach Alpha). That index (a mean across the four measures) correlates with longevity in the differential arrangement (r=.65 for the 32 kibbutzim (p< .000)). It would be a fair conclusion, therefore, to state that the number of years since adopting differential salaries increases economic inequality in a kibbutz.

Other indicators strengthen the conclusion that time since the adoption of DSA positively affects economic inequality. For the differential kibbutzim I measured net family income after state and community taxes were levied. The analogous measure for the traditional kibbutzim was a question about the size of the free personal budget that members receive from their kibbutz. For each group of kibbutzim (at the individual level of analysis) I calculated the correlation between net income (for the differential kibbutzim) or budget (for the "traditional" kibbutzim) with the respondents' number of schooling years (assuming that this can serve as a proxy for present position or occupation). The correlations were as follows: for the traditional kibbutzim, r=.011 (NS); for 1-2 years, r=.22 (p<.05); for 3-4 years, r=.34 (p<.000); for 4-5 years, r=.34 (p<.000). This is another indication that as the years pass since the adoption of differential salary arrangements, income comes to depend more strongly on personal socioeconomic status.

The second indicator I used was based on a calculation of the standard deviation of the reported net income in each of the kibbutzim with DSA (and free personal budget in the traditional kibbutzim), the mean of SDs across all kibbutzim in the same group and a comparison of these means of SDs across the four groups. The results corroborated expectations: the mean SD for the traditional kibbutzim was 954; for the 1-2 years group, 2682; for 3-4 years, 3423; and for 5-6 years, 3384. These results signify an increase in income disparity between traditional kibbutzim and those that adopted DSA in the last 1-2 or 3-4 years (while no differences appeared for the two last groups). It should also be mentioned that the mean of monthly incomes (salary or budget) did not differ across the four groups of kibbutzim – they were all around 4,100 NIS.

A third indicator is tentative but still of interest. For the three groups of kibbutzim that already have differential arrangements, I compared the percentage of families with a net income of 3,500 NIS or less with the percentage of families earning 12,000 NIS or more (only respondents with spouses were included in order to hold family size constant). I found the following results: for the 1-2 years group, 26.6% reported an income of 3,500 or less, while 2.5% reported an income of 12,000 or more; the respective figures for the 3-4 years group were 27.7% vs. 3.6%; and for the 5-6 years group they were 28.7% vs. 7.4%. These differences are not statistically significant but may be indicative of a trend of widening income disparity as the years since the adoption of differential salaries increase.

The four measures presented in Table 5 are all indicators of economic inequality. However, analysis of all 700 individuals at the individual level showed similar results for inequality on social dimensions (Leviatan, et al., 2006). For instance, while the members of traditional kibbutzim had the average highest score on a scale of self-reported level of management at the workplace as well as the highest score on a scale of "centrality" (high) or "peripherality" (low) among members of their kibbutz, the standard deviations on these dimensions among traditional kibbutz members were the smallest of all four groups.

The finding that economic inequality is a function of time since the adoption of DSA points to another conclusion: there is a jump in inequality between traditional kibbutzim and those who adopted differential salaries only one or two years earlier; then there is another jump between the latter group and the kibbutzim that had adopted DSA 3-4 years earlier. However, there were no differences, or only minor ones, between the 3-4 years group and the 5-6 years group. These results for the variables of inequality are shown as ANOVA analyses in Appendix 1. It is impossible to know at this time whether these results signify a temporary plateau, or whether inequalities reach a ceiling after 3-4 years.

DSA, inequality, and social capital (Hypotheses 2a; 2b). Another important question is whether time since the adoption of DSA also brings about changes in other characteristics not covered by measures of economic inequality.

The answer to this question calls for deciding first about the relationship of time since the adoption of DSA and level of inequality with measures of social capital; and whether both are needed to test the hypotheses in this study. Table 6 shows the correlation coefficients [r] of time since adopting DSA and the index of economic inequality with 17 expressions of social capital (6 assumed to be PHSC and 11 assumed to be PSSC). It also shows whether the relationship of social capital with DSA remains when inequality is held constant.

Several clear conclusions emerge from Table 6:

(a) Summary statistics show the mean Pearson correlation coefficient across all 17 expressions of social capital with years since the adoption of DSA to be r=-.56 with a median of r_md=-.54. For the index of inequality the respective numbers are r=-.67 and r_md=-.69. Therefore, and also because of the high inter-correlations among the variables of social capital (as described above), I introduced the GSC index (global social capital) into the last row of the table. This index shows a correlation of -.72 with DSA and -.87 with inequality. Thus, there is a strong albeit not complete overlap of years since the adoption of DSA with economic inequality in their relationships with indicators of social capital.

Table 6: Pearson coefficient correlations (r) of years since the adoption of DSA and level of inequality with 17 expressions of social capital; partial correlations of time since the adoption of DSA with the indicators of social capital while holding the index of economic inequality constant; multiple regressions of both predictors (time since the adoption of DSA and inequality) (32 kibbutzim)

	Mean (SD) a	Index of economic inequality		Inequality; Yrs. since DSA adoption held constant		Yrs. since DSA adopted	Yrs since DSA adopted; inequality held constant
Physical Social Capital; internal consistency, Alpha)
1. Level of formal education by number of years 1=<12; 2=13-14; 3=15-16; 4=17+	2.23 (31)		-.47**		—	-.39**		—b
2. Kibbutz commitment to create appropriate jobs (in general and for older members) (2 items) (alpha=.79)	2.68 (.56)		-.67***		-.40*	-.70***		-.46** R=-.75*** (multiple regression)
3. Percent of members reporting not working (1=not working; 2=working)	17.29 (13.23)		-.47**		—	-.49**		—
4. Satisfaction with kibbutz institutions' response to personal unique demands and needs	3.31(.37)		-.70***		-.50***	-.58***		—
5. Average amount in NIS a respondent pays privately for medication	86.47 (74.44)		-.57***		—	-.66***		-.47**
6. Satisfaction with health services on kibbutz. Index of 16 items) (alpha=.96)	3.55 (.43)		-.42**		—	-.56***		-.42*
Psychosocial Social Capital
7. Satisfaction with sense of feeling at home and belonging to kibbutz	3.58 (.39)		-.68***		-.59***	-.41*		—
8. Improvement (5) or deterioration (1) in social situation of kibbutz compared to two years earlier	2.45 (.36)		-.70***		-.55***	-.53**		—
9. Frequency of participation in social and cultural events on kibbutz (5=very frequent) (7 items) (alpha=.83)	2.44 (.38)		-.64***		-.34 P=.062	-.70***		-.48**
10. Satisfaction with social relations on kibbutz (4 items) (alpha=.79)	2.97 (.42)		-.72***		-.54**	-.61***		—
11. Satisfaction with the information about the kibbutz and one's level of influence in it (4 items) (alpha=.87)	2.87 (.35)		-.78***		-.71***	-.48**		—
12. Social support on kibbutz (2 items) (alpha=.82)	3.10 (.54)		-.82***		-.69***	-.63***		—
13. Frequency of organized social events on kibbutz (4 items) (alpha=.66)	2.10 (.43)		-.66***		-.54**	-.45**		—
14. Frequency of meeting other members at work and the dining hall (2 items) ( alpha=.44)	3.12 (.70)		-.78***		-.57***	-.77***		-.55*** R=-.85*** (multiple regression
15. Trust in kibbutz leadership (2 items) (alpha=.78)	3.02 (46)		-.78***		-.70***	-.46**		—
16. Trust in other members	3.10 (.33)		-.74***		.61***	-.54***		—
17. Decision-making on the kibbutz (1=a few people make all the decisions; 5=most members participate)	2.45 (.44)		-.67***		.52**	32**		—
Global Social Capital (mean across 16 variables)	2.62 (.33)		-.87***		.75***	-.72***		r=-.41*  R=-.89*** (multiple regression)

