A Closer Look at Life Goals across Adulthood: Applying A Developmental Perspective to Content,Dynamics,and Outcomes of Goal Importance and Goal Attainability

Content of goals: Age differences in goal importance and goal attainability

Conceived as motivated agents (McAdams, 2015), individuals strive to develop themselves and are inclined to expand who they are. From this humanistic perspective (Maslow, 1954; Rogers, 1995), people expand their potential by setting a motivational agenda, which moves them toward internal representations of desired future outcomes or events, represented in life goals (Austin & Vancouver, 1996). In the present study, we focused on major life goals (such as starting a family) rather than on more specific strivings (such as dating a particular person) because major life goals set the compass that directs and guides the life course (Emmons, 1986) and, accordingly, shape personality development (McAdams, 2015).

Life goals as an example of a type of characteristic adaptation (McAdams, 2015) do not emerge in a contextual vacuum. Instead, they vary with changing circumstances, role expectations, and maturation over the life course (Elder, 1995; McAdams & Pals, 2006; Roberts & Wood, 2006). Consequently, life goals represent what individuals are planning and working on while they find themselves in a certain life period (Brunstein, Dangelmayer, & Schultheiss, 1996; Cantor & Kihlstrom, 1987). According to Erikson (1968) and Havighurst (1972), each life period includes developmental tasks that arise at a certain time in personal development and that contain age–graded normative expectations, which entail specific roles, positions, and obligations, reflecting an interplay between social demands and expectations, biological development and maturation, and personality (also Freund & Baltes, 2005; Hutteman, Hennecke, Orth, Reitz, & Specht, 2014; Little, Salmela–Aro, & Phillips, 2007; Nurmi, 1992). We, consequently, argue that developmental tasks will be reflected in an age–dependent relative importance that people allocate to certain life–goal domains.

More specifically, developmental tasks in young adulthood (roughly ages 18–40 years; Staudinger & Bluck, 2001) are characterized by a focus on growth (i.e. developmental gains; Heckhausen, Dixon, & Baltes, 1989) that enables young adults to gain information and to explore who they are. This growth orientation enables young adults to acquire new skills and to reach their full potential (Ebner, Freund, & Baltes, 2006; Havighurst, 1972). Moreover, it is usually in young adulthood that people enter the workforce, establish long–lasting friendships, and commit to their first long–term romantic relationship, which makes topics of work and social relations salient for the young adult (Havighurst, 1972; Nurmi, 1992). Thus, we expect young adults’ developmental tasks to be reflected in the importance of personal–growth, social–relation, and work goals. Moreover, previous research has shown that extrinsic life goals (i.e. life goals directed at obtaining external rewards such as money, fame, and praise) tend to be expressed more in young adulthood than in other age groups (Deci & Ryan, 2000), which leads us to predict the importance of extrinsic goals (i.e. status goals) to be negatively linked to age.

In contrast to the proving grounds of young adulthood, in middle age (roughly ages 41–60 years; Staudinger & Bluck, 2001), adults pursue goals that secure, consolidate, and stabilize what has been established and that orient them toward the future of subsequent generations (Erikson, 1968; McAdams, 2015). For instance, middle–aged adults tend to value goals related to family and raising children, work, prosocial engagement, and passing traditions on to the next generation (Freund & Riediger, 2006; Havighurst, 1972; Neugarten & Datan, 1996). Accordingly, we expect middle–aged adults to give importance to prosocial–engagement, social–relation, and work goals.

The developmental tasks of older adults (roughly age 61 years and above; Staudinger & Bluck, 2001) are centred around the maintenance of a functional level in domains such as health or leisure (Ebner et al., 2006; Heckhausen et al., 1989), orienting older adults toward avoiding losses rather than acquiring potential gains (Freund, 2008; Ogilvie, Rose, & Heppen, 2001). In addition, given that the future tends to be perceived as limited in late adulthood, older adults usually follow goals that are present oriented, such as deepening close relationships (Carstensen, Isaacowitz, & Charles, 1999). Finally, older compared with younger adults tend to be more altruistic, for example, in more strongly valuing contributions to the public good and being more likely to donate money to a good cause (Freund & Blanchard–Fields, 2014). We therefore expect older adults to rate goals of health, prosocial engagement, and social relations as important.

Taken together, life stages with their normative expectations and structural opportunities yield developmental tasks that lay the ground for allocating differential importance to certain life goals (Hutteman et al., 2014). Building on this rationale, we predict that age shapes the relevance assigned to life goals: Importance of personal–growth, status, and work goals should be negatively associated with age, whereas importance of health and prosocial–engagement goals should be positively associated with age. As outlined, people of each life stage are inclined to value social–relation goals. Hence, we do not expect age differences in the importance ascribed to social–relation goals, although the motivation to value these goals might differ across age: The growth focus among young adults might shift to a focus on established and secured aspects of life in middle adulthood and to momentary rewards in older age. Despite these different motivations, social relationships are considered essential for individuals to thrive across the lifespan (e.g. Deci & Ryan, 2008) and should therefore reflect an important goal in each age group.

As well as goal importance, we investigated goal attainability, defined as the subjective perception of ‘opportunity, control, and support’ of goal pursuit (Brunstein, 1993, p. 1062). Given that goal importance and goal attainability dynamically interact with each other (Brandtstädter & Rothermund, 2002), we expect goal attainability to vary with age in a similar manner to goal importance. We predict that age effects on goal attainability should show the same age–related pattern as age effects on goal importance: Attainability of personal–growth, status, and work goals should be negatively associated with age; attainability of health and prosocial–engagement goals should be positively associated with age; and attainability of social–relation goals should not be affected by age. One can expect a differential effect in the health domain, because—although health goals become more important with increasing age—older adults often have to deal with physical health issues, health impairment, and cognitive decline, which limit the resources they can devote to attaining their goals (Reynolds & Finkel, 2016). Yet we expect age to be positively related to health–goal attainability for the following reason: If people experience a discrepancy between their goals and the likelihood of achieving these goals, they start to cope with this discrepancy (Brandtstädter & Rothermund, 2002; refer to the next section). Assuming that health goals become more important with increasing age at the same time that resources are shrinking, older adults might, for instance, rescale their aspirations in the health domain, develop more feasible health goals, and invest in these newly developed goals (Brandtstädter & Rothermund, 2002). These adjustment strategies likely render it possible not only to value health goals as important with increasing age but also to perceive them as attainable.

Goal dynamics: Association between goal importance and goal attainability over time

As indicated above, goals vary not only in their importance but also with regard to their perceived attainability (Atkinson, 1964; Tomasik, Knecht, & Freund, 2017). Self–regulation theories of development posit that people strive for control over their lives by balancing the importance and attainability of their goals (Baltes, 1987, 1997; Brandtstädter & Greve, 1994; Heckhausen, Schulz, & Wrosch, 1998). More specifically, the dual–process framework (Brandtstädter & Rothermund, 2002) proposes two modes of coping with the dynamics between the pursuit and the adjustment of goals, namely, the assimilation mode and the accommodation mode. Both modes illustrate two types of adaptive processes, which are complementary in that both intend to resolve a goal discrepancy, but they function in opposite ways (Brandtstädter, 1989; Brandtstädter & Renner, 1990; Brandtstädter & Rothermund, 2002; Rothermund & Brandtstädter, 2003).

