Intergenerational family relations and subjective well-being in old age: a cross-national study

Abstract

This study addresses the links between different dimensions of intergenerational family relations (solidarity, conflict, and ambivalence) and subjective well-being (life satisfaction, and positive and negative affects) of older people (aged 75+) using a comparative perspective from a random urban sample in five countries (Norway, England, Germany, Spain, and Israel). Comparative descriptive analyses and multivariate models are presented. Three general conclusions can be drawn. First, the country variables contributed significantly to the explained variance for all three components of well-being, where Israel was most different from the other four countries, except for positive affects, for which Spain and Israel differed from Norway, England, and Germany. However, the countries explained more of the variance for positive and negative affects compared with life satisfaction. Second, the intergenerational family solidarity dimensions contributed significantly to the explained variance for all three components of well-being, although they were differentially related to the different aspects of subjective well-being. Third, personal resources, mainly physical functioning and financial adequacy, were related to all of the well-being variables, although their relative contribution was much stronger for life satisfaction. The importance of intergenerational family relations and personal resources for the subjective well-being of older people and the importance of using multiple measures for outcome variables of well-being are discussed.

Introduction

This article examines the links between intergenerational family relations and subjective well-being of older parents studied from a comparative cross-national perspective: Norway, England, Germany, Spain, and Israel. As we witness the phenomenon of global ageing, concern about well-being of older people and its antecedents has heightened. The focus here is on three components that compose subjective well-being: life satisfaction and positive and negative affects of older parents and on three dimensions of family relations—solidarity, conflict, and ambivalence—emotional qualifications of family ties.

In light of growing ageing populations and changes in family structures, norms, and behaviors in European countries (OECD 1996), the relations between intergenerational family relations on the well-being of older parents takes on added significance. Using data from the OASIS study1 (Old age and autonomy: the role of service systems and intergenerational solidarity) of the group of adults aged 75 and older, this study investigates how various dimensions of intergenerational family solidarity, conflict, and ambivalence in older parent–child relationships are associated with well-being of the older parents in the five countries studied. Specifically, two research questions will be addressed: first, what the similarities and differences are between the five countries in terms of strength of solidarity, conflict, and ambivalence; and second, to what extent specific cultural and interpersonal family relations contribute to the explanation of the variance in subjective well-being of elders?

Families represent the primary and perhaps most crucial context within which human development takes place. Given their central role in individuals’ lives, intergenerational family processes may condition the effects of various exogenous challenges on individuals, thereby reducing or magnifying the effects of such events on individuals’ well-being. Families have been shown to buffer the effects of potentially difficult life changes and crisis-like events (Eggebeen and Davey 1998). Understanding the relationship between family relations and well-being for coping and social integration in old age is extremely valuable (Silverstein and Bengtson 1991). Increased life expectancy and decreased family size raise the importance of intergenerational ties between older parents and their adult children, especially when older parents are in need of care (Davey et al. 2005). Families today are experiencing new relationships with regard to duration and intensity of care for their older parents.

The links between intergenerational family relations and various aspects of older parents’ well-being has been studied mostly in American society (Krause et al. 1992; Lee et al. 1995; Silverstein and Bengtson 1994; Silverstein et al. 1996; Umberson 1992). An additional body of research compares intergenerational family relations between the United States and some other countries but without relating the comparisons to well-being (e.g., Bengtson and Martin 2001, comparison with German-speaking countries; Silverstein et al. 1996, comparison with Wales).

However, several European cross-national studies on this topic were recently conducted, e.g., the SHARE and the VOC research projects. The first, a 15 country Survey of Health, Aging and Retirement in Europe (SHARE) was developed in order to better understand the status of persons aged 50 and over in Europe (Boersch-Supan and Juerges 2005). The second, Value of Children and Intergenerational Relations (VOC), a cross-cultural study (Trommsdorff and Nauck 2005), was developed to study the interplay of values (general and child related) and parent–child relationships in three generations and across cultures.

Some articles have already been published using OASIS data to elaborate on the links between family relations and quality of life (Katz and Lowenstein 2003), and reciprocity in parent–child exchange and life satisfaction (Lowenstein et al. 2007). However, this paper is focused on both the affective and the cognitive components of subjective well-being, and on the wide range of intergenerational family relations, including ambivalence, which will enrich and add knowledge to the existing OASIS publications.

The study deals with cross-cultural and cross-national comparisons of five countries that belong to both different and similar welfare policy and services to older people and family traditions. A comparative approach with a strategic choice of countries may add insights that single country studies lack. For welfare-state regimes the well-known typologies of Esping-Andersen (1990, 1999) were adopted: social democratic (Norway), conservative (Germany and Spain), liberal (England), with Israel as a “mixed” model. Other indicators were also included. One of these was family legislation, that is, whether adult children have any legal responsibility towards their older parents: those in Norway and England had no legal obligation to help their older parents, whereas those in the other three countries did. A second indicator was the role of the state, expressed, for example, by the level of community services provided, in particular home-based services that might alleviate the burden on families: the level of services was high in Norway and Israel, medium in England and Germany, and low in Spain.

