Changes in personal work goals in relation to the psychosocial work environment: A two-year follow-up study

Abstract Associations between changes in the psychosocial work environment and changes in personal work goals were investigated in a two-wave, two-year longitudinal study. Psychosocial work environment was studied within the context of the Effort–Reward Imbalance model (ERI; Siegrist, 1996). The participants consisted of 423 young Finnish managers. Their most important personal work goals were categorized into seven content categories of competence, progression, well-being, job change, job security, organization, and finance at both measurement times. There were differences, especially in changes in the career opportunities factor of reward, between participants whose goals changed during the study. First, those who became engaged in job change goals also reported a decrease in rewards between T1 and T2, whereas participants who became engaged in competence or organizational goals also reported an increase in rewards over time. Second, participants who disengaged from job change goals reported an increase in reward and a reduction in effort–reward imbalance. Finally, a reduction in rewards and an increase in ERI were accompanied by disengagement from job security goals. The study highlights the central role of psychosocial factors, and in particular the reward factor of career opportunities, in goal pursuit. These results can have implications for the occupational well-being of employees.


Introduction
The impact of the psychosocial work environment on occupational health and wellbeing has been indicated by a large body of occupational stress research (for reviews, see Cooper, Dewe, & O'Driscoll, 2001;Kahn & Byosiere, 1992;Stansfeld & Candy, 2006;Tsutsumi & Kawakami, 2004;van Vegchel, de Jonge, Bosma, & Schaufeli, 2005). The present study provides a new perspective to the current occupational stress theories by extending the investigation to future-oriented, work-related aspirations of managers (i.e., personal work goals) over a two-year period (2006Á2008). Previous personal goal research has already shown that the contents of personal life goals (Salmela-Aro & Nurmi, 2004) and appraisals of personal work goals (e.g., Harris, Daniels, & Briner, 2003;Maier & Brunstein, 2001;Pomaki, Maes, & ter Doest, 2004) accounted for individual differences in occupational well-being. Moreover, the psychosocial work environment contributed to the contents of personal work goals (Hyvö nen, Feldt, Tolvanen, & Kinnunen, 2010), which in turn were associated with occupational well-being indicators of burnout and work engagement in recent crosssectional studies (Hyvö nen, Feldt, Salmela-Aro, Kinnunen, & Mäkikangas, 2009;Hyvö nen et al., 2010;see also Hyvö nen, 2011).
The added value of the present study stems from investigating the contents of self-articulated goals, which have previously drawn less research attention than goal appraisals (Nurmi, Salmela-Aro, & Aunola, 2009;Pomaki et al., 2004;Salmela-Aro & Nurmi, 2004) or, for example, goal orientation in different performance contexts (e.g., Brett & VandeWalle, 1999; for a review, see DeShon & Gillespie, 2005). To our knowledge, no previous research has specifically addressed the contents of personal work goals in a longitudinal study. Thus, the mixed methods approach (e.g., Creswell & Plano Clark, 2007) of this study can shed new light on changes in the contents of ideographic goals over time as well as on how psychosocial factors, which describe possible stressors of the work context within which participants set their goals, are associated with the pursuit of goals. Our target group comprised 433 employees, all of whom were in managerial positions and age 36 years or younger at the onset of the study in 2006. This study can therefore yield potentially valuable information about the psychosocial factors at work that guide interests and development in the early phases of careers in management. Due to the key position of managers in organizations, research into their goals can also stimulate further understanding of leadership and organizational behaviour (Bateman, O'Neill, & KenworthyU'Ren, 2002), and of their occupational well-being, which have also been found to impact also the well-being of their subordinates (e.g., van Dierendonck, Haynes, Borrill, & Stride, 2004; for a review, see Skakon, Nielsen, Borg, & Guzman, 2010).