The assimilation mode implies a purposeful and intentional change of individuals’ own behaviour or their life circumstances. That is, when important goals appear less attainable, people tend to invest more heavily in the goal pursuit and to intentionally modify the situation (assimilation mode or primary control strategies; Brandtstädter & Rothermund, 2002; Heckhausen & Schulz, 1995; Heckhausen et al., 1998; Rothermund & Brandtstädter, 2003; Wrosch, Heckhausen, & Lachman, 2000). This is typically the case in young adulthood when enough resources are available to acquire new skills, to improve existing functions, and to seek environments that offer access to new resources (Brandtstädter & Rothermund, 2002; Ebner et al., 2006). For instance, a 19–year–old woman might have the goal of becoming a successful professional swimmer, so she practices on a nearly daily basis to achieve this goal. During competitions, however, she rarely wins a medal. She realizes that her important goal appears to be becoming less attainable. As a consequence, she begins to practice with a new trainer to learn a modified swimming technique, and she joins a more competitive swimming group. She invests more heavily in her goal in that she has purposefully and intentionally changed her behaviour and life circumstances.

As evident from the above, these assimilative efforts require available resources, such as the perception of having enough time, of receiving social and/or financial support, and of being in good physical and/or mental shape (Brandtstädter & Rothermund, 2002). As people age, however, these internal and external resources may shrink (Brandtstädter & Rothermund, 2002; Freund, 2008), and the ratio of expected gains to expected losses of resources becomes less favourable (Ogilvie et al., 2001). It is, thus, the adjustment of goals to these constraints and the lowering of aspirations that contribute to the maintenance of high goal attainability and to control in older adulthood (i.e. accommodation mode or secondary control strategies; Brandtstädter & Rothermund, 2002; Heckhausen & Schulz, 1995; Heckhausen et al., 1998; Rothermund & Brandtstädter, 2003; Wrosch et al., 2000). This mode typically implies eliminating blocked goals, rescaling aspirations, or funnelling energies and investment into new, more feasible goals. For instance, a 72–year–old man might have the goal of being actively engaged in the political work of his community. For the last 20 years, he has participated in weekly meetings of his local government, given public speeches, and met various stakeholders and politicians to discuss current topics. Within the last 2 years, however, his ability to effortlessly give long speeches and travel long distances has decreased. While remaining actively engaged in community work, he has rescaled this goal. Now, he provides his expertise and knowledge from the back office through being a wise mentor for the next generation's politicians and a close advisor in speech writing.

As outlined, the balance between the assimilative and accommodative modes depends on the conditions and resources that individuals think are available in a specific life condition. We argue that age might be a moderating factor in this balance. We expect the associations between goal importance and goal attainability to be positive but to vary in how pronounced they are across adulthood. We assume that young adults more intensely invest in important goals, which will increase the perceived likelihood of attaining these goals (i.e. assimilation mode). Consequently, we predict that it is in younger age that goal importance predicts later goal attainability, and we expect this positive association to be negatively associated with age. In contrast, we hypothesize that older adults lower their aspirations and mentally decrease the importance of those goals that are no longer perceived as attainable (i.e. accommodation mode). Consequently, we predict that it is in older age that goal attainability predicts later goal importance, and we expect this positive association to be positively associated with age. Given that any life domain that is open to be modified can be addressed through these adaptive processes (Brandtstädter & Rothermund, 2002), we expect our hypotheses to apply to each of the investigated life–goal domains.

Goal outcomes: Goal importance and goal attainability as predictors of subjective well–being

It is well established that the pursuit of personally meaningful goals is advantageous for various indicators of subjective well–being (e.g. Brunstein, Schultheiss, & Maier, 1999; Deci & Ryan, 1985, 2008; Emmons, 1996; Emmons & King, 1988; C. Harris, Daniels, & Briner, 2003; Schmuck, Kasser, & Ryan, 2000). However, the study of goals for well–being has not been without controversy, because their role is ambiguous (Brandtstädter & Rothermund, 2002): While they motivate behaviour, give structure, and provide meaning, goals might also be a source of dissatisfaction when they are perceived as unattainable (especially when the goal remains important and no accommodative strategies have been applied). Thus, the link between goals and subjective well–being in light of both goal importance and goal attainability needs to be carefully investigated.

In line with previous theory (Diener, 1984; Diener et al., 1999; Lucas, Diener, & Suh, 1996), we understand subjective well–being as comprising a cognitive–evaluative component (global life satisfaction and domain–specific satisfaction) and an affective component (positive and negative affect). One way to investigate the consequences of life goals for subjective well–being is, for instance, to differentiate between intrinsically and extrinsically oriented goals (Deci & Ryan, 2000). On the basis of past theory and research (e.g. Kasser & Ryan, 1996), we define intrinsic goals as goals directed at the fulfilment of innate psychological needs such as relatedness, autonomy, and competence, including goals for meaningful relationships, community contributions, personal growth, and health. We conceptualize extrinsic goals as those directed at the desire for fame, image, and wealth, mapping on status goals. Work goals, finally, most likely reflect a combination of intrinsic and extrinsic goals.

Previous research suggests that intrinsic life goals are particularly conducive to well–being (e.g. Brunstein, 1993; Deci & Ryan, 1985). Extrinsic goals, in contrast, tend to work against people's well–being given that these goals are focused on obtaining external rewards and approbation from others, thereby giving the activity an instrumental character to achieve an intended consequence (e.g. Deci & Ryan, 1985; Kasser & Ryan, 1993, 1996; Schmuck et al., 2000). For the present research, we expect a positive association between goal importance and subjective well–being for the intrinsic goal domains and a negative association for the extrinsic goal domains.

With respect to goal attainability as a predictor of subjective well–being, we base our hypotheses on findings from research on locus of control (Rotter, 1966). While an internal locus of control implies that a person holds the belief that his or her own ability, effort, or actions determine what happens, an external locus of control reflects that fate, luck, or outside forces are responsible for what happens (Rotter, 1966). People tend to be more satisfied if they perceive a goal as attainable and feel a sense of control (e.g. Judge & Bono, 2001; Rodin, 1986; Rodin & Langer, 1977). Applying this to the present research, we expect that people who perceive their goals as attainable will indicate higher levels of subjective well–being. Here, we again differentiate between intrinsic and extrinsic goals. A previous finding demonstrated that the importance and attainability of intrinsic goals are positively linked to positive affect, while the importance and attainability of extrinsic goals are negatively linked to positive affect (Kasser & Ryan, 1996). No effects were found for the link between goal importance/attainability of intrinsic/extrinsic goals and negative affect in this study (Kasser & Ryan, 1996). Although these findings provide insights into the relevance of intrinsic and extrinsic goals for the affective component of subjective well–being (as well as, among others, vitality, depression, and physical symptoms), the study was limited in that other, more cognitive aspects of subjective well–being were not included (Kasser & Ryan, 1996). In addition, the study covered a short time span (a maximum of 7 days in Study 2), which highlights the need for a more nuanced investigation of the longitudinal prediction of the effects of goals on later subjective well–being. This leads us to expand the long–term prediction of an effect of goal attainability to diverse indicators of subjective well–being.

Finally, in line with the life–course perspective of the present study, we test whether age moderates the association between goal importance and subjective well–being and/or between goal attainability and subjective well–being. So far, there is a limited amount of research on possible age–related differences in the link between goal importance/goal attainability and subjective well–being. In their study, Kasser and Ryan (1996) found no support for a moderating effect between goal importance/goal attainability and the affective component of well–being. However, these findings bear two limitations. First, although the study sample included participants with a wide age range (18–79 years), it was a small sample of 100 adults, and age effects might have been underestimated. Second, the cognitive–evaluative component of subjective well–being was not included in the study. Hence, we seek to shed further light on possible moderating effects of age on the association between goal importance/goal attainability and different indicators of subjective well–being.