^a Detailed Means and SDs for each of the four groups of kibbutzim appear in appendices 2 & 3; *p<.05; **p<.01; ***p<.001, all two tailed; ^b NS

(b) When the index of inequality is held constant, five of the correlations of DSA with measures of social capital are still statistically significant, and three of them are of the physical kind (the kibbutz's commitment to creating appropriate jobs (2); money paid privately for medication (5); and satisfaction with health services (6)). Two are from what was a priori assumed to belong to the psychosocial kind of social capital (frequency of participation in social and cultural events (9); opportunities to meet other members at work and in the dining hall (14)). However, to an extent, these variables also indicate the kibbutz's priorities in how it invests its resources: in many of the differential kibbutzim the common dining hall does not function anymore and a large percentage of their members no longer work within the kibbutz but rather as employees of external organizations. Thus, the physical opportunities for such meetings are greatly reduced. The same is true of kibbutz cultural and social events, which have been much diminished in the differential kibbutzim. Therefore, it is not just the social aspect of not having contact with other members that has diminished, but also the physical opportunities for such contact.

(c) The results of Table 6 also show that while the index of inequality is the stronger contributor to explained variance in social capital, time since the adoption of DSA contributes to levels of social capital (particularly to its physical kind) over and above the contribution of level of equality. In addition, in some instances their contributions are uniquely independent. Indeed, multiple regression analyses show that for two expressions of social capital (the kibbutz's commitment to creating jobs and meeting members in the dining hall and at work), the combined contribution of both time since the adoption of DSA and inequality to the explained variance in levels of social capital is higher than that of each them separately. This is also the case with the global measure of social capital (GSC).

(d) One apparent reason that the correlations with time since the adoption of DSA is not even stronger is because there is little difference between the 3-4 years group and the 5-6 years group (this is shown in appendices 2 & 3, which present the results of analyses of variance for all 17 expressions of social capital). I also demonstrate this result with the index of Global Social Capital (a mean across all 16 indicators of social capital) in Table 7 below. The relationship is not linear: there is a jump between the traditional kibbutzim to the 1-2 years group; then another jump to the 3-4 years group, but no difference between the two last groups.

Table 7: Analysis of variance — the index of Global Social Capital (GSC) as a function of number of time since the adoption of DSA

	Budgetary arrangement	Number of kibbutzim	Mean	SD	F & (r)
Global Social Capital (GSC)	Traditional	11	2.92	.27	F=14.23 r=-.72 p<.000
	1-2 yrs. Differ.	5	2.70	.17
	3-4 yrs. Differ.	8	2.34	.16
	5-6 yrs. Differ.	8	2.40	.20
	Total	32	2.61	.33

Association of time since the adoption of DSA, inequality and social capital with well-being and health. The next hypotheses to be tested concern the relationships between time since the adoption of DSA and socioeconomic inequality with well-being and physical health (Hypothesis 3); the relationships between social capital of both kinds (physical and psychosocial) with well-being and physical health (Hypothesis 2d); the mediating role of social capital in the relationship between time since the adoption of DSA and socioeconomic inequality with well-being and physical health (Hypothesis 2e). I present the findings regarding these hypotheses in Table 8 (simple and partial correlations with indicators of well-being and health) and Table 9, which presents a summary of five multiple regressions (the indicators of well-being and health as dependent variables) regressed on the 17 expressions of social capital.

(A note about the multiple regression analysis: because of the very high multi-collinearity among the expressions of social capital and the small number of units for analysis (32), I performed the linear multiple regression analysis in three stages: first, I ran an analysis of the 6 expressions of PHSC as predictors. Only those predictors that entered each formula were to be included into the final analysis. The same procedure was applied to the 11 expressions of PSSC. In the third stage I ran the predictors that were entered into any of the formulas in the first two stages. The resulting regression models are shown in Table 9.)

Table 8: Relationships (Pearson's r) of all expressions of social capital, a global index of social capital (GSC), time since the adoption of DSA, and socioeconomic inequality with five measures of well being and physical health. Also, the partial relationships (r_partial) of GSC with well being and health when time since the adoption of DSA and socioeconomic inequality are held constant