Psychosocial work environment
An occupational stress model, the effortÁreward imbalance (ERI) model by Siegrist (1996), was utilized to investigate the primary psychosocial stressors in the work environment. The model is based on the reciprocal relationship between costs and gains in the workplace. Employees invest effort in fulfilling the demands and responsibilities, such as interruptions, work load, and overtime. In return for their effort, employees expect rewards, such as money, esteem, job security, and career opportunities. The model also accounts for the lack of reciprocity between efforts and rewards, which is represented as an imbalance of effort and reward at work. The lack of reciprocity might be present in situations where employees have fewer employment options (e.g., due to skills deficits or a poor job market), but also, for instance, when an employee has strategic ambitions, such as career progression . According to the ERI model, individual factors, namely overcommitment to work, could predispose the employee to investing an exceedingly high amount of effort into work, especially in a demanding work environment (e.g., De Jonge, van der Linden, Schaufeli, Peter, & Siegrist, 2008;Siegrist et al., 2004). Overcommitment is seen as the ''inability to withdraw from work'' (e.g., Siegrist et al., 2004).
The ERI model proposes that a prolonged lack of reciprocity can be detrimental to health, which has been supported by various research findings (for reviews, see Stansfeld & Candy, 2006;Tsutsumi & Kawakami, 2004;van Vegchel et al., 2005): For instance, a recent study among employees aged 35Á44 years indicated that a higher ERI was related to a lower heart rate variability, which is connected to an increased risk of coronary heart disease (Loerbroks et al., 2010). Furthermore, the ERI model has also been investigated in relation to indicators of work-related well-being, demonstrating that the components of the model contribute towards explaining variance in burnout symptoms (e.g., Dai, Collins, Yu, & Fu, 2008;Willis, O'Connor, & Smith, 2008), perceived work stress (Calnan, Wadsworth, May, Smith, & Wainwright, 2004), as well as vigour and dedication at work (Kinnunen, Feldt, & Mäkikangas, 2008). In terms of job attitudes, for instance, lower rewards were associated with lower job satisfaction (e.g., Calnan, Wainwright, & Almond, 2000), and a higher effort-reward imbalance with stronger turnover intentions (Kinnunen et al., 2008).