We have now established links between importance and attainability of life goals with life satisfaction, positive affect, and negative affect for the present investigation. Yet we want to go one step further and focus on an additional aspect of the cognitive–evaluative component of well–being: Domain–specific satisfaction. Although past findings have demonstrated a consistent association between certain life goals and well–being, they often tend to overlook that the link between goals and satisfaction can differ between life domains (Diener, 1994; Vansteenkiste et al., 2007). In other words, rating goals in the work domain as important and perceiving them as attainable might have positive ramifications for satisfaction in the work domain (e.g. Lent & Brown, 2006; Maier & Brunstein, 2001; Roberson, 1990) but not in the family domain. This is the case because work goals lead to allocation of resources to the work domain but not to the relationship and family domains (e.g. Wiese & Freund, 2005). Conversely, relationship goals might be positively linked to satisfaction in the relationship domain (e.g. Sanderson & Evans, 2001) but not to satisfaction in the occupational domain.

Given that goals require the allocation of resources to particular life domains at the expense of allocations to other domains, we expect to find thematic associations between the importance and attainability of goal domains and satisfaction within given life domains (e.g. work goals predicting satisfaction with work but not satisfaction with social relationships). In terms of life domains, we focus on (i) occupational performance, (ii) health, and (iii) interpersonal relations as salient and important domains of an adult's life. For (i) occupational performance, we consider that people of different ages are confronted with different aspects of occupational performance and investigate both satisfaction with education (likely to be present for young adults) and satisfaction with work [likely to be (or to have been) present for middle–aged and older adults]. For (ii) health, we assess overall satisfaction with one's health. Finally, for (iii) interpersonal relations, previous research has shown that social contexts change over the lifespan and, more specifically, that social networks decrease after a person's second decade of life (Wrzus, Hänel, Wagner, & Neyer, 2013). However, findings of this study also suggest taking a differentiated perspective on a person's social network given that friendship networks decrease across age but family networks remain stable (Wrzus et al., 2013). We thus acknowledge the manifold character of social contexts and differentiate between three areas of interpersonal relationships, namely, satisfaction with family life, satisfaction with one's romantic relationship, and satisfaction with friendships.

In line with the developmental focus of this paper, we again test for moderating effects of age. We hypothesize that goal–outcome associations are stronger for goal domains that correspond to developmental tasks at a particular age. This might be the case because focusing on goals that correspond to developmental tasks enables people to fulfil these tasks, which is likely experienced as rewarding. Consequently, we expect that the link between work goals and satisfaction in the work domain decreases with age, that the link between health goals and satisfaction in the health domain increases with age, and that the link between social goals and satisfaction in the social domain remains stable across adulthood.

Effects of age on goal importance and goal attainability (goal content)

Hypothesis 1a: Importance of work, status, and personal–growth goals is negatively associated with age, and importance of health and prosocial–engagement goals is positively associated with age. We do not expect age effects on importance of social–relation goals.

Hypothesis 1b: Attainability of work, status, and personal–growth goals is negatively associated with age, and attainability of health and prosocial–engagement goals is positively associated with age. We do not expect age effects on attainability of social–relation goals.

The association between goal importance and goal attainability over 2 years (goal dynamics)

Hypothesis 2a: Goal importance and goal attainability are reciprocally linked to each other; that is, goal importance predicts later goal attainability, and goal attainability predicts later goal importance.

Hypothesis 2b: We expect these associations to be moderated by age in that (i) the association between goal importance and later goal attainability is weaker with higher age, while (ii) the association between goal attainability and later goal importance is stronger with higher age.

Goal importance and goal attainability as predictors of subjective well–being (goal outcomes)

Hypothesis 3a: Importance of intrinsic life goals is positively linked to later subjective well–being (i.e. life satisfaction, and positive and negative affect), and importance of extrinsic life goals is negatively linked to later subjective well–being.

Hypothesis 3b: Attainability of intrinsic life goals is positively related to later subjective well–being (i.e. life satisfaction, and positive and negative affect), and attainability of extrinsic life goals is negatively related to later subjective well–being.

Hypothesis 3c: For domain–specific satisfaction, we expect a thematic link between goal domains and their respective satisfaction domains: Importance and attainability of work goals predict satisfaction with the occupational domain; importance and attainability of health goals predict health satisfaction; and importance and attainability of social–relation goals predict satisfaction in the social domain. These associations are stronger at the age at which the goal is valued as more important and attainable.

Sample and general procedure

Data for the present study were obtained from the longitudinal Co–Development in Personality (CoDiP) ¹ study that was conducted in the German–speaking parts of Switzerland. Ethical approval for the CoDiP study was received from the regional ethics committee of Basel (approval number: 175/09) at the University of Basel, Switzerland. Necessary supplemental materials (i.e. overview of study variables and data analysis script) are stored at a public and open–access repository (accessible through the following link: https://osf.io/s2w3n/).

Individuals from different age groups were recruited either through university and vocational schools (young adults) or through lectures given as part of a lifelong learning course aimed at seniors (older adults). Individuals were asked to invite their parents and grandparents (in the case of young adults) and their children and grandchildren (in the case of older adults) to participate. Thus, the final sample of the study included family members of three age groups (young, middle–aged, and older adults) who participated at three measurement occasions (referred to as Time 1, Time 2, and Time 3) with intervals of 2 years. At all three measurement occasions, participants provided self–reports on goal importance and subjective well–being (i.e. life satisfaction, positive affect, negative affect, and domain–specific satisfaction). Goal attainability was assessed at Time 1 and Time 2.

Participants

At Time 1, the sample included 973 individuals above age 18 years from 341 families, of whom 637 participated at Time 2 and 573 participated at Time 3. The age of participants at Time 1 ranged from 18 to 92 years (M = 43.00 years; SD = 22.08) with 57.6% identifying as female and 42.4% as male. To test for attrition effects, we compared participants who participated after Time 1 (continuers) with those who participated only at Time 1 (noncontinuers). Continuers, compared with noncontinuers, were more often female (62.7% vs. 53.8%), were older (M = 44.52 vs. M = 39.45 years), t(972) = −3.33, p = .001, and indicated lower scores in importance of work goals (M = 3.30 vs. M = 3.40), t(955) = −2.66, p = .01, and lower scores in importance of status goals (M = 2.06 vs. M = 2.22), t(969) = −4.53, p < .001. Continuers, compared with noncontinuers, also reported lower scores in attainability of status goals (M = 2.33 vs. M = 2.47), t(965) = −3.31, p = .001, positive affect (M = 3.60 vs. M = 3.68), t(972) = −2.30, p = .02, and negative affect (M = 1.72 vs. M = 1.83), t(971) = −2.67, p = .01. No other differences were statistically significant (all ps ≥ .05).

Within the sample of continuers, we further compared participants who continued to Time 3 (long–term continuers) with those who participated at Time 1 and Time 2 (short–term continuers). Long–term continuers, compared with short–term continuers, were more often female (64.6% vs. 45.5%) and were significantly more satisfied with their life at Time 1 (M = 3.99 vs. M = 3.80), t(688) = 2.84, p = .01, and at Time 2 (M = 4.01 vs. M = 3.84), t(628) = 2.45, p = .02. Long–term continuers were also more satisfied with their work domain at Time 1 (M = 7.52 vs. M = 6.87), t(357) = 4.47, p < .001, with their work domain at Time 2 (M = 7.52 vs. M = 6.87), t(415) = 2.43, p = .02, and with their health domain at Time 2 (M = 7.52 vs. M = 6.87), t(628) = 2.82, p = .01. No other differences were statistically significant (all ps ≥ .05).

