W what tomorrow brings ? examining a model of antecedents of career expectations

Purpose: Analyze if the perceptions about policies and practices on human resources management (HRM), well-being at work and expectations of organizational future act as antecedents of career expectations in the organization. Originality/gap/relevance/implications: The question “What will my future be in this organization?” is a relevant part of working life. Thus, it is important to assess which variables influence future career expectations. Based on theoretical contributions and previous surveys about career, well-being and HRM, the model tested coordinates variables that are analyzed in isolate by literature. Key methodological aspects: In this quantitative survey, 305 professionals from a government agency completed an e-questionnaire with scales validated in Brazil. Data were analyzed through structural equations models. Summary of key results: The effects of well-being at work and expected future in the organization on the career expectations were positive and moderate, in the first case, and strong in the second case. The effect of perceptions of HR policies and practices on career expectations was completely mediated by well-being at work and expectations of organizational future. Key considerations/conclusions: Results support the relations theoretically conceived or found in previous qualitative surveys. In practical terms, data suggest that improving HR policies and practices can increase well-being at work and expectations of organizational future and, finally, foster positive career expectations in the organization. Considering that the research was limited to one organization, future investigations should analyze the model in other organizations.


INTRODUCTION
Career development has undergone substantial changes in the last few decades.Socio-economic changes and new organization management models brought about a scenario where career paths became unpredictable and non--linear (Magalhães & Bendassolli, 2013;Prince, 2003).This way, if over most of the 20 th century professionals tended to work in the same organization for long times, performing the same activities and knowing beforehand each stage of career development, it is now harder and harder to find pre-established scripts.
Because of the unpredictable professional path, career future becomes an open project that is source of concerns and object of consistent monitoring.In fact, the question "What will my future be in this organization?"becomes a key question and, this way, perspectives of career development/growth are among the most appraised characteristics by workers in organizations (Tolfo, 2002).
If expectations and perceptions of future in career play relevant role in the link between professionals and their job and organization, it is worth investigating which elements influence this connection.Amazingly, investigations adopting expectations and perceptions of future in career as dependent variable are relatively scarce, and antecedents remain elusive.
On this basis, this article aimed to evaluate a model of antecedents of career expectations in the organization.This construct refers to the individuals' beliefs about their future professional path in the current organization (Vasconcellos, 2015), comprising expectations about professional objectives, remuneration, prestige, performing interesting tasks in line with personal preferences.When approaching expectations, i.e., beliefs about potential future (Oettingen & Mayer, 2002), this investigation differentiates from other studies about careers that use present-oriented variables (like satisfaction with the career) or that focus on the desired future (career aspirations/objectives) which are quite common in Brazilian qualitative surveys (Graf & Coutinho, 2010;Soares & Sestren, 2007).
The model proposed -which is detailed below -includes three antecedent variables of career expectations in the organization and, as far as the review has gone, no similar model could be found in literature.Even for one-by-one analysis of the link between each duo of variables, there are few empirical studies.When testing this model, the authors intended to deepen investigations about career expectations and, in a broader light, comprise the temporal dimension of studies about work/organizations which is typically neglected or improperly approached in this area (Sonnentag, 2012).

THEORETICAL BASES AND PROPOSED MODEL
In psychology, there is a long history of studies on past experiences, dispositional characteristics, and observation of the other's performance as antecedents of expectations (Bandura, 1977;Monga & Rao, 2006;Scheier & Carver, 1985).By specifically discussing the career expectation, the model proposed herein (Figure 1) moves towards a different direction.Firstly, by including perceptions of human resources (HR) policies and practices and well-being at work as antecedents, the model tests the role played by perceptual and affective components of the professional's current situation in the building of career expectations.Moreover, by comprising organizational future expectations, the model verifies the effect of a future-oriented variable, which is little usual in literature.The following subsections detail the development of each research hypothesis and their theoretical bases.