Indicators of well being and physical health
17 Indicators of social capital; a global index of social capital, time since the adoption of DSA and inequality.	18. Satisfaction with kibbutz life	19. Commitment to kibbutz life (3 items alpha=.73	20. Satisfaction with life in general	21. (no) symptoms (out of 16)	22. Subjective health (3 items alpha=.78)
Physical SC
1. Level of formal education by number of years	—a	—	.43*	.64***	.35*
2. The kibbutz's commitment to creating appropriate jobs for its members	.73***	.71***	.72***	—	—
3. Percent of members reporting being unemployed	—	—	.43*	.38*	.51**
4. Satisfaction with kibbutz institutions' response to personal unique demands and needs	—	.34*	—	.46**	—
5. Average amount in NIS a respondent pays privately for medication	.76**	.56**	.73**	—	—
6. Satisfaction with health services on kibbutz	.43*	.51**	.46**	—	—
Psychosocial SC
7. Satisfaction with sense of feeling at home and belonging to kibbutz	.88**	.64**	.74**	—	—
8. Improvement or deterioration in social situation of kibbutz compared to two years earlier	.68**	.64**	.67**	—	—
9. Frequency of participation in social and cultural events on kibbutz	.43*	.37*	.39*	.36*	—
10. Satisfaction with social relations on kibbutz	.69***	.60***	.74***	—	—
11. Satisfaction with the information one has about the kibbutz and one's level of influence in it	.67**	.53**	.70***	—	—
12. Social support on kibbutz	.68**	.59**	.65**	—	—
13. Frequency of organized social events on kibbutz	.404*	—	. —	—	—
14. Frequency of meeting other members at work and in the dining hall	.61***	.59***	.53***	—	—
15. Trust in kibbutz leadership	.71**	.63***	.63***	—	—
16. Trust in other members of kibbutz	.74***	.59***	.62**	—	—
17. Decision-making on the kibbutz	.65***	.50**	.68***	.39*	—
Global index of social capital (GSC)	.78***	.66***	.75***	.28 p=.11	—
Longevity in DSA	-.48**	-.56***	-.48**	-.34*	—
Index of inequality	-.61***	-.54***	-.59***	-.251 p=.165	—
Longevity-into-DSA; GSC held constant (rpartial)	—	—	—	—	—
Index-of-inequality; GSC held constant (rpartial)	—	—	—	—	—
GSC when time since the adoption of DSA is held constant (rpartial)	.72***	.45**	.67***	—	—
GSC when inequality is held constant (rpartial)	.63***	.46**	.59***	—	—
Means (and SDs) for indicators of well being and health	3.66 (.29)	2.85 (.52)	3.74 (.26)	15.18 (2.24)	3.37 (.18)

*p<.05; **p<.01; ***p<.001, two-tailed; ^a NS.

Tables 8 and 9 and some other analyses demonstrate the major findings of this study. Table 8 shows the following:

1. The correlations of time since the adoption of DSA and economic inequality with the three indicators of well-being are substantial (ranging between -.48 to -.61) and therefore support Hypothesis 3. However, when the partial correlation of time since the adoption of DSA and inequality with well-being and physical health are computed while holding the GSC index constant (by a partial r procedure), the correlations disappeared and were no longer significant. This means that the expression of social capital mediates the relationships between time since the adoption of DSA and inequality with indicators of well-being and health, as suggested by Hypothesis 2e.

Table 9: Multiple regressions of indicators of well being and health on expressions of social capital

Predictors	Satisfaction with kibbutz life		Commitment to kibbutz life		Satisfaction with life in general		(no) symptoms			Subjective health
Physical Social Capital	β	r a	β	r	β	r	β	r		β		r
Level of formal education by number of years			-.31	-.02	.29	.43	.63	.64
The kibbutz's commitment to creating appropriate jobs for its members			.59	.71
Average amount in NIS a respondent pays privately for medication							.46	.46
Percent of members reporting being unemployed										.51		.51
Psychosocial Social Capital
Improvement or deterioration in social situation of kibbutz compared to two years earlier			.37	.70
Satisfaction with sense of feeling at home and belonging to kibbutz	.88	.88			.68	.74	-.32	-.03
R	.88		.82		.79		.79		.51
R2	.78		.67		.63		.616		.26
R2adj	.77		64		.61		.58		.26
B	1.29		.59		1.61		13.10		4.17
F	104.09		19.26		24.76		14.98		10.65
Sig.	.000		.000		.000		.000		.003

^a A multiplication of β*r helps estimate the percent of variance explained by each predictor in a multiple regression analysis as based on the formula with standardized scores: since R² =r₁* β₁+r₂* β₂+…r_n* β_n, therefore_, 100*( r_n* β_n/ R²( estimates the percent of explained variance due to predictor n. No such interpretation is possible when the signs of r and β are different from each other (which happens infrequently and for very low correlation coefficients).

1. Individual expressions of social capital correlate highly with the three indicators of well-being. However, the associations are weaker for the social capital of the physical kind. While only 12 out of the possible 18 correlations for PHSC are significant and the median correlation coefficient level is r_md=.53 (p<.01), the respective numbers for PSSC are higher: 31 out of the 33 possible correlations are significant with a median of r_md=.64 (p<.001).
2. The index of GSC is strongly related (.78; .66; .75) with the three indicators of well-being, all significant at the p<.001 level.
3. When time since the adoption of DSA and inequality are held constant ( as shown at the bottom of Table 8), the relationship of GSC with well-being is still high, with a range of r (partial)=.45 to.72. This result supports Hypothesis 2d.
4. The results are different for the two indicators of health. First, the number of significant relationships is much smaller and they are all at a lower level. Yet, whatever association there is, it is stronger for the PHSC than for the PSSC. While five out of the possible 12 correlations are at an acceptable significance level for the expressions of PHSC, only two of the 22 possible correlations for the expressions of PSSC are statistically significant. Second, neither of the two health indicators is significantly related to the global measure of social capital (GSC). Thus, it is possible that social capital firstly affects well-being and only later indicators of physical health. It is possible that the first causalities of inequality and the ensuing deterioration in social capital are general indicators of well-being with physical health affected only later on.
5. Stressors and pressures in the environment that affect physical health, such as symptoms of sickness or self-evaluation of health, may take time to build up (see, for instance, Singer & Ryff, 1999; Stewart, 2006). It is possible that the reason these indicators do not show up yet is because not enough time has elapsed for the deterioration in levels of expressions of social capital to have become apparent.
6. Table 9 Shows that health and well-being are not only influenced by single expressions of social capital, but also by several of them in a cumulative way. The multiple regression coefficient R is higher for three of the dependent variables than it is for the highest single correlation in Table 8 (commitment – from .71 to .82; satisfaction with life – from .75 to .79; and (no)symptoms – from .64 to .79). It is at the same level for two indicators (satisfaction with kibbutz life and subjective health), because the one measure of social capital covers all other measures. I would suggest that this finding implies that health and well-being are affected cumulatively by several perspectives of social capital and are not determined by one perspective alone. This adds further support to Hypothesis 2d.
7. Measures of assumed PHSC explain (together) almost as much of the variance in the five indicators of well-being and health as the two measures of PSSC.
8. Finally, many of the expressions of social capital are in effect expressions at the individual level of analysis since they hinge on individual perceptions. The mechanism suggested for their effect on health is through individuals' reactions (see for instance Adler and Ostrove, 1999). If this is correct, then analyses at the individual level should be more sensitive in discovering associations of social capital expressions with indicators of physical health. Indeed, this is what we have found at the individual level of analysis (Leviatan et al., 2006): self-reported health was positively associated with number of school years; being in work (rather than unemployed); satisfaction with health services; satisfaction with the response of kibbutz institutions to unique personal demands; satisfaction with one's sense of feeling at home and belonging to kibbutz; satisfaction with the level of social support; participation in social and cultural events; and with meeting others at work or in the dining hall. The range of the correlations in these associations was between r=.10 to r=.27. All were in the expected direction and statistically significant to at least the p<.05 level with about 650 respondents. Also, at the individual level of analysis, respectable relationships appear between "satisfaction with life in general" and the indices of (no)symptoms and subjective health (r=.15, r=.26 with 650 respondents, for both p<.000) perhaps suggesting that "satisfaction with life in general" precedes in time indicators of physical health and serves as one of the factors that contributes to it  .