Regarding family culture several indicators were used. Among them women’s labor-force participation (high in Norway; medium in England, Germany, and Israel; and low in Spain but gradually increasing), which reflects the role of families and of women in the family. Another indicator was fertility rate, which shows that Israel has a much higher rate than the four European countries, explained by Jewish and Arab family traditions (Katz and Lavee 2005). As families are important source of informal support, living arrangements (percentages of elderly parents co-residing with their children) were considered indicator of family culture. Substantial differences between the countries were found where in Spain about a third of the respondents co-reside with children and much less in Norway, Israel and Germany (5, 6 and 7%, respectively). Furthermore, the above mentioned indicator of legislation on family obligation to provide care can also be viewed as reflecting family norms and culture.

The five nations in the OASIS project thus differ in their cultural and social contexts, differences that yield fruitful ground for examining the associations between family relations and subjective well-being.

Intergenerational family relations and subjective well-being

Intergenerational relationships were generally found to contribute to subjective well-being of individuals throughout their life course (Rossi and Rossi 1990), and maintaining high level of subjective well-being is considered to be an important aspect of successful ageing (Freund and Baltes 1998).

Subjective well-being refers to evaluations that people make about their lives (Shmotkin 2005), and is defined as a broad concept comprising a wide range of distinct dimensions (Kunzmann et al. 2000). As such there are different approaches to the meaning and to the measurement of this construct. One approach distinguished between positive and negative affects and defined it as the balance between the two (Bradburn 1969). Another approach is that well-being represents judgments of one’s life and reflects one’s relative dominance of positive versus negative affect (Diener 1994). According to this terminology (Diener et al. 1999), subjective well-being is a construct consisting of two components: the cognitive (satisfaction judgments) and the emotional (affective reactions). An individual with a desirable life is satisfied and experiences frequent pleasant (positive affect) emotions and infrequent unpleasant (negative affect) emotions.

Ryff posits yet, another approach to the construct of well-being and generated a multidimensional model including six distinct components of positive psychological functioning like personal growth, purpose in life and self-acceptance (Ryff 1995). We followed the conventional structure of assessing subjective well-being through life satisfaction along with the positive and negative affects. By using this approach, we followed a long tradition, including in gerontology, of seeing subjective well-being as a rough and parsimonious indicator of adjustment and mental health. In a meta-analysis of 286 empirical studies on the association of SES, social network and competence with subjective well-being in the elderly the three most often used measures were life satisfaction index, single items that assessed overall life satisfaction and affect balance scale (Pinquart and Sorensen (2000).

The conceptual framework of intergenerational solidarity represents one of several enduring attempts in family sociology to examine and develop a theory of family cohesion (Katz et al. 2005). The intergenerational solidarity framework perceives the relationships between older parents and their adult children as a primary source of mutual emotional and instrumental support. Since the early 1970s, Bengtson and colleagues have continued to develop and expand this model within the Longitudinal Study of Generations (LSOG; Bengtson and Harootyan 1994; Silverstein and Bengtson 1997). Beginning in 1985, Bengtson and others have incorporated conflict into their study, arguing that as a normative aspect of these relations it is likely to affect the perception of these relationships and the willingness of family members to assist each other.

The term “ambivalence,” reflecting contradictions and ambiguities in relationships, was introduced by Luescher and Pillemer (1998) as a valuable revived conceptual perspective for studying parent–child relations in later life. They argue that intergenerational ambivalence should be the primary topic of study of intergenerational relations that cannot be reconciled in adulthood (Pillemer and Luescher 2004). Bengtson et al. (2002) though questioned the utility of the ambivalence concept, concluded that it complemented rather than competed with the solidarity–conflict framework mainly for exploring mixed feelings. Confronting the two paradigms of solidarity–conflict and ambivalence and their impact on quality of life of older parents within the OASIS study shows that conflict had no effect and that ambivalence had a much lesser effect than the solidarity dimensions (Lowenstein 2007).

Quality of family relations, including intergenerational relations, were reported in many studies as having strong associations with subjective well-being (Fernandez-Ballesteros et al. 2001; Pinquart and Sorensen 2000). In contrast, McCamish-Svensson et al. (1999) found that neither child nor friend support was related to life satisfaction in their study of older people aged 80 and over in Sweden. Other studies indicate that support from adult children is psychologically beneficial at moderate levels, but may be harmful if it is too frequent or too infrequent (Lee et al. 1996; Silverstein et al. 1996).