PERCEPTIONS OF HR POLICIES AND PRACTICES AND WELL-BEING AT WORK
The HR management plays an increasingly strategic and relevant role to reach the organizational objectives and acquire competitive advantage (Bohlander & Snell, 2010;Jackson, Schuler, & Werner, 2012).This management is made up by policies -guidelines/proposals that guide the achievement of results intended by the organization -and practices, i.e., action paths that make management visible and operational (Demo, Neiva, Nunes, & Rozzett, 2012).Among other possibilities, HR policies and practices comprise recruitment and selection, training/development, working conditions, performance and competences evaluation, remuneration/compensation and the involvement of professionals with the organization (Demo, Neiva, Nunes, & Rozzett, 2014).Review of recent literature (Meneses, Coelho Júnior, Ferreira, Paschoal, & Isidro-Filho, 2014) shows that practices and policies of training/development and remuneration are the topics most frequently surveyed in Brazil.
In addition to the role they play in the achievement of organizational objectives, HR policies and practices influence the links between professionals and their work and organization.This survey is focused on the role played by the positive perceptions of professionals about HR policies and practices, where professionals recognize HRM as proper/beneficial, comprising their needs/ interests.The model proposed investigates the effects of such perceptions on the well-being at work, then understood as the prevalence of positive feelings over negative ones, as well as the individuals' perception hey express and develop their potential and skills at work, advancing in the achievement of their goals (Paschoal & Tamayo, 2008).
This concept combines two elements typically evoked by studies on well-being: hedonism and eudemonia.The first one emphasizes pleasure as the essence of well-being (Ryan & Deci, 2001) while the second one highlights the individuals' capacity of expressing their potential and having personal achievements in a meaningful way (Waterman, 1993).Although different, these perspectives are supplementary and positively related one another.This way, combining these enriches the study on well-being in general light (Ryan & Deci, 2001) and at work (Paschoal & Tamayo, 2008).
One could assume that well-being at work is predicted by the professionals' perceptions about HR policies and practices because these largely affect the organizational environment and the links between individuals and their activities.This proposal is supported by Warr's (2013) theoretical model that presents that elements affected by the HR policies and practices, among which remuneration, competence development, working conditions and characteristics of the activities, act as environmental antecedents of well-being.
Previous surveys have recorded positive relation between HR management and well-being at work, but define this last variable as organizational commitment and satisfaction with work (Baptiste, 2008;Horta, Demo, & Roure, 2012).In a sample from the United States, Demo and Paschoal (2013) have recently validated the English version of the Well-Being at Work Scale (Paschoal & Tamayo, 2008) and supported its nomological validity, since it reported strong positive association between well-being and perceptions of HR practices and policies.Based on the aforementioned theoretical and empirical support, the first hypothesis defines that: • Hypothesis 1 -Positive perceptions on RH policies and practices are positively related with well-being at work.

PERCEPTIONS ABOUT HR POLICIES AND PRACTICES, CAREER EXPECTATIONS IN THE ORGANIZATION AND EXPECTATIONS OF THE ORGANIZATIONAL FUTURE
If HR policies and practices are acquiring increasingly strategic and competitive relevance, these are highly likely to influence the future of organizations.However, is this relationship -which is so vaunted in literature about management -perceived by professionals?This survey approaches this relation between the professionals' perceptions about HR policies and practices and their expectations of organizational future defined as individuals' beliefs about management and effectiveness of their organizations in the future (Vasconcellos, 2015).
Few studies approach the organizational future in the workers' perspectives.However, these studies point out this variable as a relevant aspect to the link individual-work-organization (Chiu, 2002;Harris & Mossholder, 1996;Hui & Lee, 2000).The only empirical survey found about HR management and organizational future in the workers' perspectives (Y.Chen, Hsu, & Yip, 2011) showed moderate and positive correlation between perceptions on HRM effectiveness and future performance of the organization in samples of managers.The elements stressed allow supposing positive impacts from the smooth running of HR policies and practices on the organizational future.This relation is reflected not only on objective indicators, but also on the professionals' subjectivity.This way, hypothesis 2 states that: • Hypothesis 2 -Positive perceptions on RH policies and practices are positively related with the expectations of organizational future.
Moreover, one can conjecture about the link between perceptions on HR policies and practices and their own career future.Here, individuals that feel they benefit from the HR policies and practices (i.e., consider that training sessions, working conditions, compensations and recognition received are positive) probably believe in a relieving future to their organizations.In this line, when perceiving HR management as positive, professionals would tend to build beliefs according to which the organization provides the required conditions for their professional development and growth, bringing about positive expectations regarding their careers.This conjecture is supported by the theory of career motivation (London, 1983), which recognizes the importance of the organization situational features (such as environment/working conditions, compensation practices and support to training and development) in the building of career expectations.
On the empirical side, there are evidences that the existence of some HRM practices (notably, training and development) is positively related with a more favorable view on career opportunities and possibilities of professional growth (Kraimer, Seibert, Wayne, Liden, & Bravo, 2011;Veloso, Dutra, Fischer, Pimentel, Silva, & Amorim, 2011).This survey performs a more comprehensive test, approaching several HR management aspects.Anyway, based on the aforementioned theoretical and empirical contributions, hypothesis 3 assumes that: • Hypothesis 3 -Positive perceptions on RH policies and practices are positively related with the career expectations in the organization.