Personal goals within the work context
The theoretical framework of personal work goals is derived from previous research on personal action that has included personal action constructs (PAC; Little, 2007) such as personal projects (Little, 1983) and personal strivings (Emmons, 1986). These PAC units of analyses can be considered as ''middle level'' constructs in personality research, where evolutionary and personality trait theories feed into PAC theories, which together shape a person's identity and life narratives (e.g., Little, 2007; see also McAdams, 1995;McAdams & Pals, 2006;McGregor, McAdams, & Little, 2006). Personal goals can range from immediate plans to goals that last through an individual's life span (Little, 2007). Personal goals can be distinguished from goal orientation, which is often considered as a personal disposition (e.g., learning goal, performance-prove, performance-avoid orientations) which can impact the person's choice of goal contents (Brett & VandeWalle, 1999).
Personal goals have been investigated, for example, by categorizing them on the basis of goal contents into categories such as interpersonal, academic, work, intrapersonal, recreational, health, maintenance, and other (Little, 1983;Little & Gee, 2007). Another typical approach to goal analysis is to examine participants' goal appraisals on dimensions, such as importance, commitment, difficulty, or conflict (for a review, see Austin & Vancouver, 1996). More recently, the consideration of relational aspects of personal goals has also raised interest, for instance, in terms of analysing the nature of social ties related to goals (Salmela-Aro & Little, 2007). Our study pursued a closer analysis of participants' responses to an openended question regarding their work goal. A similar approach has been used by Bateman et al. (2002) to investigate the contents of multiple work goals of 75 top corporate leaders. Of the 10 work goal categories identified, the majority of goals related to the leadership role and organizational functioning (e.g., financial, customer, and operational goals; see Bateman et al., 2002). However, the personal goals category was the largest single category (18% of all goals) and referred to goals outside the business environment (e.g., career aspirations, wealth, and family). Thus, goals at work also incorporate hopes and wishes peripheral to work since boundaries between work and home are becoming increasingly blurred (Jones, Burke, & Westman, 2006).
In this two-year follow-up study, we expected to find the same goal content categories as were identified in the baseline study by Hyvö nen et al. (2009) with the same, albeit larger sample of young managers (N 0747). In this previous study, the contents of the most important personal work goals were coded into eight categories: competence (professional development and training; 28%), career progression (promotion and advancement; 21.7%), well-being (self-concerns, managing stress, job satisfaction, motivation; 13.9%), and job change (finding a new job or setting up a company; 12.6%). The smaller goal categories included job security (continuing working, a permanent employment contract; 6.8%), organization (success and performance of the project, team, department, or company; 5.1%), finance (pay rise, bonus; 3.6%), and an additional 8.3% of managers with no work goal mentioned. The goals varied from shorter term (e.g., finishing a project) to longterm goals (e.g., getting promoted within the next three years). Similar personal work goals have been identified by Wiese and Salmela-Aro (2008) in a study with 131 employees from a range of professional fields and employment positions. As opposed to the top leaders' work goals (Bateman et al., 2002), the goals focusing on the management and leadership tasks were less common among the young managers (i.e., organizational goals; 5.1%) who were still working mainly in lower or middle management positions. Instead, the young managers' goals reflected to a large extent career establishment that according to Super (1969Super ( , 1985Super ( , 1990 incorporates periods of stabilization, consolidation, and advancement. The present study: Personal work goals at the interface of person Á environment interactions The social ecological model of well-being (e.g., Little, 2000Little, , 2007 posits that personal goals reflect the continuous balancing of stable and dynamic personal and environmental features. Well-being and human flourishing is enhanced through the sustained pursuit of core goals in life. Since personal goals are influenced by social, cultural, and historical life contexts (e.g., Baltes, 1997;Freund & Riediger, 2006;Heckhausen & Schulz, 1995), which may change over time, previous personal goal research has also noted changes in goals across the life span. Opportunities and restrictions of a particular life stage channel personal goals, which can reflect the agegraded developmental tasks (Nurmi, 1992;Salmela-Aro, 2009;Salmela-Aro, Aunola, & Nurmi, 2007). Brandtstädter (2009) has approached goal pursuit and adaptation as assimilative and accommodative processes, describing how a person adapts to the discrepancy between the factual circumstances (or real self) and desired outcomes (or ideal self). In brief, a person might strive to change the situation to the desired outcome (assimilation), or accommodation could be necessary by adapting personal goals to the prevailing situation. Goal reconstructions could be fundamental in adapting to the changing challenges of the work environment and reflect ''career adaptability'' which highlights the adaptive processes required in career development (Savickas, 1997). Therefore, we expected that the personal work goals of many of the participants will have changed during the two-year follow-up period, reflecting the ongoing negotiation of the opportunities and demands of life contexts.
Within the occupational domain, the psychosocial stressors in the work environment have been found to contribute to the contents of personal work goals in cross-sectional analyses based on this dataset (Hyvönen et al., 2010). More specifically, perceiving the work environment as more strenuous (i.e., reporting high effort, low reward, or high effortÁreward imbalance) was associated with job change and well-being goals. These personal work goals were also related to the lowest level of occupational well-being (Hyvö nen et al., 2009;Hyvö nen et al., 2010). Thus, in line with personal goal theories (e.g., Brandtstädter, 2009;Little, 2007) and research (e.g., Hyvö nen et al., 2010), reducing resources in the work context could instigate goal changes, for instance towards job change goals, in order to improve adaptation and occupational well-being.
In the previous cross-sectional study with this dataset, the most favourable work environments (the highest level of rewards, and low effortÁreward imbalance), in turn, were associated with organizational goals, which also related to the highest level of occupational well-being (Hyvö nen et al., 2010). These findings paralleled, firstly, previous research on favourable psychosocial work environment and positive occupational well-being outcomes (e.g., Dai et al., 2008;Kinnunen et al., 2008), and secondly, research on goal appraisals indicating that favourable conditions for goal attainment at work predicted positive job attitudes (job satisfaction and organizational commitment) among new employees who were committed to their goals in an eight-month follow-up study (Maier & Brunstein, 2001). Thus, favourable changes in the psychosocial work environment (i.e., a reduction of psychosocial stressors) could be related to engaging in goals focused on the performance of the organization (i.e., organizational goals) reflecting the resources available that can be directed towards the managers' leadership tasks at hand.
Besides changes in the prevailing psychosocial work environment, also other changes, such as an unemployment period, changing jobs, or getting a promotion could instigate changes in the work environment and personal work goals. Therefore, these career events were also taken into consideration in the present study. In the previous study by Hyvö nen et al. (2010), overcommitment had less contribution towards goal contents and thus, we focused on effort, reward, and ERI in the present investigation. More specifically, our research questions and hypotheses were as follows: (1) To what extent have personal work goals changed during the two-year period from baseline to follow-up (2006Á2008)? We expected to find the same seven goal categories Á competence, progression, well-being, job change, job security, organization, and finance Á as were found two years earlier (see Hyvö nen et al., 2009) (Hypothesis1a). However, we also expected to see changes in personal work goals of the participants, reflecting ongoing negotiation of the opportunities and demands in the young managers' life contexts (Hypothesis 1b). (2) Is there an association between changes in the psychosocial work environment (effort, reward, and effortÁreward imbalance) and changes in personal work goals? On the basis of previous theory (e.g., Little, 2007) and research (Hyvö nen et al., 2010), we hypothesized that unfavourable changes in the work environment (low reward, high effort, and high effortÁreward imbalance) will be associated with engaging in job change or well-being goals (Hypothesis 2a). We further expected that favourable changes in the work environment (high reward and low effortÁreward imbalance) will be associated with engaging in organizational goals (Hypothesis 2b).