The present study contributes to the body of research on the associations between family relations and well-being by comparing several dimensions of subjective well-being and several dimensions of intergenerational family relations among five countries that differ in their cultural and social contexts, particularly, family traditions and welfare development.

It is expected that a significant effect of cross-cultural differences on subjective well-being and a significant effect of intergenerational family solidarity on well-being beyond cross-cultural differences will be found. Specifically, stronger connections between family solidarity and subjective well-being will be found in the more traditional countries regarding family norms and patterns of behavior. Also, in the more welfare developed countries weaker associations will be found between family relations and subjective well-being because services are available through public support system.

Method

Sampling and data collection

Data collection was based on face-to-face structured interviews with a random representative urban sample of 1,200 respondents (800 aged 25–74 and 400 aged 75+) in each of the five participating countries, totaling 6,000 respondents. In the present paper, the analysis is based on the reports of respondents aged 75 years and older who have adult children over 21 years old, living in separate households in the community. Overall response rate in all countries varied from 70 to 76%. A complete account of the OASIS model, design, and methodology is available in Lowenstein and Ogg (2003).

Description of the sample

Data in Table 1 present the comparative distribution of the background variables. The average age of respondents in the five countries is around 80 years. In all countries women outnumbered men. A higher percentage of respondents in Spain than in the other four countries were married. In Israel and in Spain, parents had on average more living adult children. Older Spaniards were the least educated. Financial adequacy shows substantial differences, with the highest level in Germany and the lowest in Spain. The percentage of people reporting a high level of functioning (81–100 points) was highest in Norway (44%) and lowest in Israel (14%). Large differences were found in living arrangements between the countries. In Spain almost a third of respondents were living with other family members, followed by England (13%). In Germany, Israel, and Norway the percentage was rather small.

Table 1.

Background characteristics of older parents (aged 75+) in five countries (N = 1,914)

	Norway	England	Germany	Spain	Israel
Age mean (SD)	81.5 (4.8)	82.3 (5.3)	81.3 (5.1)	80.4 (4.3)	80.0 (4.3)
Sex (% male)	40	32	31	35	40
Marital status (% married)	35	36	36	39	35
Number of children, mean (SD)	2.3 (1.1)	2.2 (1.3)	1.9 (1.2)	2.6 (1.4)	2.7 (1.5)
Education
Primary level (%)	30	25	13	81	38
Secondary level (%)	34	62	63	15	36
Higher (%)	36	13	24	4	26
Financial adequacy(% comfortable)	59	52	68	28	50
Physical functioning (SF36)
Low: 0–40 points	21	44	29	31	39
Intermediate: 41–80 points	35	34	47	46	47
High: 81–100 points	44	22	24	23	14
Living arrangements (% co-residence)	5	13	7	31	6
Total N	398	378	410	370	358

The comparison across samples reveals both similarities and differences in background attributes between the five countries. Therefore, in comparing intergenerational relations and subjective well-being across the five samples, we estimated equations with and without controlling for background attributes.

Measures

The variables used in this analysis are organized into three groups: outcome variables, socio-demographic variables, and intergenerational family relations variables.

Outcome variables

The outcome variable in this study is subjective well-being, using three self-reported measures: the cognitive dimension (life satisfaction) that was found to be associated with stable, long-term assessments of well-being (Chamberlain 1988); and the emotional component (positive and negative affects) that refers to the experience of pleasant or unpleasant emotions and moods (e.g., happiness). It has been empirically demonstrated that positive and negative emotions are not opposite poles of one underlying dimension but are instead two independent dimensions (Diener 1994).

Life satisfaction was measured using a global question about overall life satisfaction in the preceding 2 weeks and 13 domain-specific questions about satisfaction with living conditions, access to health services, relations with friends, oneself etc., yielding an alpha of 0.78. Response categories, on a Likert-type scale, ranged from 1 (very dissatisfied) to 5 (very satisfied).

Positive and negative affects were measured with the Positive and Negative Affect Schedule (PANAS) (Watson et al. 1988), which contains two subscales of 10 items each, one measuring positive affect and the other negative. The scales are highly consistent internally, largely uncorrelated, and stable at appropriate levels over a 2-month period. We used a 10-item version of the standard 20-item PANAS, adapted by Kercher (1992). This shorter version was found to have an appropriate factor structure, high discriminant validity, and reasonable reliability for its subscales (Hilleras et al. 1998). The positive affect contained five items describing positive feelings: excited, enthusiastic, alert, inspired, and determined. The negative affect contained five items describing negative feelings: distressed, upset, scared, nervous, and afraid. Alpha scores were 0.86 and 0.87, respectively. Respondents rated each item on a 5-point Likert-type scale from 1 (not at all) to 5 (an extreme amount) in the preceding 2 weeks; total scores ranged from 5 to 25.