WELL-BEING AT WORK AND CAREER EXPECTATIONS IN THE ORGANIZATION
Studies external to the organizational field pointed out that positive affects influence on future-oriented thoughts and behaviors (Aspinwall, 2005) and that these, in general, lead to expectations of favorable results (Monga & Rao, 2006).These authors argue that positive affects tend to activate memories of harmonious past results (i.e., positive as well) that then shape positive prospects of future.
In the organizational behavior literature, the role played by affects in the shaping of behaviors and cognitions was object of theories and models in the last few decades.The affect infusion model (Forgas & George, 2001) suggests that affects are infused and "color" the formulation of attitudes, evaluations and behaviors, notably in open situations (i.e., with no standardized response).As the building of career expectations is translated into an open cognitive process, one can assume that positive (negative) affects at work influence favorable (unfavorable) expectations about professional future.
In turn, the broaden-and-build theory (Fredrickson, 2003) predicts that the presence of positive feelings broadens the individuals' cognitive repertoire and, in the long-term, develops personal resources that favor fitting to work and organizations.Individuals with heavy loads of positive affects tend to be more integrated in social terms, creative and motivated in tasks.These characteristics suggest higher resilience, capacity and confidence to handle with eventual adverse conditions in the future (Aspinwall, 2005;Fredrickson, 2003).This way, the prevalence of positive feelings over negative ones (well-being) at work can lead individuals to perceive themselves as more skilled and fit to the work/organization, giving rise to more favorable expectations about their careers.
The empirical surveys show evidences that positive affects are positively correlated with wage expectations of individuals in initial career stages (O'Neill, Stanley, & O'Reilly, 2011).Here, the proposal is to expand the research by O'Neill et al. (2011) considering that this investigation goes beyond wage expectations, including professionals at different professional stages and assessing affective states (these authors grounded their questionnaires on affectivity traits).
The previous argument about association between well-being at work and career expectations is focused on affects, although the concept of well-being adopted includes professional achievement.In this regard, a recent exploratory research showed that professional achievement is strongly and positively related with the individuals' vision on their career future, which is a variable assessed by one single item where respondents informed to which extent their future in the current organization seemed to be favorable (Vasconcellos & Neiva, 2014).Supported by the points raised, the fourth hypothesis postulates that: • Hypothesis 4 -Well-being at work is positively related with career expectations in the organization.