Participants and procedure
The study was conducted using both a questionnaire and a qualitative approach. The questionnaire study was conducted with two measuring points, Spring 2006 (T1) and 2008 (T2). In January 2006, the sample was taken from the membership registers of two Finnish national labour unions (the Union of Salaried Employees and the Union of Professional Engineers). The original sample included 1904 members who were all younger than 36 years and whose professional title referred to a management position. Questionnaires were posted to the home addresses and 933 questionnaires were returned. Of the respondents, 186 were not in management or in employment (e.g., they were on maternity leave, studying, or had been unemployed for over three months) and therefore, these respondents were excluded from the final sample. The response rate was 43.4% in 2006 (for more detail, see Hyvönen et al., 2009). During the data collection in 2006, 126 participants had indicated that they no longer wished to participate in the research and therefore the follow-up questionnaires in 2008 were sent to 621 participants. In total, 433 questionnaires were returned, which yielded a response rate of 69.7%. That is, of the original sample (N 0747) in 2006, 58.0% of participants also responded at the follow-up study in 2008. Of those 433 respondents, seven were unemployed and three had not responded to the study variables, and thus were omitted from the final sample (N 0423). At T1, the average age of the participants was 31 years (range 24Á35 years, SD03.2 years), a large majority of them were men (83.9%), and 7.8% were in upper, 49.4% in middle, and 42.8% in lower management. A large majority of the participants had a permanent employment contract (93.3%), and 30.3% had experienced unemployment or lay-offs during the period following their graduation up to 2006. Between T1 and T2, 7.0% of participants had experienced career disruptions (unemployment or lay-offs) and 28.9% (n0118) reported that they had changed jobs on their own initiation. In addition, 34.8% (n 0142) of participants had been promoted since 2006.

Attrition analyses
The attrition analysis showed that the respondents did not differ in terms of gender from nonrespondents at T1 (see also Hyvö nen et al., 2009). The data of the nonrespondents' age were only available for the members of the Union of Salaried Employees; these respondents (n 0331) did not differ from nonrespondents (n0379) in age. At T2, the respondents (n0433) did not differ from the nonrespondents (n0314) in terms of gender, x 2 (1)03.79, ns; managerial level, x 2 (2)00.62, ns; employment contract, x 2 (1)00.09, ns; or career disruptions before 2006, x 2 (1) 00.76, ns. No significant differences emerged in relation to effort, t(744)0(0.24, ns; reward, t(745)00.73, ns; or effortÁreward imbalance, t(744)0(0.53, ns. However, the x 2 -test indicated that participants with job security goals and those with no work goals at T1 were slightly underrepresented among the respondents who participated at T2 and overrepresented among the respondents who had only participated at T1, x 2 (7) 017.2, pB.05.

Measures
Personal work goals were inquired about with an open-ended question: ''Write down your most important personal goal that relates to your work or career'' (Hyvö nen et al., 2009;Hyvö nen et al., 2010). At T1, three independent coders thematically categorized the participants' responses using a generic and data-driven qualitative analysis that did not rely on preset categories (e.g., Creswell & Plano Clark, 2007). Seven content categories of goals were found: competence, progression, well-being, job change, job security, organization, and finance. In addition to these, a category was formed which consisted of participants who had either not mentioned a work goal or mentioned a goal unrelated to work or career (i.e., the ''no work goal'' category). Each participant could be in only one of the eight goal categories (for further details about the different stages of coding, see Hyvönen et al., 2009). A fourth independent coder applied the categorization outlined by the first three coders and the intercoder agreement of the goal categories was 92%. The AC 1 coefficient was .92 (CI 00.88, 0.94), indicating an excellent inter-coder agreement (Gwet, 2008). The same goal categories were also found at T2 and no new categories emerged. At T2, the inter-coder agreement of the goal categories was 94% between two coders and the AC 1 coefficient .94 (CI 00.90, 0.96). Of the two coders, one had been involved in the coding at T1, but the other coder had not. These two coders decided the most suitable categories for the remaining 6% of goals together, which had been coded into different categories during the first stage of independent coding at T2.
Effort, reward, and effort-reward imbalance were measured by a scale developed by Siegrist et al. (2004). The good construct and discriminant validity of the Finnish version of the ERI scale has been reported previously by Kinnunen et al. (2008). Effort was assessed with five items describing the demands in the workplace (e.g., ''I have constant time pressure due to a heavy work load''). If the respondent answered the question affirmatively, they were asked to rate the impact of effort from ''not at all distressed'' to ''very distressed''. The scale ranged from 1 to 5: (1) does not apply; (2) does apply, but I am not at all distressed; (3) does apply, and I am somewhat distressed; (4) does apply, and I am distressed; (5) does apply, and I am very distressed. A higher mean score of effort indicates more effort invested at work. The internal consistency (Cronbach's alpha) for effort was .88 (M03.16; SD 00.98) at T1 and .90 (M03.05; SD 01.00) at T2.
Reward was assessed with 11 items describing esteem (5 items, e.g., ''I receive the respect I deserve from my superiors''), career opportunities (4 items, e.g., ''Considering all my efforts and achievements, my salary/income is adequate''), and job security (2 items, e.g., ''My job security is poor'', reverse scored). A similar rating procedure was used as described above for the effort scale, and a higher mean score of reward indicates more rewards received at work. The Cronbach's alphas for reward were .86 (M 04.04; SD00.73) at T1 and .88 (M 04.13; SD00.73) at T2.
The imbalance of effort and reward is described as an ERI ratio. The ERI ratio is calculated by first multiplying the sum score of reward with a correction factor (see Niedhammer, Tek, Starke, & Siegrist, 2004;Siegrist et al., 2004). Because 5 items were used to assess effort compared to 11 items to assess reward, the correction factor in this study was 0.4545. The sum score of effort is then divided by the corrected sum score of reward. A score close to ''0'' indicates favourable conditions, where received rewards outweigh the effort invested at work. In turn, a score over ''1'' indicates unfavourable conditions, where more effort is spent than rewards expected or received in return. As recommended in previous studies Siegrist et al., 2004), a continuous variable of the ERI ratio was used for the analyses. The mean of the ERI ratio was 0.83 (SD00.39) at T1 and 0.80 (SD00.40) at T2.
Background variables measured at T1 included gender (male/female), managerial level (upper/middle/lower), employment contract (permanent/fixed-term), and career disruptions before T1 (no/some periods of unemployment or lay-offs since graduation). On the basis of the previous, cross-sectional study with this sample of participants (Hyvö nen et al., 2009;Hyvönen et al., 2010), these background variables were related to the outcome measures used, and therefore, were also controlled for in this study. Career events measured at T2 incorporated information about the workrelated experiences of participants between T1 and T2 and included three dichotomous variables: career disruptions T1ÁT2 (no/some periods of unemployment or lay-offs); job changes on one's own initiation T1ÁT2 (no/yes); and promotions T1ÁT2 (no/yes).