Measures

Data analysis approach

Given that our sample included participants from the same family (i.e. young–adult children, parents, and grandparents), we first tested for interrelations between family members on the key variables. Intraclass correlation coefficients ranging between 0.14 (life satisfaction) and 0.55 (importance of social–relation goals) supported the assumption of nonindependence of data. Consequently, to conduct statistical analyses, we used a multilevel modelling approach that takes nonindependence of data into account. ² We applied a two–level approach, in which Level 2 represents the family and Level 1 the individual. Controlling for variation between families on Level 2, Level 1 represents individuals’ variations on the relevant key variables. Given that variations between families were not the focus of the present paper, we present results on Level 1. Multilevel analyses were conducted by using the lme4 package in R (Bates, Maechler, Bolker, & Walker, 2015), and figures were created by making use of the effects package (Fox et al., 2018) and the ggplot package in R (Wickham, 2016). Johnson–Neyman analyses to explore the regions of significance for the age moderations were conducted by using the jtools package in R (Long, 2018). For all analyses, age was used as continuous variable. Life–goal predictors, age, and outcomes were grand mean centred. Missing values were handled with the maximum likelihood estimation approach.

For the hypotheses on goal content (Hypotheses 1a and 1b), we applied multilevel regression analyses in two separate models. In the first model, the dependent variable was goal importance; in the second model, the dependent variable was goal attainability. In both models, the predictor was age (continuous linear effects and squared effects). Results represent effects at Time 1. ³

For the hypotheses on goal dynamics (Hypotheses 2a and 2b), we applied multilevel regression analyses in two separate models for goal importance and goal attainability. In the first model, we tested the stability of the variable of interest (i.e. earlier goal importance on later goal importance as well as earlier goal attainability on later goal attainability). In the second model, we regressed goal importance at Time 1 on goal attainability at Time 2, and vice versa, controlling for the 2–year stability of the later variable. In both models, we included the interaction effect of goal importance and age as well as the interaction effect of goal attainability and age in the model.

For the hypotheses on goal outcomes (Hypotheses 3a, 3b, and 3c), we conducted multilevel regression analyses for each outcome variable in separate models. The dependent variables were life satisfaction, positive affect, negative affect, and domain–specific satisfaction (i.e. work, education, health, family, relationship, and friendship satisfaction). We calculated models for outcomes at Time 2 and Time 3, controlling for the stability of the outcome. In each model, the predictors were goal importance at Time 1, goal attainability at Time 1, the interaction between goal importance and age, and the interaction between goal attainability and age. ⁴

We are aware of the problems associated with multiple testing and note that the present study includes a considerable number of analyses. However, rather than lowering the p level and narrowing the confidence interval (CI), we have decided to report all analyses at the conventional p level of 5% and at a CI of 95% and to interpret those results that show a consistent and robust pattern (Perneger, 1998).

Descriptive statistics and preliminary analyses

Table 1 shows the means and standard deviations of goal importance, goal attainability, and well–being at Time 1, as well as their intercorrelations. As evident from this table, with one exception, all goal domains were intercorrelated with each other in both their importance and their attainability. The exception was status goals, which showed fewer correlations for both importance and attainability. The table also reveals that most goal domains were positively related to indicators of subjective well–being with the exception of status goals and work goals, for which fewer correlations were observed. Means and standard deviations of goal importance and goal attainability at Time 2 as well as means and standard deviations of well–being at Time 2 and Time 3 are provided in Table S7.

OPEN IN VIEWER

Table 1

Means and standard deviations of goal importance, goal attainability, and subjective well–being at Time 1 and their intercorrelations

Variable	M	SD	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	19	20
Goal importance
1 Health	3.63	0.40
2 Personal growth	3.51	0.39	.28
3 Prosocial engagement	3.11	0.49	.25	.34
4 Social relations	3.71	0.35	.36	.33	.33
5 Status	2.11	0.51	.17	.17	−.03	.17
6 Work	3.33	0.57	.17	.27	.06	.27	.40
Goal attainability
7 Health	3.00	0.50	.38	.12	.10	.13	.16	.17
8 Personal growth	3.18	0.42	.14	.44	.18	.18	.05	.20	.38
9 Prosocial engagement	2.89	0.46	.14	.20	.60	.26	−.04	.07	.34	.44
10 Social relations	3.33	0.44	.24	.19	.27	.55	.08	.16	.36	.44	.43
11 Status	2.38	0.63	.07	.03	.00	.11	.29	.19	.21	.26	.30	.23
12 Work	3.03	0.58	.10	.17	.10	.12	.18	.61	.29	.43	.31	.35	.42
Subjective well–being
13 Life satisfaction	3.94	0.69	.15	.12	.14	.20	−.04	.09	.29	.39	.25	.41	.14	.29
14 Positive affect	3.62	0.55	.19	.27	.17	.16	.13	.25	.32	.46	.31	.33	.24	.34	.41
15 Negative affect	1.75	0.57	−.05	.02	−.05	.01	.26	.09	−.13	−.25	−.16	−.18	−.03	−.13	−.43	−.17
16 Work satisfaction	7.29	2.11	.12	.09	.11	.11	.03	.22	.19	.21	.21	.20	.15	.37	.38	.33	−.23
17 Education satisfaction	7.37	2.05	.04	.08	.06	.09	−.05	.16	.09	.23	.15	.18	.11	.28	.33	.26	−.18	.53
18 Health satisfaction	7.50	2.15	.11	.08	−.02	.06	.05	.17	.42	.19	.10	.18	.08	.15	.28	.28	−.15	.18	.12
19 Family satisfaction	7.96	2.03	.14	.07	.13	.26	−.02	.03	.10	.19	.15	.41	.06	.08	.44	.24	−.22	.26	.18	.19
20 Friendship satisfaction	8.10	1.92	.08	.04	.14	.22	−.01	.03	.18	.20	.19	.36	.11	.12	.40	.23	−.22	.26	.17	.18	.35
21 Relationship satisfaction	8.23	2.02	.09	.04	.06	.31	−.04	.08	.09	.18	.11	.51	.08	.14	.32	.10	−.16	.09	.18	.18	.47	.31

Note: N = 973. Correlation coefficients in bold are significant (p < .05). Life–goal domains were as follows: health goals, personal–growth goals, prosocial–engagement goals (community and generativity goals), social–relation goals (family and relationship goals), status goals, and work goals. Goal importance and goal attainability were assessed on a 4–point Likert scale (from 1 to 4); life satisfaction, positive affect, and negative affect were assessed on a 5–point Likert scale (from 1 to 5); and domain–specific satisfaction was assessed on an 11–point Likert scale (from 0 to 10).

Content of goals: Age effects on goal importance and goal attainability

Linear and squared age effects on goal importance and goal attainability are presented in Table 2. Age exhibited significant effects on goal importance in all life–goal domains: Age had a negative linear effect on importance of personal–growth, social–relation, status, and work goals, whereas it had a positive linear effect on importance of health and prosocial–engagement goals. Further, age had a positive squared effect on status goals, while it exhibited a negative squared effect on work goals. No other significant squared effects were observed (all ps > .05). Figure 1 illustrates the linear and squared age effects on goal importance. Except for the negative age effect on social–relation goals, the present results support Hypothesis 1a.