EXPECTATIONS OF ORGANIZATIONAL FUTURE AND CAREER EXPECTATIONS IN THE ORGANIZATION
Studying the link between the organization's future and career's future is crucial in the current context where globalization, technological advances and need for innovation compel organizations to continuously change (Bohlander & Snell, 2010).By means of interviews, the participants' discourse in some surveys showed the interweaving between organizational and professional futures so that concerns, risks and opportunities to individuals emerged intertwined with the expected further configuration of the organization (Lips-Wiersma & Hall, 2007;Margolis & Hansen, 2003;Ullrich, Wieseke, & van Dick, 2005).
For example, Margolis and Hansen (2003), investigating an airline in merging process, report that professionals expected increase in business, change of the organizational culture and improvement of the organization's reputation, while fostering expectations of career development and increase of salaries and benefits.In this line, one can assume that professionals that see favorable perspectives to the organization -like positive changes in management and expansion of operations -will probably envisage more positive future for themselves because those expectations can unfold in benefits to them.
Exploratory studies like that by Margolis and Hansen (2003) suggest the connection between expectations of organizational future and career future.The authors emphasize how literature has gaps regarding the understanding about career future projects and recommend new surveys on this topic.In compliance with the recommendation, this study analyzes the link between expectations of organizational future and of career in the organization, now thought quantitative analyses.Hypothesis 5, grounded on the aforementioned studies, defines that: • Hypothesis 5 -Expectations of organizational future are positively related with the career expectations in the organization.
A review of the literature that grounded the aforementioned hypotheses shows indications that positive perceptions about HR policies and practices are related to well-being at work and expectation of organizational future that, in turn, work as antecedents of career expectation.These indications support the proposition of a mediation model (Figure 1).As the previous literature analysis also suggests direct effect of positive perceptions about HR policies and practices on the career expectations (London, 1983: Veloso et al., 2011), the study suggests that mediating variables explain this effect only partially.As we have not found in literature a model that analyzed these variables as a whole, there is no reason to suppose full mediation.Anyway, the possibility of full mediation was analyzed in the results section.
Gender, age, organizational tenure and position type (managerial or not) were contemplated as control variables; however, previous surveys show that only the last variable is consistently related with expectations of career and of organizational future (Chay & Aryee, 1999; G. Chen et al., 2011;Hui & Lee, 2000;Stroh & Reilly, 1997;Veloso et al., 2011).This way, only position type was included as control to prevent the elaboration of an unnecessarily complex model.Although being not used as control, the variable organizational tenure was used in a supplementary analysis.Previous studies showed that organizational tenure moderates the relation of variables relevant to organizational behavior (Ng & Feldman, 2011;Wright & Bonett, 2002), including of attitudes at work and career expectations (G.Chen et al., 2011).Then, the study verified if the model preserved the same characteristics between professionals with different organizational tenure.
Based on the hypotheses proposed by the model, this survey intends to contribute with literature in different ways.To that, it has firstly evaluated the links between variables derived from previous theoretical constructions.Moreover, it analyzed if the results of previous studies, based on different samples (mostly foreign samples) and different methodological approaches, could be sustained in a new investigation.Finally, and above all, it coordinated the propositions/ findings of studies from different sectors of literature in a single and original model.This way the particular relations between variables can be better analyzed with more accuracy, since these are jointly tested, i.e., considering the reciprocal implications.

PARTICIPANTS
The initial sample comprised 332 respondents.After treating outliers and missing data the final sample was composed by 305 respondents.In that sample, mean age was 44.04 years (SD = 9.50), organizational tenure was 13.36 years (SD = 9.91), 74% were men and 30% held managerial positions.Education was equal to or higher than high school, and all the sample members worked to a government agency.The percentage of groups in the sample (including in layers by age and organizational tenure) were similar to those found all over the organization.To α = 0.05, this sample offers statistical power higher than 80% in all models reported, considering ε 0 ≤ 0.05 and ε a = 0.08, respectively, as null hypothesis and Room Mean Square Error of Approximation (RMSEA) value tested.

PROCEDURES
The respondents, randomly selected from a list with all professionals working at the organization, received an e-mail (sent by the HRM area of the organization) with information about the study, instructions, research contact and link of access to the questionnaire.It was completed individually, in a voluntary and confidential way.

Expectations of career in the organization and of organizational future
The study used the reduced version of the Career Expectations in the Organization Scale (Ceos) and the Organizational Future Expectations Scale (Ofes) validated by Vasconcellos (2015).In both scales, participants mark on 7-point scale the probability of the positive situation described in the item to come to happen in five years (1 = not probable to 7 = extremely probable).The Ceos statements refer to the individual's future career situation in that organization (example: "I will achieve my career objectives defined to this period") and those of the Ofes refer to the organization's future situation (example: "The working environment will be improved by the organization management").The Ceos is composed by the factors Professional Achievements and Career-Personal Life Relation; however, as result of negotiation with the organization and of the need for reducing the questionnaire, this study employed only the Professional Achievements factor (six items and α = 0.90).Both Ofes factors were used: Management and Organizational Environment (four items and α = 0.92) and Organizational Effectiveness (five items and α = 0.92).