Analyses
Spearman correlation coefficients were calculated for categorical variables (gender, managerial level, employment contract, career disruptions before T1, career disruptions T1ÁT2, job changes T1ÁT2, promotions T1ÁT2) and Pearson correlation coefficients for continuous variables (effort, reward, ERI ratio). The changes in the most important personal work goals (i.e., competence, progression, well-being, job change, job security, organization, and finance) were investigated by forming two goal change groups for each goal category: (1) Participants who had not mentioned a particular goal at T1, but mentioned it at T2 (i.e., they became engaged in the goal); (2) Participants who mentioned a goal at T1, but not at T2 (i.e., disengaged from the goal). Thus, the further analyses excluded those participants who mentioned similar goals at both measurement points and those participants who had not mentioned a work goal. Separate analyses were performed for the two types of goal change groups in order to compare the differences between the goal categories.
The two goal change groups were investigated in relation to changes in the investigated ERI components (effort, reward, ERI ratio, esteem, career opportunities, and job security) by means of the Analyses of Covariance (ANCOVA) where all background variables could be controlled for and partial eta-squares (h 2 ) reported. In order to compare the degree of change in the ERI components, difference scores were computed to describe the change between T1 and T2. The difference scores (i.e., the dependent variables in ANCOVA analyses) were formed by deducting the mean score of an ERI component at T1 from the mean score at T2 (e.g., reward at T1 was deducted from reward at T2). The goal categories at T2 were the betweenÁsubjects factors when calculating the differences in the degree of change in the ERI components among the participants who engaged in a goal, whereas the goal categories at T1 were the betweenÁsubjects factors when calculating the differences among the participants who disengaged from a goal. Bonferroni pairwise comparisons were calculated to detect which goal categories differed from each other in the degree of change in relation to the investigated ERI component.

Descriptive results
Table 1 depicts the intercorrelations among background variables, effort, reward, and ERI ratio. Female gender, lower managerial levels, and having a fixed-term employment contract were related to lower effort and ERI ratio at T1, but only lower managerial levels were related to lower effort at T2. Job changes T1ÁT2 related to lower reward at T1 and to lower effort and ERI ratio at T2. In addition, experiencing career disruptions T1ÁT2 correlated with lower effort and ERI ratio at T2. Furthermore, promotions T1ÁT2 correlated with higher effort and reward at T2. Of the background variables, only career disruptions before T1 had no association with the ERI components. Nevertheless, career disruptions before T1 were controlled for in the further analyses because of its associations with personal work goals in the previous study (see Hyvö nen et al., 2009).