Socio-demographic variables

Background attributes include the following: sex (1 = female, 0 = male); marital status of older parents (1 = married or unmarried cohabiting, 0 = not married or cohabiting); number of living adult children (older than 21 years); education, measured by the highest level attained on a 3-point scale (primary, secondary, and higher); financial adequacy, measured by perceived current financial situation (1 = comfortable, 0 = not comfortable); and ADL physical functioning, measured by the shortened version of the SF-36 with 12 items (Ware and Sherbourne 1992); the total score of the scale ranges from 1 to 100, a higher score indicating better functioning. These attributes were included because they were found to be associated with well-being among the elderly (McNeil et al. 1986; Pinquart and Sorensen 2000).

Intergenerational relations variables

Intergenerational solidarity was defined as a multidimensional structure that reflects behavioral, affectual, cognitive, and structural components of family relations (Bengtson and Schrader 1982). Solidarity items were selected from the Longitudinal Study of Generations (LSOG). The instrument contains 54 items related to the respondents’ children along the following dimensions of solidarity: (1) Proximity: the geographic distance that might constrain or facilitate interaction, measured on a 6-point scale from 1 (living 3 or more hours’ traveling distance away) to 6 (living together). (2) Association: frequency of face-to-face contact, coded as 1 (several times a year) to 6 (daily or more often). (3) Affect: feelings of emotional intimacy between family members, using three questions such as “How close do you feel to (this child)”? The questions were coded from 1 (not at all) to 6 (extremely). (4) Consensus: degree of similarity in opinions and values, coded as 1 (not at all similar) to 6 (extremely similar). (5) Functional: instrumental assistance operationalized as receiving or providing help from/to at least one child in the areas of shopping and transportation, household chores, house repair and gardening, personal care/child care, financial assistance, and emotional support. Factor analysis performed for all countries (pooled samples) revealed a two-factor structure. The first element includes the structural–behavioral dimension (proximity and associational), labeled Solidarity S, with factor loadings of 0.90 and 0.87. The second element includes the affective–cognitive dimension (affectual and consensus), labeled Solidarity A, with factor loadings of 0.75 and 0.98. Instrumental assistance (help provided and received) did not emerge in the factor structure and was thus used separately in the analyses.

Conflict was measured by 3 items related to the degree of conflict or tension, criticism, and arguments between the generations, coded as 1 (none at all) to 6 (a great deal). A mean score was used in the analyses. Internal consistency of the conflict measure was 0.77.

Ambivalence was measured by three items based on those designed by Luescher and Pillemer (1998). Respondents were asked, for example, “Sometimes family members can have mixed feelings in their relationships. Thinking about your relationships with your child, how often do you have such mixed feelings?” The responses ranged from 1 (never) to 6 (very often). Internal consistency for the ambivalence measure was 0.70.

Data analysis

The present analysis was carried out in two stages. In the first stage, descriptive statistics of intergenerational solidarity and subjective well-being (life satisfaction and positive and negative affects) were carried out from a comparative perspective. In the second stage, three multiple regressions were constructed on the outcome variables to examine which of the three blocks of variables had the strongest impact on well-being: countries, family solidarity–conflict dimensions and ambivalence, and the background attributes.

Results

Comparing the intergenerational family dimensions in the five countries

Table 2 presents the distribution of the solidarity–conflict and ambivalence dimensions in a comparative cross-national perspective. The data show that the level of Solidarity S (structural–behavioral) was similar in four of the countries (mean 3.6–3.8), and was highest in Spain (4.5). Duncan multiple-range tests on Solidarity S show that Spain was significantly different from the four other countries. Solidarity A (affective–cognitive) was high in all countries (mean 4.2–4.7 on a 6-point scale), and was highest in Israel (4.7). For Solidarity A, Israel was by itself, England and Norway formed the middle group, and Germany and Spain formed the lowest group. The exchange of help was relatively low: Spain was highest (mean 2.1) for help received from adult children, and Norway and Israel were lowest (mean 1.4). Duncan tests showed that Germany, England, and Spain were grouped together for more help received, and Norway and Israel formed a second group. No difference between the countries was found for help provided (mean 0.8–1.0).

Table 2.