Positive perceptions about human resources policies and practices
Three factors of the short version of the Human Resources Policies and Practices Scale were used (Demo et al., 2014): Training, Development and Education (three items and α = 0.81), Remuneration and Compensations (four items and α = 0.74) and Working Conditions (five items and α = 0.77).The scale has six factors, but applying it integrally would increase the questionnaire size.These factors were selected because they reflect core and traditional aspects of HR management, and all items fit into the institutional context of the company.For illustration purposes, here is an example of an item of the Training, Development and ,ation factor: "The organization I work for fosters learning and knowledge production".Items were answered using agreement scale (1 = totally disagree up to 5 = totally agree).All items used are elaborated in a positive way.

Well-being at work
The three factors of the Well-Being at Work Scale (Paschoal & Tamayo, 2008) were applied: Positive Affects (nine items and α = 0.95); Negative Affects (12 items and α = 0.93) and Fulfillment (nine items and α = 0.94)."In the last six months, my work has made me glad" is an example of Positive Affects factor items.The items on affects were answered in intensity scale (1 = not a bit up to 5 = extremely) and the items of the Fulfillment factor (example: "At my work, I do what I really like to do") through the agreement scale (1 = fully disagree up to 5 = fully agree).

Demographic variables
The questionnaire will also include items on gender, age, organizational tenure and position type of the respondent (managerial or non-managerial).

DATA ANALYSIS
The investigation had two stages: confirmatory factor analysis (to test the measurement model) and the structural model evaluation (to test the hypotheses).In the models, the six items of the construct "career expectations in the organization" were counted as empirical variables.The remainder constructs used the mean of their factors.This aggregation strategy is common in literature because it enables the emergence of variations with interval, normal and reliable distribution (Hall, Snell, & Foust, 1999).Models were estimated in the AMOS 18 software by maximum likelihood.As multivariate normality has not been rigorously achieved (the Mardia's test critical ratio was equal to 11.57), 500 sub-samples were generated through bootstrapping to evaluate the stability of parameters (Byrne, 2010).The sign, magnitude and significance of coefficients were observed in the tests of hypotheses, as well as goodness-of-fit of models based on the following indicators: Goodness-of-Fit Index (GFI), Comparative Fit Index (CFI), Standardized Root Mean Residual (SRMR), and RMSEA.The model proposed was compared with alternative models, because this practice is a stricter test than the isolate analysis of its goodness-of-fit (Hair, Black, Babin, Anderson, & Tatham, 2010).