OPEN IN VIEWER

Table 2

Linear and squared effects of age on goal importance and goal attainability at Time 1

Variable	Linear effects				Squared effects
	b	SE	95% CI	p	b	SE	95% CI	p
Goal importance
Health	0.02	0.006	[0.01, 0.03]	.003	0.003	0.003	[−0.003, 0.008]	.32
Personal growth	−0.02	0.005	[−0.03, −0.01]	<.001	−0.002	0.003	[−0.007, 0.004]	.59
Prosocial engagement	0.04	0.007	[0.02, 0.05]	<.001	−0.001	0.003	[−0.009, 0.006]	.69
Social relations	−0.03	0.005	[−0.03, −0.02]	<.001	−0.002	0.003	[−0.007, 0.002]	.47
Status	−0.07	0.006	[−0.09, −0.06]	<.001	0.01	0.003	[0.003, 0.018]	.004
Work	−0.11	0.008	[−0.12, −0.09]	<.001	−0.001	0.004	[−0.02, −0.002]	.02
Goal attainability
Health	−0.01	0.007	[−0.02, 0.01]	.29	0.00001	0.003	[−0.007, 0.008]	.97
Personal growth	−0.02	0.001	[−0.03, 0.01]	.003	−0.008	0.003	[−0.01, −0.002]	.01
Prosocial engagement	0.02	0.006	[0.01, 0.04]	<.001	−0.01	0.003	[−0.02, −0.01]	<.001
Social relations	−0.01	0.006	[−0.02, 0.01]	.08	−0.004	0.003	[−0.01, 0.002]	.19
Status	−0.05	0.008	[−0.06, −0.02]	<.001	−0.02	0.004	[−0.03, −0.01]	<.001
Work	−0.06	0.008	[−0.07, −0.04]	<.001	−0.02	0.004	[−0.02, −0.01]	<.001

Note: N = 973. CI, confidence interval. Significant results (p < .05) are shown in bold. Age is scaled in decades. For goal importance, testing Model 1 (only linear effects) against Model 2 (linear and squared effects), the combined model fit the data significantly better in the domains of status goals, χ²(1, 973) = 8.13, p = .004, and work goals, χ²(1, 973) = 5.61, p = .02. For goal attainability, testing Model 1 (only linear effects) against Model 2 (linear and squared effects), results revealed that Model 2 fit the data significantly better than did Model 1 in the domains of personal growth, χ²(1, 973) = 6.38, p = .01, prosocial engagement, χ²(1, 973) = 15.76, p < .001, status, χ²(1, 973) = 9.93, p = .002, and work, χ²(1, 973) = 11.21, p < .001.

OPEN IN VIEWER

Figure 1

Effects of age on goal importance in six life–goal domains. Areas in light blue display the 95% confidence intervals. For Models A, B, C, and D, linear effects fit the data sufficiently, while for Models E and F, the combined model of linear and squared effects fit the data better. [Colour figure can be viewed at wileyonlinelibrary.com]

For goal attainability, age exhibited fewer effects: Age had a negative linear effect on the attainability of personal–growth, status, and work goals and a positive effect on the attainability of prosocial–engagement goals. Further, age exhibited a negative squared effect on personal–growth, prosocial–engagement, status, and work goals. Neither linear nor squared age effects were observed for attainability of social–relation and health goals. Figure 2 illustrates the linear and squared effects that age exhibited on goal attainability. Except for the nonsignificant effect on health goals, these findings support Hypothesis 1b.

OPEN IN VIEWER

Figure 2

Effects of age on goal attainability in six life–goal domains. Areas in light blue display the 95% confidence intervals. For Models A and D, linear effects fit the data sufficiently, while for Models B, C, E, and F, the combined model of linear and squared effects fit the data better. [Colour figure can be viewed at wileyonlinelibrary.com]

Goal dynamics: Association between goal importance and goal attainability over time

Next, we tested the reciprocal association between goal importance and goal attainability over a time span of 2 years, which is shown in Table 3. Goal importance and goal attainability were fairly stable in all life–goal domains across the two measurement occasions. In addition, goal importance and goal attainability predicted each other significantly over time: Earlier goal attainability positively predicted later goal importance in the life–goal domains of health, personal growth, and social relations. For the inverse direction—earlier goal importance on later goal attainability—all associations were significant in a positive direction.

OPEN IN VIEWER

Table 3

Multilevel regression analyses predicting goal importance and goal attainability at Time 2 (T2) from goal importance and goal attainability at Time 1 (T1)

Life–goal domain	Stability effects								Attainability ↔ Importance
	Importance T1 → Importance T2				Attainability T1 → Attainability T2				Attainability T1 → Importance T2				Importance T1 → Attainability T2
	b	SE	95% CI	p	b	SE	95% CI	p	b	SE	95% CI	p	b	SE	95% CI	p
Health	0.58	0.03	[0.52, 0.64]	<.001	0.66	0.03	[0.59, 0.72]	<.001	0.10	0.03	[0.04, 0.15]	<.001	0.10	0.04	[0.02, 0.18]	.02
Personal growth	0.61	0.03	[0.54, 0.68]	<.001	0.52	0.04	[0.45, 0.59]	<.001	0.11	0.04	[0.04, 0.18]	.002	0.12	0.04	[0.04, 0.21]	.005
Prosocial engagement	0.70	0.03	[0.65, 0.77]	<.001	0.55	0.03	[0.48, 0.62]	<.001	−0.01	0.04	[−0.08, 0.07]	.92	0.13	0.04	[0.05, 0.21]	.001
Social relations	0.75	0.04	[0.67, 0.82]	<.001	0.70	0.03	[0.64, 0.76]	<.001	0.08	0.03	[0.02, 0.15]	.01	0.10	0.05	[0.02, 0.21]	.02
Status	0.73	0.03	[0.68, 0.79]	<.001	0.48	0.04	[0.40, 0.55]	<.001	0.02	0.02	[−0.03, 0.05]	.41	0.18	0.05	[0.08, 0.28]	<.001
Work	0.55	0.05	[0.46, 0.63]	<.001	0.57	0.04	[0.49, 0.66]	<.001	0.08	0.05	[−0.01, 0.18]	.10	0.19	0.05	[0.09, 0.29]	<.001

Note: N = 637. CI, confidence interval. Significant results (p < .05) are presented in bold. In each model, predictors were goal importance and/or goal attainability, age, and interaction effects with age. In predicting later goal importance or goal attainability, we controlled for the stability of the respective outcome measure. For stability effects of goal importance, explained variance associated with fixed effects was R²_health = .34, R²_{personal growth} = .36, R²_{prosocial engagement} = .48, R²_{social relations} = .44, R²_status = .59, and R²_work = .33. For stability effects of goal attainability, explained variance associated with fixed effects was R²_health = .28, R²_{personal growth} = .36, R²_{prosocial engagement} = .32, R²_{social relations} = .48, R²_status = .23, and R²_work = .29. For the predictive effect of earlier goal attainability on later goal importance, explained variance associated with fixed effects was R²_health = .36, R²_{personal growth} = .37, R²_{prosocial engagement} = .48, R²_{social relations} = .45, R²_status = .59, and R²_work = .33. For the predictive effect of earlier goal importance on later goal attainability, explained variance associated with fixed effects was R²_health = .43, R²_{personal growth} = .29, R²_{prosocial engagement} = .33, R²_{social relations} = .48, R²_status = .25, and R²_work = .31.

Goal outcomes: Goal importance and goal attainability as predictors of subjective well–being

Addressing the predictive validity of goal importance and goal attainability on subjective well–being, we tested predictive effects on later subjective well–being (i.e. life satisfaction, positive affect, negative affect, and domain–specific satisfaction) after intervals of 2 and 4 years. Here, we provide tables for the prediction of well–being at Time 2, while tables for the prediction of well–being at Time 3 are shown in Tables S8, S9, and S10.

Life satisfaction

The first section of Table 4 shows the predictive effects of goal importance and goal attainability on life satisfaction at Time 2, controlling for life satisfaction at Time 1. While no main effect was observed of earlier goal importance on later life satisfaction, earlier goal attainability showed significant predictive validity: Attainability of health, personal–growth, prosocial–engagement, and social–relation goals was positively associated with life satisfaction at Time 2. For life satisfaction at Time 3, earlier attainability of personal–growth goals (b = 0.24, p < .001, 95% CI [0.12, 0.36]) and social–relation goals (b = 0.20, p = .001, 95% CI [0.08, 0.31]) remained a significant positive predictor, while attainability of health and prosocial–engagement goals was no longer significantly predictive (Table S8).