Summary and Discussion

I shall now summarize the extent of support for each of the hypotheses in the results of the study.

Hypothesis 1(a): The community's leadership led changes to abandon "qualitative equality" as a guiding principle.

The findings do not allow clear causal inferences in this regard as the study does not have a longitudinal design for the relevant variables for this hypothesis. However, the findings enable us to make a reasonable speculation. Recall that the leadership of the differential kibbutzim is in favor of abandoning "qualitative equality" more than the rank and file members (see Table 4). The opposite is true for the leadership of the traditional kibbutzim. An alternative interpretation might suggest that the leaders of the differential kibbutzim are more adaptable than the rank and file to the normative values in their communities. Therefore, when their kibbutzim moved to DSA, they immediately adopted values that are more appropriate for the new structure. Following the same rationale, the argument regarding the leadership of the traditional kibbutzim would be that they are the most adaptable members of their communities. However, this argument does not accord with what we know about leadership in general and its role in communities, organizations, or societies. The research literature on leadership generally sees leaders as influencing their organization, and not the other way around (e.g. Bass, 1985; Bass et al., 2003; Yukl, 2002). Also, who is primarily responsible for setting or changing social structures in organizations and societies if not the leadership? The rank and file? The masses? There have been such examples in history, but not without a revolution. In the case of the kibbutzim there was no revolution. I therefore conclude that although the first hypothesis is not directly supported, it is still supported by inference. Kibbutz leaders are the important instigators of structural sociological aspects of their society.

Hypothesis 1(b): Kibbutzim that decided to adopt DSA will show higher levels of socioeconomic inequality among their members; the longer a kibbutz has had DSA arrangements, the higher the level of socioeconomic inequality.

The findings regarding this hypothesis are clear. Table 5 indicates a strong relationship between time since the adoption of DSA with four measures of inequality, and the index composed of the four measures displays an even stronger relationship (r=.65). Also, while inequality is mostly measured by responses about perceptions – a "soft" measurement – DSA is a "hard" objective measure, and so one does not expect the methodological confounding of the two measures in such a way that would have raised their level of statistical association. In addition, the correlations reported are not higher because inequality does not increase (in these data) beyond the third year in DSA. However, could it be that the direction of causality is the opposite of that suggested, such that socioeconomic inequality causes the transformation in DSA? This cannot be the case, as, by law, no kibbutz could have instituted the equity principle as a function of position at work or in the community as the basis of its remuneration system before the shift to DSA. A related issue is the question of treating the data as if it is longitudinal, with the number of years since the adoption of DSA relating to the same kibbutzim, while in fact the research design is cross-sectional and compares kibbutzim that instituted DSA at different times in the past. This might be a problem if one could successfully argue that differences already existed among the groups of kibbutzim before they adopted DSA, and that those differences correlated with inequality and expressions of social capital (as this argument also relates to other hypotheses below). This is not the case for inequality, as explained earlier. As for social capital, at least one study has shown that social capital deterioration occurs after (and not before) major structural changes (Leviatan, 1998). Nonetheless, the data does not have a definite answer to the question of why inequality increases as time passes since the adoption of DSA. I could offer only speculations based on nonsystematic observations which would need to be validated. However, it is clear that DSA, which is a major social structural phenomenon, and time since the adoption of DSA, which is a measure of accumulated time, determine levels of socioeconomic inequality, and Hypothesis 1(b) is supported.

Hypothesis 1(c): The number of years since the implementation of DSA directly affects aspects of social capital, particularly of the physical kind, which involve the commitment of resources by the community (and not only via the intervention of socioeconomic inequality).

This hypothesis is also supported as demonstrated by findings in Table 6. Time since the adoption of DSA contributes directly over and above the contribution of socioeconomic inequality to the explained variance of five out of the 17 expressions of social capital; three of these are of the PHSC kind and the other two have at least some component of material investment for the sake of members, an investment based on social priorities. This direct unique contribution of time since the adoption of DSA is also recorded in relation to the index of Global Social Capital. It is important to discuss an alternative interpretation of the effect of time since the adoption of DSA on expressions of social capital. The transfer from a society based mostly on the principle of "qualitative equality" to one based on DSA must constitute a major and stressful change for its members – even for those who supported it. As a result, perceptions of low levels of social capital might not represent a genuine reduction in social capital, but rather a reaction to change itself and the need to readjust to the new social structure. If that interpretation is valid, we should have seen kibbutzim that are 1-2 years into DSA to show the lowest levels of social capital, but after that they should have returned to the levels of the traditional kibbutzim that had not experienced such major social changes. However, that is not the case. The level of social capital in kibbutzim that adopted DSA three years ago remains below those just 1-2 years into DSA, and certainly below the level of the traditional kibbutzim.

1. Hypotheses 2(a-c): The level of socioeconomic inequality resulting from the adoption of DSA by kibbutz communities is negatively associated with (a) Expressions of PHSC in those communities; (b) Expressions of PSSC in those communities;   And (c) Associations (a) & (b) are independent of one another.

Once more, the findings presented in Table 6 offers strong support for these hypotheses. The strong simple Pearson's correlation coefficients of inequality with PHSC and PSSC and with the index of GSC presented in Table 6 provide evidence for this, and the fact that most of the relationships stay even when DSA is held constant. However, the index of inequality has some limitations when related to expressions of social capital as measured here since both share the same methodology (questionnaire).Thus, some part of the common variance is apparently due to this shared methodology. An alternative is to look at time since the adoption of DSA as a proxy for inequality. In this case, the strength of the relationships with expressions of social capital of both kinds is only very slightly reduced.

1. Hypothesis 2(d): Expressions of social capital of both kinds independently and cumulatively affect levels of health and well-being.