Changes in personal work goals
The distribution of personal work goals in 2006 and 2008 is shown in Table 2. In line with our expectations (H1a), we found the same goal categories in 2008 as in the first study phase in 2006. The sizes of the categories also remained similar in both study phases except for the category for organizational goals, which almost doubled in size. Our expectation regarding goal change (H1b) also received support since the majority of participants (67%; n 0285) changed their most important personal work goal during the two-year period of the study. As can also be seen in Table 2, the percentages of participants with similar goals in both study phases were fairly low, ranging from 12.5% (participants with no work goal) to 45.5% (participants with competence goals). These participants were omitted from further analyses, which concentrated on investigating changes in personal work goals. Thus (after the participants who had not mentioned a personal work goal at either occasion had been excluded), there were 273 participants who were engaging in various personal work goals at T2 that they had not mentioned at T1, and 264 who disengaged from their original goals mentioned at T1.

Changes in the ERI components and personal work goals
The results of ANCOVAs can be seen in Table 3 showing the relationships between goal change groups and changes in the ERI components in addition to adjusted means and standard deviations from both study phases. These analyses were adjusted for covariates of background variables (gender, managerial level, employment contract, career disruptions before T1, career disruptions T1ÁT2, job changes T1ÁT2, and promotions T1ÁT2). Among participants who engaged in goals, the goal categories at T2 were significantly associated with the degree of change in reward. In addition, among the participants who disengaged from goals, the goal   categories at T1 were significantly associated with the degree of change in effort, reward, and ERI ratio. First, our expectation H2a received partial support. Unfavourable changes in the work environment (i.e., a decrease in rewards) were observed among participants engaging in job change goals at T2, who differed significantly from participants who had become engaged in competence or organizational goals in at T2. That is, participants who engaged in competence or organizational goals reported an increase in reward, a finding that was also partly in line with our expectations (H2b) regarding favourable changes in relation to organizational goals.
Second, favourable changes in the work environment, that is, an increase in reward and a reduction in ERI ratio, was characteristic of participants who disengaged from job change goals as being the most important personal work goal at T1. These participants differed from participants disengaging from job security goals, a change that was linked to a reduction in reward and an increase in ERI ratio. The degree of positive change in rewards for participants who disengaged from job change goals over the period of the study was also higher than for those who disengaged from competence goals as being the most important personal work goal at T2; only a slight reduction in rewards was observed to be linked to disengagement from competence goals. Bonferroni comparisons did not detect significant differences between the goal categories in the degree of change in effort for participants who had disengaged from goals by T2, although the F-value reached significance (see Table 3).
Since the reward component of the ERI scale is formed by three sub scales (esteem, career opportunities, and job security), further ANCOVAs were calculated to test the associations between the goal categories and the specific reward factors. Of those factors, changes in career opportunities showed the strongest associations with changes in personal work goals. First, among participants who became engaged in goals, the goal categories at T2 were significantly associated with the degree of change in career opportunities, F(7,268)03.35, pB.01, partial h 2 0.08, whereas no significant associations were found between reward factors and the degree of change in esteem, F(7,268) 01.33, ns, or in job security, F(7,268) 01.70, ns. Bonferroni comparisons yielded the same results for career opportunities as for the total mean score of reward; participants who became engaged in job change goals reported a significantly higher degree of change in career opportunities (i.e., a decrease in career opportunities) than participants who engaged in competence or organizational goals; they reported an increase in career opportunities.
Second, among participants who disengaged from goals, the goal categories at T1 were significantly associated with the degree of change in esteem, F(7,268)02.33, pB.05, partial h 2 0.06, and career opportunities, F(7,268) 02.71, pB.01, partial h 2 0.07; but no significant association was found with the degree of change in job security, F(7,268) 01.57, ns. Again, Bonferroni comparisons showed the same significant pairwise comparisons for career opportunities as for the total mean score of reward; participants who disengaged from job change goals reported a higher degree of change (an increase) in career opportunities than participants who disengaged from competence or job security goals, which were linked to a reduction in career opportunities. However, Bonferroni comparisons did not detect significant differences between the goal categories in the degree of change in esteem among participants who disengaged from goals, although the F-value reached significance.

Discussion
This study shed light on the relations between the psychosocial work environment and personal work goals over a two-year follow-up period, and provided support for the following expectations. The same categories of personal work goals were found at T2 as at T1 without any new goal categories emerging, but the personal work goals were observed to change. The clearest results on the association between changes in psychosocial work environment and personal work goals were seen in regard to a reduction in rewards Á particularly in terms of a reduction in career opportunities Á observed among participants who became engaged in job change goals over the two-year period. This is in contrast to the participants who became engaged in competence or organizational goals, who reported an increase in rewards.
Furthermore, favourable changes were reported in the psychosocial work environment, including an increase especially in the career opportunities factor of the reward component as well as a reduction in ERI among participants who disengaged from job change goals. This finding was highlighted in relation to the participants who disengaged from competence goals, and particularly in relation to those who disengaged from job security goals, who reported a notable reduction in reward and an increase in effortÁreward imbalance. Thus, work characteristics measured within the context of the ERI model (Siegrist, 1996) were found to be associated with personal work goals over the two-year period. This finding can also be considered important from the perspective of occupational well-being. For example, on the basis of previous cross-sectional analyses with this dataset, the contents of personal work goals are also associated with the occupational well-being indicators of burnout and work engagement (Hyvö nen et al., 2009;Hyvö nen et al., 2010). Therefore, occupational well-being could be supported by taking into account the pursuit of personal work goals at the interface of environmental and personal interactions.