Means and standard deviations of intergenerational family relations

Intergenerational family relations	Country [means (SD)]
	Norway	England	Germany	Spain	Israel
Solidarity S (proxi. + association)	3.6 (1.4)	3.8 (1.4)	3.7 (1.4)	4.5 (1.3)	3.8 (1.2)
Solidarity A (affect + consensus)	4.4 (0.9)	4.5 (1.0)	4.2 (0.9)	4.2 (0.8)	4.7 (0.9)
Help receiveda	1.4 (1.3)	1.9 (1.7)	1.9 (1.6)	2.1 (1.7)	1.4 (1.4)
Help provideda	1.0 (1.1)	0.9 (1.2)	0.9 (1.2)	0.8 (1.2)	0.8 (1.0)
Conflict	1.4 (0.7)	1.3 (0.6)	1.6 (0.7)	1.4 (0.7)	1.7 (0.8)
Ambivalence	1.7 (0.03)	1.3 (0.04)	1.6 (0.03)	1.5 (0.04)	1.5 (0.03)
Total N	398	378	410	370	358

Source: Lowenstein (2007). Mean scores on a scale of 1–6, with 6 indicating high feelings of solidarity, conflict, or ambivalence

^aReceiving or providing help from or to at least one child, in at least one of the following areas: shopping, transportation, household chores, house repair and gardening, and personal care

Levels of conflict were low in all countries (mean 1.3–1.7, with a score of 6 indicating high levels). Duncan tests for conflict show that two groups were formed: Israel and Germany; and Spain, Norway, and England. Levels of ambivalence were also low (mean 1.3–1.7, with a score of 6 indicating high levels). Duncan tests for ambivalence show that three groups were formed: Norway and Germany formed the highest group, followed by Spain and Israel, and England the lowest.

Table 3 presents correlations between the solidarity dimensions, conflict, and ambivalence. Inter-correlations between the six components were low-to-moderate (ranging from 0.018 to 0.436). Though most of them were significant, they can be used as independent variables in the regression models.

Table 3.

Correlations of intergenerational family relations

	Solidarity S	Solidarity A	Help received	Help provided	Conflict	Ambivalence
Solidarity S		0.149**	0.284***	0.125**	−0.033	−0.070**
Solidarity A			0.134**	0.098**	−0.358***	−0.369***
Help received				0.287***	−0.029	−0.031
Help provided					0.018	0.047*
Conflict						0.436***

* P < 0.05, **P < 0.01, ***P < 0.001

Comparing well-being in the five countries

Table 4 presents the means and standard deviations of life satisfaction and positive and negative affects in the five countries. Life satisfaction was medium to high in all countries (mean 3.3–3.8). Duncan tests for life satisfaction show that three groups were formed: Norway (M = 3.8), England (M = 3.7), and Germany (M = 3.7) showing the highest life satisfaction; Spain in the middle (M = 3.5); and Israel the lowest (M = 3.3). As for positive affect, Duncan tests showed that the German respondents formed one group (M = 2.9), followed by England (M = 2.3) and Norway (M = 2.3), with Israel (M = 2.1) and Spain (M = 2.0) forming the group with the lowest positive affect. For negative affect, three groups were revealed, with the Israeli respondents forming one group (M = 2.3); Spain (M = 1.8), Germany (M = 1.7), and England (M = 1.7) the second; and Norway (M = 1.6) the third.

Table 4.

Means and standard deviations of well-being measures

Well-being	Norway [means (SD)]	England [means (SD)]	Germany [means (SD)]	Spain [means (SD)]	Israel [means (SD)]
Life satisfaction	3.8 (0.03)	3.7 (0.03)	3.7 (0.03)	3.5 (0.03)	3.3 (0.03)
Positive affect	2.3 (0.04)	2.3 (0.05)	2.9 (0.04)	2.0 (0.05)	2.1 (0.04)
Negative affect	1.6 (0.04)	1.7 (0.04)	1.7 (0.04)	1.8 (0.04)	2.3 (0.04)
Total N	398	377	410	370	358

Life satisfaction scale was: 1 = very poor to 5 = very good; positive and negative affect scores were: 1 = not at all to 5 = an extreme amount

The correlations between positive emotions and negative emotions (0.136) and between positive emotions and life satisfaction (0.264) were rather small. The correlation between negative emotions and life satisfaction was medium (−0.404).

To study the associations between countries, intergenerational family relationships, and personal attributes and the three variables of subjective well-being (life satisfaction, positive and negative affects), three block-recursive regressions were performed with three models. Country variables were entered into the equations first (Israel as the reference and dummy-coded the remaining countries accordingly), followed by a second model in which the family relations variables were added, and finally the background attributes were entered.

Life satisfaction, countries, intergenerational family relations, and background attributes

Table 5 presents the results of the first multiple-regression analysis. The regression data indicate that the overall explained variance of life satisfaction by countries, intergenerational concepts, and background attributes was 45%. The most powerful predictors of life satisfaction are background attributes (i.e., physical functioning, financial adequacy, and level of education, all higher), which contributed 30% to the explained variance. Country variables contributed another 9%. Model 1, which contained the country variables only, shows that Norway, England, Germany, and Spain differed from Israel, where respondents reported lower levels of life satisfaction. When the intergenerational relations variables were added, they accounted for another 6% of the explained variance, with Solidarity A factor and help provided showing positive relationships with life satisfaction, whereas help received and ambivalence showed negative relationships. Conflict and Solidarity S were not significant predictors. When personal variables were accounted for, only two of the intergenerational family relations dimensions remained significant: Solidarity A as a positive predictor of life satisfaction, and ambivalence as a negative predictor.