Support for this hypothesis is offered in Tables 8 and 9, but social capital affects well-being more than the indicators of physical health. Table 8 demonstrates how the three indicators of well-being are affected by levels of social capital, but also how PHSC is more closely related to the health indicators than PSSC. While the two indicators of health did not relate to the index of GSC, they did relate to several of the individual expressions of social capital, particularly to the expressions from the PHSC group (Tables 8 and 9). Table 9, the multiple regression analysis, adds further support to Hypothesis 2(d). In addition, I have shown that expressions of social capital at the individual level of analysis show a much stronger relationship and suggest that perhaps this is the level of analysis that is more appropriate for such analysis. I also suggest that indicators of health might be affected by levels of well-being and follow them in time.

1. Hypothesis 2(e): Expressions of social capital of both kinds act as mediating variables between level of inequality and expressions of health and well-being.

A clear support of the hypothesis is given with the analyses of the data in which GSC is held constant while calculating relationships of time since the adoption of DSA and inequality with well-being and health. Holding GSC as constant turns the relationship to becomes non-significant.

Hypothesis 3: The level of socioeconomic inequality caused by the adoption of DSA among kibbutz communities is negatively associated with average expressions of physical health and well-being of members in these communities.

This is the overarching hypothesis of the current study. Indeed, as shown in Table 8, the relationships hypothesized do indeed exist. However, as explained in the summary points of the previous hypotheses, they are spurious and totally mediated by the expressions of social capital.

(8) General summary of findings. Major social structural changes – such as the shift to DSA – negatively affect both levels of equality and social capital. The longer these structures persist, the more detrimental their effect. This is true – based on the data presented here – up to a point (three years), when it reaches a plateau. Whether this plateau will change in the future is not yet known. It would appear that the local leadership is responsible for changes in social structure and thus for levels of inequality in the communities, and also for the levels of expressions of social capital. Time since the adoption of DSA is responsible for the level of inequality and is also directly responsible for some expressions of social capital, especially those that require the investment of material resources by the community. Thus, the extension of research "upstream" into more general social structures and processes (as suggested, for instance, by House, 2001; Muntaner & Lynch, 1999; Coburn, 2000) can be seen to have paid off: the causal flow does not start with inequality.

Level of socioeconomic inequality is mostly responsible for the psychosocial expressions of social capital. Both "physical" and "psychosocial" expressions of social capital are strongly determined by time since the adoption of DSA and socioeconomic inequality. This is true for most of the 17 different expressions of social capital employed in this study. Thus, it seems that theory should incorporate both expressions of social capital, as suggested by Kawachi (2002; 2004) and others. Social capital affects levels of well-being and health and serves as a mediating variable between time since the adoption of DSA and inequality, and the indicators of well-being and health.

The general model suggested by Chart 1 is strongly upheld with some minor modifications. Iin particular, the separation of indicators of health from those of well-being, as it seems that the former react differently than the latter to expressions of social capital.

In addition, all findings are the result of a research design that was very attentive to critiques of methodological problems in previous studies. Specifically, units of the study were equated for all imported aspects that could enter as partially confounding variables and the number of units was methodologically comfortable.

Several additional discussion points are now in order.

(9) Short time to exhibit effects. It is remarkable that the effects of the social structure of DSA occur so fast after it is implemented.  A major deterioration in levels of equality and in expressions of social capital is already apparent within only 1-2 years of adopting DSA. The question is whether such rapid changes could also take place in reverse, that is, social structural changes that would increase the level of equality and change priorities in the investment of material resources for the benefit of all members of society. Since such changes are the responsibility of the leadership (the bearers of power in the political system) I see no reason why not. Investment in education in Israel during the years 1990-2006 provides an anecdotic example of this idea. There were four changes of government during that period of time, each instigating a change in governmental ideology: in 1992 (Likud party (neo-liberal ideology in economic affairs) to Labor party), 1996 (Labor to Likud), 1999 (Likud to Labor), and 2001 (Labor to Likud). Political changes in the government also changed the direction of investment in education: when Labor took office there was an increase in relative (to GDP) expenditure for education; when Likud took office there was a decrease.

1. Absolute low levels of inequality. It is also remarkable that such strong effects of inequality were discovered while differences in absolute inequality levels are small. Even though throughout the paper I have referred to the kibbutzim with DSA as communities with high level of inequality among their members, to a certain extent this is inaccurate. While there is socioeconomic inequality in these communities, this is only in comparison to traditional kibbutzim. That difference is relative but – when one looks at differences in inequality among communities or societies elsewhere – not major. Also, members of the differential communities are still much more equal to one another as compared to the society around them. Nonetheless, the strong effects of that relative inequality have already been recorded. This could be seen as supporting the claim regarding the importance of the concept of relative inequality.

(11) The possibility of investing in material social capital though without reducing inequality. Some researchers (e.g. House, 2001) have suggested that inequality does not necessarily have to be reduced in order for social capital to be strongly expressed (particularly the material kind but also some aspects of the psychosocial kind). This is because investments in social capital can be made by the government independently of levels of inequality. While this is theoretically possible it does not appear to be very likely. The kibbutz experience shows that it does not work even in societies where a social structure of strong solidarity was deeply embedded. While it is likely that many of the leaders are still – in their hearts – supporters of at least some measure of solidarity for their communities, once the social structure has changed it follows its own dynamics, expressed in major deteriorations of social capital of both kinds. Moreover, the dynamics are such that the strongest and most powerful do not care about the weakest.

1. The boundaries and size of units where inequality counts. This study gives a clear answer to the question of the appropriate boundaries and size of communities when studying the effect of inequality and social capital on well-being and health. As explained in the introduction, kibbutzim are very small communities, yet the results of the current study are very strong. As proposed in the introduction, it is not size per se that counts, but rather certain special characteristics of the community. These might include its perceived (by members) and actual ability to autonomously decide about investment in PHSC; its perceived and actual ability to autonomously develop opportunities and institutions for PSSC; members' closeness to one another and their mutual knowledge of each other such that they serve as significant comparisons for each other; individuals carrying out most, and their most important, roles within the boundaries of the same social unit, implying that individuals cannot (psychologically) seek comparisons with members of other social units. Therefore, when planning research into the effects of intra-community levels of inequality in social entities, one should try to estimate the standing of these entities based on the criteria enumerated above and to only expect results regarding entities that meet at least some of these criteria.
2. Future research

(a) Future research into kibbutzim should focus on comparing the two kinds of kibbutzim ("traditionals" and "differentials"); in particular, changes in social capital among the differentials imply heightened mortality rates. However, such a study will have to wait a few more years for the health expressions to take hold and be translated into mortality.