ANALYSIS
Descriptive statistics and correlations are shown in Table 1.Expectations about the career and the organizational future are positively related with the remainder constructs (except for the factor of Negative Affects of well-being at work).The confirmatory factor analysis results are presented in Table 2.The model proposed (four factors) showed good fit in all indicators.All correlations between constructs and factor loads were significant (p < 0.01).These last ranged between 0.55 and 0.87 and were virtually identical to the mean of factor loads of the bootstrapping sub-samples (the difference was lower than 0.01 for all cases).These results provide evidence of stability of loading.a Career expectations and expectations of organizational future as unique factor (because these constructs are oriented to the future), and the remainder constructs as individual factors.
b Career expectations and expectations of organizational future as unique factor and remainder constructs in the other factor.c Expectations of organizational future and positive perceptions of HR policies and practices as a unique factor (because these are constructs oriented to the organization), and remainder constructs in the other factor.** p < 0.01.
Source: Elaborated by the authors.
The reliability was measured through the composite reliability (also known as Jöreskog rho) as proposed by Chin (1998).All reliability measures and mean variances extracted from factors exceeded the reference values (0.75 and 0.50, respectively) indicated by Hair et al. (2010).Therefore, the model proposed also showed evidences of convergent validity.The discriminant validity was based on the comparison of alternative models with logical sense in the confirmatory factor analysis (Table 2).This practice is very usual in research focused on testing the relation between variables rather than on the validation of instruments (Strobel, Tumasjan, Spörrle, & Welpe, 2013;Tierney & Farmer, 2002).In this comparison, the alternative models were statistically inferior (by the differences of χ 2 ) and expressed goodness-of-fit indicators worst than the proposed model.Since all variables were obtained in the same questionnaire, there was the risk of common method bias.Although the questionnaire has been designed to mitigate it ensuring anonymity of responses and adopting different introductions and scales to each variable (Podsakoff, MacKenzie, Lee, & Podsakoff, 2003), the study observed to which extent the risk persisted.The techniques to find this bias are limited/imprecise and there is no consensus in literature about which is the best one (Podsakoff et al., 2003;Richardson, Simmering, & Sturman, 2009).The single-factor test (also known as Harman's test) was selected to this survey (Woszczynski & Whitman, 2004).It is a well-known technique where an exploratory factor analysis with no shifting for all items of the model is performed.If the first factor explains more than 50% of the variance, there are indications of bias (Podsakoff et al., 2003), but it did not happen in the proposed model.A variation of this technique became very popular in literature recently and uses the data shown in Table 2: a unique factor model is generated in the confirmatory factor analysis, and goodness-of-fit is evaluated (Podsakoff et al., 2003).As the unique factor model resulted in poor goodness-of-fit, it is unlikely to have high risk of common method bias.

STRUCTURAL MODEL AND TEST OF HYPOTHESIS
The structural model proposed recorded good fit: χ  Hypotheses 1 and 2 predict that positive perceptions of HR policies and practices were positively related with well-being at work and expectations of organizational future.Both were supported.On the other hand, hypothesis 3 (positive perceptions of HR policies and practices are positively related with expectations of career in the organization) was not supported.The effect was weak and insignificant.In this sense, data suggest that the relation between perceptions of HR policies and practices and expectations of career in the organization found in the measurement model (r = 0.77; p < 0.01) was fully mediated by well-being at work and expectations of organizational future.
Hypotheses 4 and 5 indicated, respectively, that well-being at work and expectations of organizational future are positively related with career expectations in the organization.The hypotheses were supported with moderate magnitude effect in the first case and strong in the second one.The control variable of position type supported weak (lower than 0.10) and insignificant (p > 0.05) effects in well-being at work and variables of expectations.
As expected, the proposed model was compared with alternative models.Based on the non-acceptance of hypothesis 3 in the alternative model 1 (A1), the direct relation of positive perceptions of HR policies and practices in career expectations in the organization was suppressed.The difference between models, Δχ 2 (1) = 1.56, was insignificant (p = 0,21).The goodness-of-fit indicators for model A1 (GFI = 0.91, CFI = 0.94, RMSEA = 0.08 and SRMR = 0.05) were virtually identical to those of the proposed model.Therefore, the exclusion of this relation did not worsen the model goodness-of-fit, which is a new evidence of full mediation between these variables.In model A1, the control variable position type replicated the same standard of relations found in the proposed model.
The alternative model 2 (A2) added to the proposal model the direct effect of well-being at work on expectations of organizational future.Although this hypothesis could not be supported by literature, there was a chance that well-being at work could influence the expectations of organizational future.The difference between models, Δχ 2 (1) = 0.01, p = 0.92, was not significant.The remainder goodness-of-fit indicators of model A2 were virtually identical to those of the proposed model.Therefore, the addition of this new connection did not improve the goodness-of-fit of the model and made it unnecessarily more complex.
Comparing the models, A1 was adopted as the final model for the study.This has the same levels of goodness-of-fit as the proposed model, but with two advantages: greater parsimony and statistical significance in all structural coefficients between constructs.Model A1 is shown in Figure 3, corroborating the support of hypotheses 1, 2, 4 and 5 of the proposed model.The indirect standardized effect of positive perceptions of HR policies and practices on career expectations in the organization reached 0.72 (0.46 via expectations of organizational future and 0.26 via well-being at work).