Table 5.

Standardized regression coefficients for life satisfaction, including countries, intergenerational family relations and background attributes (N = 1,913)

	Model 1	Model 2	Model 3
Norway	0.336***	0.362***	0.202***
England	0.287***	0.297***	0.274***
Germany	0.276***	0.340***	0.202***
Spain	0.152***	0.187***	0.212***
Solidarity S (proxi. + association)		0.015	0.013
Solidarity A (affect + consensus)		0.135***	0.078***
Help received		−0.133***	0.021
Help provided		−0.123***	0.035
Conflict		−0.015	0.002
Ambivalence		−0.129***	−0.083***
Sex			0.012
Age			0.002
Marital status			−0.002
No. of children			0.018
Education			0.049*
Financial adequacy			0.301***
Physical functioning			0.453***
Total R 2	0.089	0.147	0.446
R 2 change		0.063	0.300

Countries as dummies with Israel as the reference

* P < 0.05, **P < 0.01, ***P < 0.001

Positive affect, countries, intergenerational family relations, and background attributes

Table 6 presents the results of the second multiple-regression analysis. The regression data indicate that the overall explained variance of positive affect by countries, intergenerational dimensions, and background attributes was 21%. The most powerful predictors of positive affect are country of origin, which explained 13% of the variance. Model 1, which contained the country variables only, shows that Norway, England, and Germany differed from Israel, whereas Spain was similar to it. In other words, respondents in Norway, England, and Germany had higher levels of positive affect than did those in Israel and Spain. When the intergenerational relations variables were added, they accounted for another 5% of the explained variance, with Solidarity A factor and help provided showing positive relationships with positive affect, whereas conflict showed negative relationships. These intergenerational family relations dimensions remained significant when personal variables were accounted for (model 3). All of the background attributes, except sex, were significantly associated with positive affect and explained 4% of the variance. They included age (younger), marital status (married or cohabiting), number of children, education, financial adequacy, and physical functioning (all higher).

Table 6.

Standardized regression coefficients for positive affect, including countries, intergenerational family relations and background attributes (N = 1,908)

	Model 1	Model 2	Model 3
Norway	0.079**	0.119***	0.097**
England	0.095**	0.142***	0.158***
Germany	0.379***	0.437***	0.427***
Spain	−0.042	0.047	0.074*
Solidarity S (proxi. + association)		−0.042	−0.035
Solidarity A (affect + consensus)		0.200***	0.183***
Help received		−0.045	−0.018
Help provided		0.106***	−0.084***
Conflict		−0.124***	−0.114***
Ambivalence		0.030	0.032
Sex			0.027
Age			−0.078**
Marital Status			−0.070**
No. of children			0.088***
Education			0.086***
Financial adequacy			0.065*
Physical functioning			0.093***
Total R 2	0.133	0.176	0.209
R 2 change		0.046	0.037

Countries as dummies with Israel as the reference

* P < 0.05, **P < 0.01, ***P < 0.001

Negative affect, countries, intergenerational family relations, and background attributes

Table 7 presents the results of the third multiple-regression analysis. The regression data indicate that the overall explained variance of negative affect by countries, intergenerational dimensions, and background attributes was 20%. The most powerful predictors of negative affect are country of origin, which explained 8% of the variance. Model 1, which contained the country variables only, shows that Norway, England, Germany, and Spain differed from Israel. In other words, in all four countries the level of negative affect is lower than in Israel. When the intergenerational relations variables were added, they accounted for another 5% of the explained variance, with conflict and ambivalence showing positive relationships with negative affect. These intergenerational family relations dimensions remained significant when personal variables were accounted for (model 3). As for background attributes, financial adequacy and physical functioning were negatively related to negative affect, and sex (female) was positively related to it.

Table 7.