OPEN IN VIEWER

Table 4

Multilevel regression analyses predicting subjective well–being at Time 2 from goal importance and goal attainability at Time 1

Variable	Life satisfaction				Positive affect				Negative affect
	b	SE	95% CI	p	b	SE	95% CI	p	b	SE	95% CI	p
Goal importance
Health	−0.10	0.05	[−0.21, 0.01]	.06	0.01	0.05	[−0.09, 0.10]	.78	0.03	0.04	[−0.07, 0.12]	.57
Personal growth	−0.09	0.06	[−0.20, 0.03]	.12	−0.13	0.05	[−0.23, −0.03]	.01	0.15	0.05	[0.05, 0.25]	.001
Prosocial engagement	−0.09	0.05	[−0.20, 0.02]	.08	−0.07	0.05	[−0.15, 0.03]	.17	0.09	0.04	[0.01, 0.18]	.04
Social relations	−0.09	0.07	[−0.23, 0.05]	.21	−0.14	0.06	[−0.27, −0.01]	.03	0.10	0.06	[−0.01, 0.21]	.11
Status	0.02	0.04	[−0.06, 0.10]	.64	−0.03	0.04	[−0.11, 0.05]	.42	0.04	0.04	[−0.03, 0.11]	.28
Work	0.06	0.05	[−0.03, 0.15]	.24	0.01	0.04	[−0.06, 0.19]	.79	0.01	0.04	[−0.07, 0.10]	.78
Goal attainability
Health	0.15	0.05	[0.07, 0.24]	.001	0.11	0.04	[0.03, 0.19]	.01	−0.11	0.04	[−0.19, −0.04]	.002
Personal growth	0.18	0.06	[0.05, 0.28]	.002	0.18	0.05	[0.07, 0.27]	<.001	−0.19	0.04	[−0.28, −0.10]	<.001
Prosocial engagement	0.12	0.06	[0.01, 0.23]	.04	0.14	0.05	[0.05, 0.24]	.01	−0.12	0.05	[−0.22, −0.03]	.01
Social relations	0.20	0.06	[0.07, 0.33]	<.001	0.20	0.05	[0.09, 0.31]	<.001	−0.16	0.05	[−0.25, −0.07]	<.001
Work	0.06	0.05	[−0.03, 0.15]	.20	0.16	0.04	[0.07, 0.24]	<.001	0.001	0.04	[−0.08, 0.08]	.97
Status	0.03	0.03	[−0.05, 0.09]	.41	0.08	0.03	[0.02, 0.14]	.01	−0.02	0.03	[−0.07, 0.04]	.53

Note: N = 637. CI, confidence interval. Significant results (p < .05) are presented in bold. In each model, predictors were goal importance, goal attainability, age, and interaction effects with age. Results are controlled for the stability of the outcome measure. For life satisfaction, explained variance associated with fixed effects was R² = .51 for each model. For positive affect, explained variance associated with fixed effects was R² = .48 for each model. For negative affect, explained variance associated with fixed effects was R²_health = .44, R²_{personal growth} = .45, R²_{prosocial engagement} = .44, R²_{social relations} = .44, R²_status = .43, and R²_work = .43.

Age moderations

We observed one significant age moderation for the association between earlier goal importance/goal attainability and later life satisfaction: Age moderated the link between earlier work–goal importance and life satisfaction at Time 2 (b = −0.004, p = .04, 95% CI [−0.008, −0.001]). We found that among participants younger than 21.85 years, higher importance of work goals predicted higher life satisfaction and that this effect was more pronounced the younger participants were.

Satisfaction with occupational domain

Tables 5 and 6 display the predictive effects of goal importance and goal attainability on domain–specific satisfaction at Time 2. For the occupational domain at Time 2 (Table 5), work satisfaction was not predicted by goal importance in any goal domain, but work satisfaction was positively predicted by earlier attainability of personal–growth and work goals. Educational satisfaction was also not predicted by goal importance, but it was positively predicted by attainability of status goals. For satisfaction with the occupational domain at Time 3, goal importance was again not predictive, but attainability of health goals (b = 0.82, p < .001, 95% CI [0.43, 1.24]) and attainability of personal–growth goals (b = 0.60, p = .02, 95% CI [0.04, 1.14]) were significant positive predictors of later work satisfaction (Table S9). Attainability of these two goal domains was also positively predictive of educational satisfaction at Time 3 (health goals: b = 0.94, p = .002, 95% CI [0.34, 1.53]; and personal–growth goals: b = 0.80, p = .04, 95% CI [0.05, 1.64]).

OPEN IN VIEWER

Table 5

Multilevel regression analyses predicting domain–specific satisfaction (work satisfaction, educational satisfaction, and health satisfaction) at Time 2 from goal importance and goal attainability at Time 1

Variable	Work satisfaction				Educational satisfaction				Health satisfaction
	b	SE	95% CI	p	b	SE	95% CI	p	b	SE	95% CI	p
Goal importance
Health	−0.10	0.34	[−0.80, 0.59]	.77	−0.01	0.28	[−0.54, 0.49]	.98	−0.11	0.21	[−0.50, 0.32]	.59
Personal growth	−0.68	0.38	[−1.43, 0.08]	.08	−0.32	0.31	[−0.94, 0.31]	.29	−0.46	0.23	[−0.93, −0.03]	.04
Prosocial engagement	−0.20	0.33	[−0.92, 0.51]	.54	−0.32	0.27	[−0.87, 0.24]	.25	−0.07	0.20	[−0.48, 0.30]	.72
Social relations	0.02	0.46	[−0.79, 1.03]	.96	0.20	0.38	[−0.55, 0.97]	.59	−0.24	0.27	[−0.75, 0.35]	.38
Status	0.17	0.28	[−0.46, 0.75]	.55	−0.34	0.24	[−0.78, 0.11]	.15	−0.02	0.17	[−0.40, 0.31]	.89
Work	−0.05	0.33	[−0.70, 0.50]	.87	−0.01	0.28	[−0.56, 0.55]	.97	0.05	0.19	[−0.31, 0.38]	.79
Goal attainability
Health	0.46	0.29	[−0.17, 1.02]	.13	0.32	0.24	[−0.12, 0.79]	.19	0.49	0.18	[0.11, 0.81]	.01
Personal growth	0.87	0.35	[0.18, 1.59]	.01	0.26	0.28	[−0.26, 0.83]	.36	0.40	0.20	[0.01, 0.84]	.04
Prosocial engagement	0.46	0.36	[−0.27, 1.19]	.21	0.12	0.29	[−0.42, 0.73]	.69	0.35	0.35	[−0.11, 0.75]	.11
Social relations	0.32	0.38	[−0.44, 1.04]	.40	0.08	0.30	[−0.51, 0.72]	.78	0.59	0.21	[0.13, 1.05]	.01
Status	−0.01	0.20	[−0.43, 0.35]	.96	0.35	0.16	[0.01, 0.64]	.03	0.18	0.13	[−0.06, 0.42]	.16
Work	0.91	0.31	[0.24, 1.55]	.004	0.46	0.28	[−0.11, 0.96]	.08	0.21	0.18	[−0.19, 0.58]	.26

Note: N_work = 294; N_education = 331; N_health = 624. CI, confidence interval. Significant results (p < .05) are presented in bold. In each model, predictors were goal importance, goal attainability, age, and interaction effects with age. Results are controlled for the stability of the outcome measure. For work satisfaction, explained variance associated with fixed effects was R²_health = .21, R²_{personal growth} = .22, R²_{prosocial engagement} = .20, R²_{social relations} = .21, R²_status = .20, and R²_work = .24. For educational satisfaction, explained variance associated with fixed effects was R²_health = .15, R²_{personal growth} = .15, R²_{prosocial engagement} = .15, R²_{social relations} = .16, R²_status = .17, and R²_work = .16. For health satisfaction, explained variance associated with fixed effects was R²_health = .36, R²_{personal growth} = .35, R²_{prosocial engagement} = .36, R²_{social relations} = .36, R²_status = .35, and R²_work = .36.