(b) Research should focus on more refined measures of physical health than those used in the current study. The study we executed as a pilot (Leviatan and Salm, 2007) made use of measures of health that included (in addition to what was used here) analyses of medical records and evaluation of health status by a physician or nurse. That study showed differences in the combined health indicators after 1-2 years of DSA (more positive health for the "traditionals"). In the current study we did not use the extra two measures both because of cost considerations and also because we thought that the measures we did use correlated strongly enough with the ones we left out. This turned out to be wrong, and future studies should not repeat this mistake.

(c) Research should focus on the processes that local leaders in the differential kibbutzim go through – from being supporters of the unique ideology of kibbutzim (equality, solidarity, community, collectivism) to becoming leaders of neo-liberal ideology and individualism. I assume that the current leaders of the differential kibbutzim must have been supporters of traditional kibbutz ideology in the recent past (because they were in leadership positions when their communities were based on these values). This is particularly interesting given that the literature on values and value change states that values are a very central and stable characteristic of a person (e.g. Rokeach, 1973; Shwartz, 1994). If this is so, how are we to account for the rapid change in values among kibbutz leaders? Perhaps the commitment to kibbutz values never ran very deep? Or might there be some other explanation?

(d) I plan a parallel analysis of the effects of inequality on health and well-being at the individual level of analysis with the data of our study (some of which has already been carried out; see Leviatan et al., 2006). Such analyses should improve our understanding of the processes and insights reported here at the individual level of analysis.

1. Practical implications. I could have written a full length advocacy paper about the social and political implications of this study for kibbutz society and society in general. However, such a paper would only repeat material written by many other researchers in the field. I shall therefore conclude with the words of hope written by one of them (and with whom I fully agree): "As research on the socioeconomic determinants of health progresses, and public understanding of the issues increases, the demand for social reform will become unstoppable. Growing knowledge changes both the morality and the rationality of the status quo. It turns official inaction into culpable negligence" (Wilkinson, 1996:25).

Bibliography

Adar, G., Goldemberg, J. & Leviatan U. (2005). Health systems in kibbutzim. Institute for Kibbutz Research, University of Haifa, 189a (Hebrew).

Adler, N. E. & Ostrove, J. M. (1999). Socioeconomic status and health: what we know and what we don't. In: Adler, N. E., Marmot, M., McEven, B. S. & Stewart, J. (eds.). Socioeconomic status and health in Industrial nations: social, psychological, and biological pathways. New York, Annals of the New York Academy of Sciences. Vol. 896, 3-15.

Bass, B.M., Avolio, B.J., Jung, D.I., & Berson, Y. (2003). Predicting unit performance by assessing transformational and transactional leadership. Journal of Applied Psychology, 88(2), 207-218.

Bass, B. M. (1985). Leadership and performance beyond expectations. New York: Free Press.

Carmel, S., Lapidot, A. & Motarn, A. (1995). Social support, life events, health and satisfaction with life among the aged in kibbutz and city. Gerontology 70, 6-18 (Hebrew).

Carmel, S., Motarn, A. & Shemi, G. (1996). Predicting subjective health evaluation among the aged in city and kibbutz. Chevra and Revacha 16(3), 361-380 (Hebrew).

Catalano, R. (1998). An epidemiological perspective. Journal of Community & Applied Social Psychology, 8, 165-168.

Central Bureau of Statistics, Israel (1979;1986;1997). Statistical Abstracts of Israel.

Cohen, A. (2006). Salary disparities in kibbutzim reaches 357%. Haaretz Daily, www.haaretz.co.il ,16.10.2006. (Hebrew).

Cohen-Mansfield, J. & Leviatan, U. (1992). The importance of the work role for aged kibbutz members. Canadian Journal of Community Mental Health, 11(2), 35-48.

Coburn, D. (2000).Income inequality, social cohesion, and the health statues of populations: the role of neo-liberalism. Science and Medicine, 51, 135-146.

Getz, S.(1990-2004). Annual Survey of Changes in Kibbutzim. Institute for Social Research of the Kibbutz, University of Haifa (Hebrew).

Harel, Y. (1990). The new Kibbutz. Jerusalem, Keter Publishing (Hebrew).

Homans G. C. (1961). Social Behavior: Its Elementary Forms. New-York: Harcourt, Brace & World.

House J. S. (2001). Relating social inequalities in health and income. J. of Health Politices Policy and Law, 26(3), 523-532.

Judge, K., Mulligan, J.A. & M. Benzeval. (1998). Income inequality and population health. Social Science and Medicine, 46 (4-5), 567-579.

Kaplan, G. A., Pamuk, E.R., Lynch, J.W., Cohen, R.D., & Balfour, J. L. (1996). Inequality in income and mortality in the United States: analysis of mortality and potential pathways. BMJ, 312, 999-1003.

Kawachi, I. & Kennedy, B. P. (1997). Socioeconomic determinants of health: Health and social cohesion: Why care about income inequality? British Medical Journal, 314, 1037-1047.

Kawachi, I. & Kennedy, B. P. (1999). Income inequality and health: Pathways and mechanisms. Health Services Research, 34, 915-930.

Kawachi, I. (1999). Social capital and community effects on population and individual health. In: Adler, N. E., Marmot, M., McEven, B. S. & Stewart, J. (eds.). Socioeconomic status and health in Industrial nations: social, psychological, and biological pathways. New York, Annals of the New York Academy of Sciences. Vol. 896, 120-130.

Kawachi, I., Kenedy,B. P. & Wilkinson, R. G. (1999). Social inequalities and health: a reader. New York: The New Press.

Kawachi I., Subramanian S.V., & Almeida-Filho N. (2002). A glossary for health inequalities. J. Epidemiol Community Health, 56, 647-52.

Kawachi,I., Kim, D., Coutts, A. & Subramanian, S.V. ( 2004). Commentary: Reconciling the three accounts of social capital. International Journal of Epidemiology, 33(4), 682-690.

Kennedy, B.P., Kawachi, I. & Prothrow-Stith, D. (1996). Income distribution and mortality: cross sectional ecological study of the Robin Hood index in United States. British Medical Journal, 312, 1004-1007.

Leviatan, U. (1983). Work and aging in the kibbutz, Aging and Work 6 (3), 215-226.

Leviatan, U. (1998). The relationship of privatization of the food budget with social climate in kibbutzim, Hakkibutz , November, 1998. (Hebrew).

Leviatan, U. & Cohen, J.(1985). Gender differences in life expectancy among kibbutz members. Social Science and Medicine, 21(5), 545-51.

Leviatan, U. & Salm, CH. (2007). Socioeconomic inequality and inequalities in health among kibbutz elderly, in Estes, R. J. (ed): Advancing Quality of Life in Turbulent World, Kluwer Academic Publishers B. V., Dordrecht : The Netherlands, 169-190.