Prominent personal work goals among young managers
Personal work goals in the different goal categories changed as expected on the basis of previous theories proposing that contextual features Á with changing demands and opportunities Á are relevant in the pursuit of personal goals (e.g., Little, 2007;Salmela-Aro, 2009). However, the most prominent personal work goals were those towards competence and progression. Around 40% of the participants in those categories focused on competence or career advancement at both measurement times. These goals could reflect vocational development related to career establishment in young adulthood (e.g., Super, 1969Super, ,1985; see also Savickas, 1997). Especially for young managers, the pursuit of a career can be a primary goal in their current life context and they may feel that career decisions are central to their future life design: For example, work-related personal goals become a central focus of this age group in addition to goals related to family and health . The assimilation processes (Brandtstädter, 2009), for instance persistence towards professional development, could therefore be particularly typical in young adulthood. Furthermore, competence and progression goals have been linked to reasonably good rewards and occupational well-being when compared to, for example, well-being and job change goals (Hyvö nen et al., 2010). There might be less pressure to change these goals as a result of psychosocial stressors at work. In accord with the social ecological model (Little, 2000(Little, , 2007, these participants might represent the employees in managerial positions in young adulthood working in an environment where the pursuit of personally salient, age-relevant goals can be sustained, which is also manifested in advantageous occupational well-being. Another noteworthy observation was that the number of participants who mentioned organizational goals had almost doubled by T2 (n 024 at T1, and n 043 at T2). Through increasing work experience and career progression, young managers can gain a wider perspective of their responsibilities and role within the organization. That is, the increasing number of goals towards performance and success of the team or organization that we observed could signify that more participants had reached a level where they felt satisfied with their competence and professional position. In a previous cross-sectional study with this sample of young managers, organizational goals were associated with the most favourable work environment and with the highest level of occupational well-being (Hyvönen et al., 2010).

Changes in psychosocial work environment and personal work goals
Of the different features of the psychosocial work environment investigated, rewards from employment, and particularly career opportunities, showed the strongest associations with changes in personal work goals. The reward factor of career opportunities includes the participants' perceptions of having adequate salary and a position that matches their education and training background, as well as good future work prospects and possibilities for promotion. Attention was especially drawn towards participants with job change goals. This study suggests that a reduction in career opportunities over the two-year study period was linked to engaging in job change goals instead of engaging in personal work goals focused on competence and organization.
This finding reiterated the results of the previous cross-sectional study with this dataset (Hyvö nen et al., 2010), in which rewards contributed to focusing on job change goals. Job change goals could be a response to a reduction in resources that, according to Brandtstädter (2009), can facilitate accommodative processes towards alternative goals. In a similar vein, previous research has found that perceiving a work environment as unfavourable for attaining personal goals predicted lower organizational commitment among those employees who were committed to their goals (Maier & Brunstein, 2001). Therefore, it is possible that the participants experienced their work environment as increasingly unsupportive of the attainment of personal work goals, for instance towards competence or career progression, and began to re-evaluate their options. Moreover, on the basis of previous research, these increasing psychosocial stressors at work, as manifested in reducing rewards, could have a direct impact on occupational well-being (e.g., Dai et al., 2008;Willis et al., 2008), as well as an indirect deleterious effect through job change goals on (higher) burnout and (lower) work engagement (Hyvö nen et al., 2010). A change towards a more favourable work environment was in turn observed among those participants who disengaged from job change goals when compared to participants disengaging from competence or job security goals. That is, an increase in rewards, and particularly in career opportunities, was reported in conjunction with a reduction in ERI in participants disengaging from job change goals.
The participants engaging in competence or organizational goals considered their work environment as more rewarding at T2. This result was also partly in line with the cross-sectional analyses based on the same dataset (Hyvö nen et al., 2010) in which engaging in organizational goals was associated with the highest level of reward and occupational well-being. Similarly, competence goals also were connected to a favourable psychosocial work environment and a good level of occupational well-being. The current study therefore provided further longitudinal evidence for the key role of rewards, and particularly of career opportunities, in the workplace; higher rewards were associated with goals that encourage commitment to the organization, as well as to professional development and training. This study gave less support for the role of effort in changes in personal work goals. Although statistically significant differences in changes in effort were detected among participants who disengaged from goals, no significant differences emerged in pairwise comparisons between the personal work goals. This could be because managers had accepted the fact that their job was psychologically demanding (Kinnunen et al., 2008), and therefore career opportunities might be more pivotal in regard to personal work goals.
These changes in personal work goals Á that is, disengaging from job change and/ or engaging in competence or organizational goals Á appear generally to be positive processes reflecting goal accommodation (Brandtstädter, 2009) and career adaptability (Savickas, 1997). A change in the psychosocial stressors of the work environment (e.g., by changing jobs or through improvements in the existing workplace) could be creating new opportunities for directing personal resources at work. Alternatively, it is possible that participants who began focusing on more performance-oriented goals perceived an alleviation in the stressors at work and an improvement in occupational well-being.
In contrast, the participants who disengaged from job security goals between T1 and T2 reported a reasonably favourable psychosocial work environment at T1, but perceived unfavourable changes in it by T2, seen as a reduction in reward and an increase in effortÁreward imbalance. According to the cross-sectional analyses by Hyvö nen et al. (2009) with this sample, participants with job security goals were also more likely to be in fixed-term employment and had experienced career disruptions. These participants might perceive that their expectations regarding career opportunities cannot be met in the current work environment, which could partly account for these changes. This finding could also reflect a view that continuing with the same employer would no longer be advantageous. Through the qualitative inspection of the personal work goals of these participants, better opportunities or a less stressful working environment were mentioned frequently (i.e., over the study period participants shifted towards progression, well-being, and financial goals).