OPEN IN VIEWER

Table 6

Multilevel regression analyses predicting domain–specific satisfaction (family satisfaction, romantic relationship satisfaction, and friendship satisfaction) at Time 2 from goal importance and goal attainability at Time 1

Variable	Family satisfaction				Romantic relationship satisfaction				Friendship satisfaction
	b	SE	95% CI	p	b	SE	95% CI	p	b	SE	95% CI	p
Goal importance
Health	−0.07	0.18	[−0.42, 0.28]	.69	0.13	0.23	[−0.28, 0.62]	.58	0.18	0.18	[−0.19, 0.57]	.32
Personal growth	−0.24	0.19	[−0.64, 0.14]	.20	−0.17	0.25	[−0.49, 0.28]	.49	−0.27	0.19	[−0.68, 0.11]	.16
Prosocial engagement	0.11	0.17	[−0.23, 0.44]	.50	0.13	0.23	[−0.32, 0.59]	.59	0.05	0.78	[−0.28, 0.39]	.17
Social relations	−0.06	0.23	[−0.40, 0.55]	.80	0.25	0.32	[−0.35, 0.85]	.44	0.07	0.24	[−0.38, 0.49]	.76
Status	−0.11	0.14	[−0.40, 0.18]	.45	−0.07	0.19	[−0.49, 0.32]	.73	−0.19	0.15	[−0.49, 0.10]	.20
Work	0.12	0.16	[−0.17, 0.45]	.46	−0.20	0.23	[−0.64, 0.25]	.39	−0.20	0.16	[−0.53, 0.15]	.22
Goal attainability
Health	0.16	0.14	[−0.11, 0.45]	.25	0.17	0.20	[−0.21, 0.57]	.41	0.06	0.15	[−0.25, 0.34]	.66
Personal growth	0.27	0.17	[−0.09, 0.61]	.12	0.91	0.23	[−0.32, 0.23]	<.001	0.48	0.18	[0.08, 0.83]	.01
Prosocial engagement	0.21	0.21	[−0.15, 0.61]	.25	0.25	0.24	[−0.27, 0.71]	.29	−0.03	0.19	[−0.39, 0.34]	.86
Social relations	0.73	0.73	[0.37, 1.09]	<.001	0.81	0.28	[0.24, 1.35]	.004	0.49	0.19	[0.16, 0.84]	.01
Status	−0.02	0.11	[−0.23, 0.18]	.82	−0.04	0.15	[−0.33, 0.23]	.77	0.03	0.11	[−0.22, 0.24]	.79
Work	−0.02	0.16	[−0.32, 0.27]	.91	0.41	0.21	[−0.01, 0.82]	.06	0.09	0.16	[−0.21, 0.40]	.58

Note: N_family = 615; N_{romantic relationship} = 418; N_friendship = 622. CI, confidence interval. Significant results (p < .05) are presented in bold. In each model, predictors were goal importance, goal attainability, age, and interaction effects with age. Results are controlled for the stability of the outcome measure. For family satisfaction, explained variance associated with fixed effects was R²_health = .35, R²_{personal growth} = .35, R²_{prosocial engagement} = .35, R²_{social relations} = .37, R²_status = .34, and R²_work = .34. For relationship satisfaction, explained variance associated with fixed effects was R²_health = .31, R²_{personal growth} = .35, R²_{prosocial engagement} = .28, R²_{social relations} = .31, R²_status = .28, and R²_work = .30. For friendship satisfaction, explained variance associated with fixed effects was R²_health = .29, R²_{personal growth} = .35, R²_{prosocial engagement} = .28, R²_{social relations} = .29, R²_status = .28, and R²_work = .28.

Age moderations

Testing moderating effects of age for the association between earlier goal importance/goal attainability and later occupational satisfaction (i.e. work satisfaction and educational satisfaction), we found no significant effect (all ps > .05).

Content of goals: Age differences in goal importance and goal attainability

Baltes (1987) differentiated between three factors that may have an influence on goal content: Nonnormative factors, normative history–graded factors, and normative age–graded factors. Focusing on the latter of these aspects, and in line with Hypothesis 1a, the present findings revealed that age predicted which goals people rate as important: With higher age, participants rated goals of personal growth, status, and work as less important, while they rated goals of health and prosocial engagement as more important. Participants also rated social–relation goals as less important with higher age, which was not in line with our hypothesis and will be discussed later. With regard to goal attainability and in line with Hypothesis 1b, participants perceived goals of personal growth, status, and work as less attainable later in life, while they valued goals of prosocial engagement as more attainable. No significant effects were found for the predicted effects on health–goal attainability, and no effects were found for the attainability of social–relation goals, indicating that people of all ages perceived their health goals and their social–relation goals as equally attainable.

Our findings correspond to the findings of Nurmi's (1992) study of three age groups that was conducted 25 years ago in Finland: Goals reflect what is possible and normative for people of different ages given their physical, cognitive, and social resources. Goals also reflect what is desirable for people of different ages, underscoring the role of age–related norms that make certain life tasks salient. Hence, the present results underline the integral part that age plays in the life goals that people value as important and perceive as attainable, situating the present findings in the context of developmental–task theory and supporting the concept of the postulated social–biological, age–based tasks (Hutteman et al., 2014). Our findings also highlight the concept of a social clock (Heckhausen, 1999; Neugarten, 1972), which means that norms and demands provide a temporal and contextual setting on which to orient one's goal pursuit in each life stage (Freund & Baltes, 2005). Overall, we maintain that the development over the adult lifespan is not a passive process but an active and dynamic interaction between the person and the norms, constraints, and tasks imposed by an age–graded environment (Baltes, Lindenberger, & Staudinger, 1998).

Goal dynamics: Association between goal importance and goal attainability over time

Both goal importance and goal attainability showed 2–year stabilities of moderate size that are comparable or slightly higher than previously reported in longitudinal studies (Lüdtke, Trautwein, & Husemann, 2009; Roberts, O'Donnell, & Robins, 2004). But goal importance and goal attainability are also dynamic constructs that adapt over time and alter along with changing life circumstances and key developmental tasks (Freund & Ebner, 2005). It has been suggested that one's goal pursuit needs to be stable to be attainable but flexible at the same time to adjust to new circumstances (Brandtstädter & Rothermund, 2002), which is referred to as the stability–flexibility dilemma (e.g. Bak & Brandtstädter, 1998). Given the present findings, we conclude that both the stability component and the flexibility component were present. More specially, we found a positive link between earlier goal importance and later goal attainability in all six life–goal domains, even when controlling for the stability of goal attainability. In other words, people who rate their goals as important and are motivated to accomplish these goals perceive these goals as more attainable (Koo & Fishbach, 2008).