SUPPLEMENTARY ANALYSIS: DOES ORGANIZATIONAL TENURE CHANGE THE MODEL?
As supplementary analysis the study tried to evaluate if model A1 kept the same characteristics despite the professionals' organizational tenure.Inspired by the segmentation adopted by Ng and Feldman (2011) two groups have been assembled: 1. professionals with short/medium organizational tenure (up to 10 years) here named as non-veterans (45% of the sample) and 2. professionals with longer organizational tenure (more than 10 years) here named as veterans (55% of the sample).
A moderation test was performed using multi-group analysis, which is a technique that demands invariance of configuration and metric between groups in the measurement model (Hair et al., 2010).The factor structures were equal to both groups (with all factor loadings reaching p < 0.01) and goodness-of-fit indicators with no restriction were acceptable: χ 2 (166) = 358.73,p < 0.01, GFI = = 0.87, CFI = 0.93, RMSEA = 0.06 and SRMR = 0.06, suggesting invariance of configuration.It is worth mentioning that the sample division into multi-group analysis tends to reduce the ratios of degree of freedom/sample size, which is a condition that naturally reduces the GFI values (Hooper, Coughlan, & Mullen, 2008).In any manner, metric invariance was also reached, since there was no significant difference between the model without restrictions and the model that Based on the model of metric invariance, the structural model without restrictions (where structural coefficients to both groups were freely estimated) was compared pair-to-pair (via Δχ 2 ) with four restricted structural models with the imposition of equality to one of the structural coefficients of model A1.Two coefficients indicated moderation.In the first case, the model that imposed equality in the connection between expectations of organizational future and career expectations showed significant difference against the model without this restriction: Δχ 2 (1) = 7.51, p < 0.01.When freely estimated between groups, the effect (standardized coefficient) of expectations of organizational future on career expectations was strong between non-veterans (0.71; p < 0.01) and moderate to strong between veterans (0.47; p < 0.01).In the second case, the difference between the model that imposed equality in the connection of well-being at work and career expectations and the model without this restriction was near of the statistical significance: Δχ 2 (1) = 4.77, p = 0.03.In the model without this restriction, the standardized effect of well-being at work on the career expectations was weak to moderate among non-veterans (0.23; p = 0.01) and moderate to strong among veterans (0.48; p < 0.01).