Leviatan, U., Adar, G., & Goldemberg, J. (2006) Socio-Economic Inequality in Kibbutzim and Its Effect on Members' Health (Study Report. Institute for kibbutz Research, University of Haifa, 189b. (Hebrew).

Leviatan, U., Cohen-Mansfield J., & Jaffe-Katz, A. (1986). Life expectancy of Kibbutz members, International J. of Aging and Human Development, 23(3), 195-205.

Leviatan, U., Adar, G. & Am-ad, Tz. (1981). Aging in kibbutz: factors that determine satisfaction with Life. Hakibbutz, 8, 16-42 (Hebrew).

Leviatan, U., Oliver, H., & Quarter, J. (1998). Crisis in the Israeli Kibbutz: Meeting the Challenge of Changing Times. Praeger Publishers.

Leviatan, U. (1988). The Community as a source of social support for the individual, Gerontology 41/42, 28-38 (Hebrew).

Leviatan, U. (1999). Contribution of social arrangements to the attainment of successful aging — the experience of the Israeli kibbutz. Journal of Gerontology: Psychological Sciences 54b, 205-213.

Leviatan, U. (2003). Is it the End of Utopia? The Israeli Kibbutz at the Twenty-First Century, Center for Study of Co-operatives, University of Saskatchewan.

Lynch, J. W. & Kaplan, G. A. (1997). Understanding how inequality in the distribution of income affects health. Journal of Health Psychology, 2(3), 297-314.

Lynch J.W., Davey S.G., Kaplan G.A., & House J.S. (2000). Income inequality and mortality: importance to health of individual income, psychosocial environment, or material conditions. BMJ, 320, 1200-1204

Marmot, M. (1999). Epidemiology of socioeconomic status and health: are determinants within countries the same as between countries. In: Adler, N. E., Marmot, M., McEven, B. S. & Stewart, J. (eds.). Socioeconomic status and health in Industrial nations: social, psychological, and biological pathways. New York, Annals of the New York Academy of Sciences. Vol. 896, 16-29.

Marmot, M. & Wilkinson, R.G. (2001). Psychosocial and material pathways in the relation between income and health: a response to Lynch et al. BMJ, 322, 1233-1236.

Mellor, J. M. & Milyo J. (2001). Reexamining the evidence of an association between income inequality and health., J. of Ssocial Politics, Policy and Law 26(3), 487-521.

Muntaner, C. & Lynch, J. (1999). Income inequality, social cohesion, and class relations: a critiques of Wilkinson's neo-Durkheimian research program. International Journal of Health Services, 29(1), 59-81.

Platt, S. (1998). A medical sociology perspective, Journal of Community & Applied Social Psychology, 8, 170-172.

Putnam R. (2000). Bowling alone: The collapse and revival of American community. New York: Simon and Schuster.

Putnam, R. D. (1993). Making Democracy Work. Princeton: Princeton University Press.

Robert, S. A. & House, J. S. (2000). Socioeconomic inequalities in health: integrating individual–, community–, and societal level theory and research. In Albrech G., Fitzpatric, R. & Scrimshaw, S. (eds.), Handbook of social studies in health and medicine. Ch. 18.1, 115-135.

Robert, S. A. (1999). Socioeconomic position and health : the independent contribution of community socioeconomic context. Ann. Rev. Of Sociology 25, 489-516.

Rokeach, M. (1973). The Nature of Human Values. NY: The free Press

Rosner, M., & Getz, S. (1994). Towards a theory of changes in the kibbutz. Journal of Rural Cooperation, 22 (1-2), 41-62.

Schnittker, J. & McLeod, J.D. (2005). The social psychology of health disparities, Ann.Rev. of Socio., 31, 75-103.

Singer, B. & Ryff, C. D. (1999). Hierarchies of life histories and associated health risks In: Adler, N. E., Marmot, M., McEven, B. S. & Stewart, J. (eds.). Socioeconomic status and health in Industrial nations: social, psychological, and biological pathways. New York, Annals of the New York Academy of Sciences. Vol. 896, 96-115.

Shwartz, S. (1994). Structure of human values. J. of Social Issues, 50 (4), 18-45.

Stewart, J.A. (2006). The detrimental effects of allostasis: allostatic load as a measure of cumulative stress. J Physiol Anthropol, 25, 133-145.

Tannenbaum, A.l., Rosner, M., Kavic, B., Vianello, M., & Weise, G. (1974). Hierarch in Organization. San Francisco: Jossey Bass.

Wilkinson R.G. (1996). Unhealthy societies: the afflictions of inequality. London: Routledge.

Wilkinson, R. G .(1992). Income distribution and life expectancy. British Medical Journal, 304, 165-172.

Wilkinson, R. G. & Picket, K. E. (2006). Income inequality and population health a review and explanation of the evidence. Social Science and Medicine, 62, 1768-84.

Yukl, G. (2002). Leadership in organizations (5^th ed.). Upper Saddle River, NJ: Prentice Hall.

Appendix 1: Comparison, by analysis of variance and Pearson correlation coefficients, of four groups of kibbutzim on four measures of economic inequality

Expressions of Physical Social Capital	Budgetary arrangement	Number of kibbutzim	Mean	SD	F & (r)
Evaluation of extent of economic differences among members in kibbutz	Traditional	11	3.21	.53	4.52** (.54**)
	1-2 yrs. Differ.	5	3.52	.16
	3-4 yrs. Differ.	8	3.79	.35
	5-6 yrs. Differ.	8	3.78	.28
	Total	32	3.54	.46
Satisfaction with degree of equality in kibbutz	Traditional	11	3.15	.54	5.46** (-.58**)
	1-2 yrs. Differ.	5	2.78	.27
	3-4 yrs. Differ.	8	2.45	.32
	5-6 yrs. Differ.	8	2.44	.48
	Total	32	2.74	.53
Mean estimated relative level of family income compared to other members	Traditional	11	3.18	.25	5.83** (-.59**)
	1-2 yrs. Differ.	5	3.02	.15
	3-4 yrs. Differ.	8	2.78	.22
	5-6 yrs. Differ.	8	2.80	.28
	Total	32	2.96	.29
SD of perceived family monthly income – compared to other members in kibbutz	Traditional	11	.65	.17	2.89* (.45**)
	1-2 yrs. Differ.	5	.77	.25
	3-4 yrs. Differ.	8	.85	.14
	5-6 yrs. Differ.	8	.84	.12
	Total	32	.76	.18