Standardized regression coefficients for negative affect, including countries, intergenerational family relations and background attributes (N = 1,911)

	Model 1	Model 2	Model 3
Norway	−0.340***	−0.324***	−0.248***
England	−0.253***	−0.215***	−0.212***
Germany	−0.276***	−0.277***	−0.220***
Spain	−0.242***	−0.215***	−0.227***
Solidarity S (proxi. + association)		0.006	0.008
Solidarity A (affect + consensus)		0.036	0.063
Help received		0.044	−0.038
Help provided		−0.044	0.003
Conflict		0.115***	0.121***
Ambivalence		0.150***	0.128***
Sex			0.066
Age			−0.031
Marital Status			−0.018
No. of children			0.026
Education			−0.030
Financial adequacy			−0.079**
Physical functioning			−0.233***
Total R 2	0.081	0.128	0.201
R 2 change		0.047	0.073

Countries as dummies with Israel as the reference

* P < 0.05, **P < 0.01, ***P < 0.001

Discussion

This study investigated similarities and differences of intergenerational family relations and subjective well-being in five countries: Norway, England, Germany, Spain, and Israel. It also explored, from a comparative cross-national perspective for the group of older people aged 75 and over, the links between family relations and three indicators of older people’ subjective well-being: life satisfaction, and positive and negative affects.

Three general conclusions can be drawn from this study. First, the country variables contributed significantly to the explained variance for all three components of subjective well-being; Israel was most different from the other four countries, except for positive affect, where Spain and Israel differed from Norway, England, and Germany. However, the countries explained more of the variance for positive and negative affects compared with life satisfaction. Second, the intergenerational family solidarity dimensions contributed significantly to the explained variance for all three components of well-being, although they were differentially related to the different aspects of subjective well-being. Third, personal resources, mainly physical functioning and financial adequacy, were related to all well-being variables, although their relative contribution was much stronger for life satisfaction.

The first conclusion relates to the comparative perspective. The participating countries represent different contexts and opportunity structures for family life and elder care. They face similar challenges in this area but are inclined towards different solutions. Germany and Spain are welfare states that tend to favor family responsibility and give the state a subsidiary role (in Germany) and even a residual role (in Spain) (Daatland and Lowenstein 2005). Both countries specify legal obligations between generations but provide relatively low levels of social services, although they provide high levels of medical services. This is reflected in the higher proportion of help received by older parents. By comparison, Norway enacts individualist social policies, imposes no legal obligations between generations, and provides high levels of social care services. The mixed Israeli model is illustrated by legal family obligations, as in Spain and Germany, with high service levels, as in Norway. Therefore, in Norway and Israel the levels of help received from children were relatively low. Higher intergenerational family solidarity was reflected differently in the more familistic countries; whereas Israel was highest in the emotional dimension of intergenerational family solidarity, Spain was the highest in proximity and contact. Some of the differences between the countries can be attributed to the specific history, development and political situation. For example, mix relations were found in Israel and in Spain. The higher rates of Solidarity A (emotional relations and consensus) as well as conflict relations found in Israel indicate contradicting trends. The family in Israel is being pulled in opposite directions by two main forces one that prods the family toward greater modernization and Westernalization, while the other acts to strengthen traditional values and norms (Katz and Lavee 2005). It is suggested that the geo-political situation is one of the salient explanations of the centrality and closeness of family relations in Israel. In Spain, Solidarity S (living arrangements and contact with children) was highest, but Solidarity A (emotional relations and consensus) was low. This mixture may be due to enduring norms of traditional respect for elders, combined with a rapid modernization process, reflected, for example, in low fertility rates (Katz et al. 2005).

The respondents aged 75 and older demonstrated relatively high levels of life satisfaction and positive affect and low levels of negative affect, except in Israel, where well-being was the lowest. This corresponds to research on subjective well-being in old age—the “stability despite loss paradox” (Kunzmann et al. 2000)—but it is not congruent with recent results on life satisfaction showing that age-related decline in life satisfaction appeared accelerated for the old-old (Schilling 2005). Results for positive and negative affects are somewhat similar for very old Swedish respondents in a study by Hilleras et al. (1998). For Israeli respondents the findings may be related to the stress of war, terrorist acts, and security threats. The regression analyses suggest that cultural context influences well-being. However, the relative contribution of the cultural context was higher for positive and negative affects. Thus, cultural variability is more relevant to general affective states, whereas life satisfaction, which is a domain-specific evaluation of subjective living situation, is more relevant to the individual level.

The second conclusion concerns the solidarity–conflict and ambivalence dimensions. The explained variance by family relations was similar for all three components of well-being; however, the relative contribution of intergenerational family relations was higher for the emotional components of well-being (positive and negative affects), consistent with findings by O’Connor (1995). Solidarity A (affectual and consensus) and help provided were the significant predictors for both life satisfaction and positive affect; conflict had an impact on negative affect. This corresponds roughly to findings showing that positive affect is positively related to social contacts (e.g., Diener 1994). Solidarity S (proximity and association) and help received did not affect any of the subjective well-being measures. However, support provided by older people was found to contribute to their well-being. Because the provision of support entails contact, the Solidarity S factor was conceivably confounded with this dimension (Lowenstein 2007). Although some studies have reported a relationship between structural–behavioral solidarity and subjective well-being, mainly during stressful events (e.g., Silverstein and Bengtson 1994), the findings are mixed.