However, the inverse link of goal attainability on later goal importance was less consistent across life–goal domains. Goal attainability was predictive of importance of health, personal–growth, and social–relation goals but not of prosocial–engagement, status, and work goals. It is possible that goal attainability might not have been the leading mechanism for later goal importance in these domains for two reasons: First, in the case of work goals, work importance is emphasized by society during major parts of the lifespan, which makes work an important and socially desirable endeavour for most people. Thus, work goals need, in some sense, to be important irrespective of whether these goals are perceived as attainable. Second, in the case of prosocial–engagement and status goals, these life–goal domains might also reflect the personality of the individual who pursues them rather than the goals’ attainability. The perceived importance of these goals might therefore be fuelled less by their attainability than by their pursuer's ideals and values, that is, by other aspects of that individual's personality. Future research is needed to shed light on these preliminary explanations for why goal attainability was not consistently predictive of later goal importance, above and beyond previous goal importance.

In summary, we maintain that goal importance and goal attainably are fairly stable across time and share a longitudinal association, which is bidirectional for most goal domains. Except for a moderating age effect in the status–goal domain, we found no moderating effect on the association of age, which leads us to assume similar goal dynamics for people across adulthood. The bidirectionality might help people exhibit high control over their development and supports the dual–process framework of an assimilative mode and accommodative mode (e.g. Brandtstädter & Rothermund, 2002; Brandtstädter, Wentura, & Rothermund, 1999): If people rate a certain life goal as important, they invest more in this goal, which makes the goal more likely to be perceived as attainable. In contrast, if a goal appears less attainable, people devalue the importance of this life goal. We thus maintain that goal dynamics help people adjust goals to personal and contextual resources and that this dynamic is largely age independent.

Goal outcomes: Goal importance and goal attainability as predictors of subjective well–being

Life goals guide and motivate a person's behaviour, which is likely to relate to that person's well–being (Deci & Ryan, 2000). In the present study, we tested the predictive effects of goal importance and goal attainability on cognitive–evaluative (i.e. life satisfaction and domain–specific satisfaction) and on affective components of subjective well–being (i.e. positive and negative affect). We tested these effects for both 2–year and 4–year intervals.

Strengths and limitations

One of the strengths of the present research is the sample covering major parts of the adult lifespan, which enabled us to test the concept of developmental tasks in a large age–heterogeneous sample. Second, our study spans multiple measurement points, providing us with an insight into the longitudinal interplay between goal importance, goal attainability, and subjective well–being. Third, most research investigating the links between goals and satisfaction in the social domain (e.g. relationship satisfaction) has focused on relationship–specific goals. We extended past research by including goals across important life domains as predictors of satisfaction in social and other domains. Fourth, as well as clustering life goals into extrinsic and intrinsic goal components, as has often been done when using the Aspiration Index (e.g. Kasser & Ryan, 1996), we factor analysed goals and grouped them into six thematic domains to obtain a more fine–grained picture of life goals and their predictive validity on outcomes.

Some limitations should be borne in mind when interpreting the results. First, even though our sample was large and age heterogeneous, it represents a convenience sample with rather cognitively fit older adults. In addition, people from different cultural, socio–economic, and educational backgrounds might evaluate life goals and their attainability differently.

Second, we exclusively refer to self–report measures. Future research would benefit from including multi–method approaches by gathering additional information, such as observational data, other–reports, objective major life events occurring in a given life stage, or experience sampling and combining these data with self–reports.

Third, owing to the model misfits mentioned in Footnote 2, it is a limitation of this study that analyses were conducted by making use of manifest variables rather than latent variables. Hence, future studies would benefit from (i) improving their measurement of the study's key variables, which (ii) would make it possible to create latent variables. This, in turn, would allow to apply an MSEM approach to account for measurement errors and to calculate cross–lagged models with multiple outcomes (e.g. life satisfaction at Time 2 and life satisfaction at Time 3 predicted by goal importance and goal attainability at Time 1).

Fourth, we assessed people's life goals in that we explicitly asked participants about the importance and attainability of various life–goal domains. In doing so, we might have limited participants’ option to mention life–goal domains that were not captured by the pre–existing goal domains. Hence, age–related differences in goal domains are only applicable to the goal domains that we asked for, and age–related effects occurring in other domains might have been masked. In a similar vein, our framing of goals was on a more abstract (e.g. ‘pursuing one's own occupational career’) than concrete (e.g. ‘becoming a successful medical doctor in the area of cardiology’) phrasing level (Little, 1989). An abstract phrasing might leave more room for individual interpretation, which might imply different meanings for different people and hence suggest different implications for a person's concrete assimilation and accommodation mode (Brandtstädter & Rothermund, 2002) as well as subjective well–being.

Fifth, we did not prompt participants to prioritize certain life goals in their evaluations. In everyday life, people might experience conflict between their life goals, such as between the work and family domains (Wiese & Freund, 2005). This conflict might be particularly prominent and relevant in the so–called rush hour of life (Bittman & Wajcman, 2000), which characterizes young adulthood as a life stage in which several life goals are pursued simultaneously. Future research might more closely look at the hierarchy of goals that a person holds and how this prioritization might change across the lifespan and is differently related to well–being.

Sixth, taking into account ontogenetic and historical contextualism, we cannot preclude that our findings would be better explained by cohort than age (e.g. Grob, Krings, & Bangerter, 2001; Staudinger & Bluck, 2001).

Finally, it was beyond the scope of the present study to answer questions pertaining to explanatory mechanisms: Why are goal importance and goal attainability reciprocally linked to each other and what mechanisms, for instance, translate social–relation goals into satisfaction with the social domain? Investigating such processes might further provide an understanding of how life goals are embedded in people's everyday life across adulthood.

Future steps

Limitations reveal impetus for future research. While our study included goal importance and goal attainability, it cannot speak to concrete goal progress, which would add to the characteristics of goal–appraisal dimensions (Brunstein et al., 1999). In this attempt, future studies might also employ a microfocus on goal processes. Studying daily or weekly life–goal processes could provide a deeper insight into the concrete implementation and progress of life goals in daily life.

A second approach worth studying is to include implicit goals and motives (e.g. Schultheiss, 2001) so as to thoroughly understand the role of motivational processes across adulthood and in the realms of different life domains. It could be argued that the pursuit of explicit goals that are not in alignment with implicit goals is unlikely to nourish well–being (Brunstein, Schultheiss, & Grässman, 1998).

A third point of further inquiry is to measure goal–relevant support that is provided by close others. It has been argued that these social resources stimulate the setting of life goals and assist with the achievement of these goals (e.g. Brunstein et al., 1996; Diener & Fujita, 1995; Fitzsimons, Finkel, & VanDellen, 2015). In particular, as close others might foster or hinder one's personal development across the lifespan (e.g. Bühler, Weidmann, Kumashiro, & Grob, 2018; Drigotas, 2002), future studies might include emotional and instrumental support that is provided by a person's social network for approaching a certain goal and might investigate how this support changes across the lifespan.

Fourth and finally, future research might benefit from asking participants in an open format for their life goals (for a similar approach, see Massey, Gebhardt, & Garnefski, 2009). For instance, within the domain of work, results might vary depending on whether, for instance, work goals refer to ‘getting promoted’ compared with ‘working less’. Future research might investigate how these open–format goals are related to age and promote later satisfaction within a given domain.

Conclusion

As Allport (1955) put it, mature striving is linked to long–range goals. With the aim of positioning life goals in the context of adult development, we applied a developmental perspective to goal content, the dynamics of goal importance and goal attainability, and goal outcomes. Given the results, we conclude that age matters for life goals in the following ways: First, people rate those life goals as important and attainable that correspond to the developmental tasks of their current life stage. Second, goal–related regulatory dynamics (such as balancing goal importance and goal attainability) seem to support people in the pursuit of their personal life goals across the entirety of adulthood and, consequently, to actively shape personality development. Finally, major life goals—particularly their attainability—seem to be of long–term importance for cognitive and affective components of well–being and show thematic predictive validity for domain–specific satisfaction across the adult lifespan.