Limitations and directions for future research
Several limitations restrict the inferences that can be drawn from the findings. First, in terms of the generalizability of the results, having a specific target population (i.e., young Finnish managers who were mainly men and employed in technical fields, with permanent employment contracts), impacts on the extent to which these findings can be applied to other employees in other countries with different employment opportunities. For instance, organizational goals in this research refer quite specifically to leadership and managerial tasks, which were not identified as an independent category of personal work goals in the study by Wiese and Salmela-Aro (2008). The second main limitation of the research was that only the most important personal work goals were investigated, although managers are likely to have multiple personal work goals (Bateman et al., 2002). Therefore, the findings of this study only relate to the participants' central focus at the workplace and do not add to knowledge regarding their more complex goal structures; participants' personal work goals are likely to be interrelated and hierarchical in nature. Goal priorities can change at different points in a career and other personal work or life goals could also be important in terms of the work environment and occupational well-being. Third, this study cannot establish causal relationships between the psychosocial stressors at work and personal work goals, since we were only able to examine concurrent changes in psychosocial work environment and work goals. It should also be noted that drawing inferences on the basis of some of these findings should be done with caution due to the small group sizes, such as the group of participants disengaging from job security goals.
One of the most informative avenues for future research would be to follow changes in the psychosocial work environment and personal work goals over a longer period of time. This should include an investigation of the interrelatedness and hierarchy of personal work and life goals in order to build a more complete picture of the relationship between work environment and goal structures. For instance, it could be that for some participants (e.g., those with job change goals) accumulating work stressors could jeopardize career opportunities, development and well-being in the long run. Also, having organizational goals could reflect development that has been promoted through longer work experience. In addition to important career events, other life events and transitions, such as the transition to parenthood, could have a significant impact on personal work goals in this agegroup of participants and should be taken into consideration. An important perspective could also be offered by dispositional goal orientation (DeShon & Gillespie, 2005) and goal appraisals. For example, goal appraisals have been found to relate to the contents of work goals (Hyvö nen et al., 2009), as well as to wellbeing over and above the characteristics of the work environment (Pomaki et al., 2004).

Conclusions
Our findings suggest that the occupational stress theory of ERI (Siegrist, 1996) and Little's social ecological model (2000Little's social ecological model ( , 2007 can provide a theoretical backdrop for investigating the relationship between psychosocial stressors at work and the personal work goals of employees. Overall, it was found that participants' personal work goals changed over a two-year period in the early phases of their career. Associations between the psychosocial work environment and changes in personal work goals were most apparent in relation to job change goals. That is, unfavourable changes in the work environment (a reduction in career opportunities) were related to becoming engaged in job change goals, whereas favourable changes (an increase in career opportunities and a decrease in effortÁreward imbalance) were related to disengaging from job change goals. A practical implication of these results could be that, particularly by providing adequate professional status and salary, as well as good prospects for work and promotion, organizations could support young managers' endeavours in further professional development and training, as well as their commitment and focus on organizational goals, in the early stages of their careers. This is likely to have implications not only for the organization but also for the well-being of the employees.