Three main dimensions of solidarity–conflict were linked to the three variables of well-being: affectual solidarity, help provided, and conflict. It was difficult to find studies that directly examine the impact of the dimensions of family solidarity on components of subjective well-being, although some studies relate to it indirectly. Family relations, size of social networks, number of social contacts, and quality of these contacts are all reported as determinants of subjective well-being among older people (Antonucci et al. 1996; Fernandez-Ballesteros et al. 2001; Krause et al. 1992). Research on affectual solidarity revealed, for example, that it prolonged the lives of older parents who had experienced personal loss (Silverstein and Bengtson 1991). Furthermore, comparative research for the US and India showed that emotional support directly and indirectly contributed to well-being (Venkatraman 1995).

The majority of studies on informal supportive relationships focused on the benefits of receiving assistance from others to well-being and quality of life, but the effect of providing support on well-being is less studied. Krause et al. 1992 and Krause and Shaw 2000 argue that there has been surprisingly little research on the role of feelings of value and usefulness to others within more proximal social networks (e.g., family and friends) in older adults although the opportunity to provide social support to others is thought to enhance mental well-being in older adults. They found that providing informal support to others, appear to exert positive effect in old life and that helping others tends to bolster the self-esteem of old people regardless of their SES standing (Krause et al. 1992; Krause and Shaw 2000). Similar results were also obtained in non-western society, among elderly Koreans (Kim et al. 2000). Apparently, providing support to adult children and grandchildren can enhance the self-esteem of older adults and contribute to feelings of independence, which in turn affect well-being.

Conflict affected mostly negative affect but was also negatively related to positive affect. Research has consistently shown that the positive and negative dimensions of emotional well-being appear to be independent of each other and to correlate with different types of variables, suggesting that the etiologies of positive and negative affects may be different. This implies that a person who reports high positive affect can report either a high or low level of negative affect, and vice versa (e.g., Stallings et al. 1997).

Bengtson et al. (2002) saw conflict as a natural part of human life (the basic assumption of conflict theory) and as representing a separate dimension of intergenerational family relations. This perspective also emerged in the current study, where the findings confirmed the importance of including the conflict dimension when studying parent–child relations in later life. Silverstein et al. (1996) considered the possible negative effects of too much solidarity, which may cause conflict and affect well-being.

Ambivalence contributed negatively to the cognitive aspect of life satisfaction and positively to negative affect. Empirical support for ambivalence has been provided by Luescher (2004) and by Pillemer and Suitor (2002). Other scholars have attempted to measure ambivalence in parent–child relations in adulthood to provide an empirical assessment, with mixed results (Fingerman and Hay 2004; Wilson et al. 2003). Pillemer (2004) found certain correlations between subscales of ambivalence and positive affect, and the summary item correlation of ambivalence approached significance for negative affect. He concluded that there “is a possibility that intergenerational ambivalence is negatively related to psychological well-being (Pillemer 2004, p. 129).

The third conclusion related to the impact of some of the life-context variables such as physical functioning, education, and financial adequacy on subjective well-being. The impact of personal resources was different for the three dimensions of subjective well-being. The strongest impact of these variables was for life satisfaction (30 out of 45% of the variance was explained by these variables). Although family ties have generally been found to affect the subjective well-being of individuals throughout the life course (Rossi and Rossi 1990), personal resources such as socio-economic status (education, financial situation) and physical functioning, relating to basic needs, are perceived as more important than family solidarity to the well-being of old people (Aquino et al. 1996; Fernandez-Ballesteros et al. 2001; Pinquart and Sorensen 2000).

There are several potential limitations to the present analysis. First, the data refer only to groups of people aged 75 and older; second, the data is cross-sectional and show a static situation of family relations; third, conflict and ambivalent relations that were reported to have low levels may be better revealed using qualitative data collection.

In sum, the findings illuminate the importance of intergenerational family relations and personal resources for the cognitive and emotional components of subjective well-being of adults aged 75 and older. However, the different patterns of impact on the three measures of well-being used in this study show that these constructs represent different entities and confirm the importance of using multiple measures for the outcome variables. The findings show the usefulness of comparative cross-national analyses in understanding similarities and differences between them.

Policy recommendations should relate to these empirical results. The affectual dimension of intergenerational family solidarity contributed to the two components of subjective well-being (life satisfaction and positive affect), whereas conflict contributed to negative affect, and ambivalence to both negative affect and (negatively) to life satisfaction. Policy should be directed to strengthen positive and close emotional bonds between older parents and their adult children. Furthermore, the association between help provided and positive affect and life satisfaction suggests that the ability to be an active provider in exchange relations enhances these desired outcomes. In order to reach this goal, more should be invested in building, protecting, and strengthening individual resources of old people in economic, educational and health spheres.