DISCUSSION
Results have shown the positive effects of positive perceptions about HR policies and practices, well-being at work and expectations of organizational future on the career expectations in the organization, besides exhibiting the great capacity of the model to explain this last variable.However, in opposition to expectations, the positive perceptions of HR policies and practices had mainly indirect effect on career expectations.
Well-being at work showed moderate and positive effect on career expectations.Probably that is so because individuals with higher positive affects in the sample are more optimistic in the cognitive processing that generates expectations (Forgas & George, 2001) and/or because individuals with good levels of well-being (including professional fulfillment) feel to be more skilled/confident to deal with future challenges (Fredrickson, 2003).In this last case, the favorable working situation in the present can be projected (extended) to the future.The expectations of organizational future, in turn, showed strong and positive effect on the career expectations exceeding in magnitude the effect of well-being at work.This way, when dealing with career in the future, professionals assign high weight to the future organizational context.
The comparison of models of professionals at different stages of the career in the organization showed that the effect of well-being at work on the career expectations is bigger among veterans.Because of their longer organizational tenure, veterans probably envisage their future as a narrower path with less possibilities of change than the non-veteran professionals.Therefore, they are more likely to believe in future in career as an extension of the current situation of well-being (or discomfort), which could explain the huger effect of the last variable on the career expectations.The non-veteran professionals, in turn, probably assume there are more chances of change, weakening this relation.
The comparison between models also showed that the effect of expectations of organizational future on career expectations is greater among non-veterans.That is so probably because most of non-veterans have not yet consolidated their niche of work in the organization and, this way, their future in career would be more strongly dependent on growth or organizational change.Veterans would be little less dependent on these elements because they have already won place in the organization.
This research contributes with the literature firstly because it tests the relations proposed or identified by previous studies.In this line, data supported the relation between positive perceptions of HR policies and practices with well-being at work (Demo & Paschoal, 2013) and with expectations of organizational future (Y.Chen et al., 2011).In the first case, results have empirically supported Warr's theoretical model (2013) according to which well-being at work is conditioned by environmental antecedents related to HR policies and practices.Likewise, results support the relation theoretically conceived between career expectations in the organization and well-being at work (Forgas & George, 2001;Fredrickson, 2003) and between career expectations in the organization and of organizational future highlighted in previous qualitative surveys (Margolis & Hansen, 2003;Ullrich et al., 2005).In this last case, the quantitative model provided additional contribution because it allowed measuring the magnitude of this relation.
Secondly, above corroborating previous surveys, this study contributes by showing the dynamic of coordination between antecedent variables.This is only possible when variables are jointly tested in the model.This initiative becomes even more relevant in face of the shortage of investigations adopting expectations of career in the organization as dependent variable.
In practical terms, data shed light on the HR management power to influence -both directly and indirectly -variables relevant to the organizational life.Improved HR policies and practices seem to increase well-being at work and expectations of organizational future, and this effect unfolds in more favorable career expectations.As in the organization investigated the means of factors of well-being at work (referring here to positive affects and fulfillment) and of expectations of organizational future are not high (Table 1), there is a margin to rise these variables working on the HR management.
It is worth observing the insertion of expectations of organizational future in the model.On one hand, expectations of organizational future strongly predict career expectations in the organization, showing the need for special care by managers when communicating the strategic planning and organizational perspectives.On the other hand, the expectations of organizational future were strongly and positively influenced by the perceptions about the HRM work suggesting this is a key point to the future of the institution.It must be observed that the organization surveyed is a government agency and, as such, will continue existing despite the economic crises that deeply affect the private initiative.In this environment relatively protected against external market conditions, HRM practices that ensure good working conditions and professional qualification are appraised as crucial to improve the organization in the future.
Among the research limitations, its cross-sectional nature, impossibility of establishing causal relations between variables, and the fact that the model was tested in a government agency (institutional environment different from private corporations, for example) are worth of notice.Other limitations concern the use of non-probabilistic sample (questionnaire was sent to a random list of professionals, but was answered only by those who wanted) and the exclusive use of self--report measurements, potentially influenced by biases like social desirability.It is also worth mentioning that variables not comprised in the model can play significant role to explain the expectations of career, among which traits of personality, performance and engagement at work and commitment with the career.Besides analyzing the effect of these variables, future surveys should investigate the model in other organizations (including the private initiative) to find out if results found in this study are replicated.
The question "What will my future be in this organization?" is part of the everyday life of professionals.This survey tried to identify/understand the factors that trigger favorable or unfavorable responses to that question.Results pointed out the relevance of both present-oriented situational variables and future--oriented variables.However, there is a lot of room in the organizational scenario to be explored regarding how variables of different time periods (past, present and future) interact and influence the subjectivity of workers.New investigations should help explaining the intertemporal dynamic of the organizational life.
figure 1STRUCTURAL MODEL PROPOSED Note: CEO = Career Expectations in the Organization; EOF = Expectations of Organizational Future; PPHRP = Positive Perceptions of Human Resources Policies and Practices; TDE = Training, Development and Education; WBW = Well-Being at Work.These were the options of response: ECO and EOF (1 = not at all likely up to 7 = extremely likely); PPHRPP and WBW Fulfillment (1 = totally disagree up to 5 = totally agree); and, WBW Positive/Negative Affects (1 = not a bit up to 5 = extremely).Correlations obtained from the mean of items of each sub-scale.Correlations equal to or higher than 0.15 (in module) are significant to α = 0.01; between 0.11 and 0.14 are significant considering α = 0.05.a Position type: 0 = non-managerial and 1 = managerial.Source: Elaborated by the authors.
figure 2STANDARDIZED COEFFICIENTS OF THE PROPOSED STRUCTURAL MODEL figure 3STANDARDIZED COEFFICIENTS OF THE ALTERNATIVE MODEL 1 (A1)

table 2 INDICATORS OF GOODNESS-OF-FIT OF MODELS IN THE CONFIRMATORY FACTOR ANALYSIS
Note: CI = Confidence interval 90%.