Appendix 2: Comparison, by analysis of variance and Pearson correlation coefficients, of four groups of kibbutzim on six measures of PHSC

Expressions of Physical Social Capital	Budgetary arrangement	Number of kibbutzim	Mean	SD	F & (r)
1.Level of formal education by number of years 1=-12; 2=13-14; 3=15-16; 4=17+	Traditional	11	2.33	.16	3.05* (.-39*)
	1-2 yrs. Differ.	5	2.48	.15
	3-4 yrs. Differ.	8	2.08	.35
	5-6 yrs. Differ.	8	2.08	.39
	Total	32	2.23	.31
2. Kibbutz commitment to create appropriate jobs for its members (in general and for older members) (2 items)	Traditional	11	3.18	.57	9.21** (-.70***)
	1-2 yrs. Differ.	5	2.77	.28
	3-4 yrs. Differ.	8	2.42	.38
	5-6 yrs. Differ.	8	2.24	.24
	Total	32	2.68	.56
3.Percent of members reporting being unemployed	Traditional	11	8.96	8.54	2.98* (.49**)
	1-2 yrs. Differ.	5	16.83	13.87
	3-4 yrs. Differ.	8	22.77	18.55
	5-6 yrs. Differ.	8	23.56	11.43
	Total	32	17.29	13.23
4.Satisfaction with kibbutz institutions' response to personal unique demands and needs	Traditional	11	3.3.9	.36	4.77** (-.58**)
	1-2 yrs. Differ.	5	3.29	.32
	3-4 yrs. Differ.	8	3.00	.32
	5-6 yrs. Differ.	8	2.87	.22
	Total	32	3.31	.37
5. Average amount in NIS a respondent pays privately for medications	Traditional	11	14.19	11.54	22.26** (.66***)
	1-2 yrs. Differ.	5	71.57	44.89
	3-4 yrs. Differ.	8	171.65	65.99
	5-6 yrs. Differ.	8	109.99	39.41
	Total	32	86.47	74.44
6.Satisfaction with health services on kibbutz (an index of 16 items)	Traditional	11	3.91	.26	7.11** (-56**)
	1-2 yrs. Differ.	5	3.54	.35
	3-4 yrs. Differ.	8	3.21	.40
	5-6 yrs. Differ.	8	3.39	.37
	Total	32	3.55	.43

Appendix 3: Analysis of variance (F) and Pearson coefficients of correlations [r] to relate 12 single expressions of psychosocial social capital with length of time adopting the DSA. Four groups of kibbutzim.

Social Capital :	Budgetary arrangement	Number of kibbutzim	Mean	SD	F ( r)
7.Satisfaction with home feeling and belonging to kibbutz	Traditional	11	3.82	41.	3.05* (-.41*)
	1-2 yrs. Differ.	5	3.56	22.
	3-4 yrs. Differ.	8	3.34	31.
	5-6 yrs. Differ.	8	3.48	38.
	Total	32	3.58	39.
8.Improvement (5) or deterioration(1) in social situation of kibbutz compared to two years earlier	Traditional	11	2.70	.43	3.79* (-.53**)
	1-2 yrs. Differ.	5	2.47	12.
	3-4 yrs. Differ.	8	2.33	23.
	5-6 yrs. Differ.	8	2.22	.31
	Total	32	2.45	.36
9. Frequency of participation in social and cultural events on kibbutz (5- very frequent). (7 items)	Traditional	11	2.75	.21	11.01** (-.70**)
	1-2 yrs. Differ.	5	2.66	.22
	3-4 yrs. Differ.	8	2.18	.33
	5-6 yrs. Differ.	8	2.15	.32
	Total	32	2.44	.38
10.Satisfaction with social relations on kibbutz ( 4 items)	Traditional	11	3.31	.43	6.29** (-.61***)
	1-2 yrs. Differ.	5	3.03	.21
	3-4 yrs. Differ.	8	2.73	.16
	5-6 yrs. Differ.	8	2.72	.40
	Total	32	2.97	.42
11.Satisfaction with influence and information one has in kibbutz (4 items)	Traditional	11	3.09	.35	4.65** (-.48**)
	1-2 yrs. Differ.	5	3.03	.10
	3-4 yrs. Differ.	8	2.61	.33
	5-6 yrs. Differ.	8	2.76	.26
	Total	32	2.87	.35
12.Social support on kibbutz (4 items)	Traditional	11	3.44	.56	5.36** (-.63**)
	1-2 yrs. Differ.	5	3.22	.27
	3-4 yrs. Differ.	8	2.65	.24
	5-6 yrs. Differ.	8	2.90	.57
	Total	32	3.10	.54
13.Frequency of organized social events on kibbutz (4 items)	Traditional	11	2.38	.38	3.99* (-.45**)
	1-2 yrs. Differ.	5	2.16	.28
	3-4 yrs. Differ.	8	1.79	.46
	5-6 yrs. Differ.	8	1.98	.33
	Total	32	2.10	.43
14.Frequency of meeting other members at work and the dining hall (2 items)	Traditional	11	3.56	.78	7.28** (-.77**)
	1-2 yrs. Differ.	5	3.42	.31
	3-4 yrs. Differ.	8	2.53	.29
	5-6 yrs. Differ.	8	2.74	.50
	Total	32	3.12	.70
15.Trust in kibbutz leadership (2 items)	Traditional	11	3.27	.49	4.32* (-.46**)
	1-2 yrs. Differ.	5	3.17	.40
	3-4 yrs. Differ.	8	2.68	.39
	5-6 yrs. Differ.	8	2.83	.25
	Total	32	3.02	.46
16.Trust in other members of kibbutz	Traditional	11	3.33	.28	4.90** (-.54**)
	1-2 yrs. Differ.	5	3.16	.26
	3-4 yrs. Differ.	8	2.89	.26
	5-6 yrs. Differ.	8	2.99	.30
	Total	32	3.10	.33
17.Ways of decision making on kibbutz (5- most members participate; 1-just few make all decisions	Traditional	11	2.75	.47	5.13** (-.51**)
	1-2 yrs. Differ.	5	2.60	.42
	3-4 yrs. Differ.	8	2.12	.32
	5-6 yrs. Differ.	8	2.29	.15
	Total	32	2.45	.44

---

This study is part of a larger research project conducted by the Institute for Social Research of the Kibbutz and partly supported by "The National Institute for Health Policy and Health Services"

Dr. Gila Adar and Jana Goldemberg of the Institute for Social Research of the Kibbutz and Michal Rotem of the Department of Sociology consisted the research team for the larger project and contributed to this study