The Implications of Direct Participation for Organisational Commitment, Job Satisfaction and Affective Psychological Well‐Being: A Longitudinal Analysis

The article examines the implications of direct participation for employees' organisational commitment, job satisfaction and affective psychological well&#8208;being. It focuses on both task discretion and organisational participation. Applying fixed effect models to nationally representative longitudinal data, the study provides a more rigorous assessment of the conflicting claims for the effects of participation that have hitherto been based primarily on cross&#8208;sectional evidence. Further, it tests a range of mechanisms by which direct participation leads to improved employee outcomes. Contrary to the critical literature, it shows that even after controlling for unobserved individual heterogeneity, both forms of direct participation have positive effects for employees' organisational commitment and well&#8208;being. The effects of task discretion are primarily direct, reflecting the intrinsic importance of personal control over the job task; in contrast, those of organisational participation derive to a greater extent from its indirect effect on the quality of working conditions.


INTRODUCTION
In its broad sense, direct participation refers to the capacity of employees to personally influence decision-making at work whether at the level of the work task or the wider organisation. The importance of participation has been highlighted both by researchers concerned with the quality of work (Boxall and Macky, 2014;Felstead et al., 2016;Gallie, 2013;Heller et al., 1998;Macky and Boxall, 2008;Wood and de Menenes, 2011;Zhou, 2009) and by those interested in performance and productivity (Appelbaum et al., 2000;Lawler et al., 1995;Wood et al., 2012). The belief that participation can have positive effects for both employees' well-being and their motivation has also given it an important place in the policy priorities of the European Union, most notably through Directive 2002/14EC on the Information and Consultation of Employees (Budd and Zagelmeyer, 2010). While the UK initially showed limited enthusiasm to implement such measures, the Information and Consultation of Employees regulations were adopted in 2004 and progressively extended to different sectors of industry in the period 2005 to 2008. Moreover, the importance of improving mechanisms for employee participation was reaffirmed in the Green Paper on Corporate Governance Reform (Department of Business, Energy and Industrial Strategy, 2016).
Yet the evidence for the effects of direct participation has been a matter of considerable debate, in particular with respect to the implications for employee well-being. While some researchers have concluded that direct participation has positive effects for both organisational commitment and employee well-being, others have argued that it promotes organisational commitment at the cost of well-being (Harley, 1999). Over recent decades, evidence has been extended from case studies to representative surveys of the wider workforce. But the causal conclusions of even wider representative studies have remained necessarily tentative given the crosssectional nature of the evidence on which they have been predominantly based. Further, there have been few attempts to assess the relative importance of mechanisms linking direct participation to employees' work attitudes and wellbeing.
The present study draws on new longitudinal data from the British Skills and Employment Surveys to assess conflicting claims about the effects of direct participation advanced by previous studies. At the same time, it seeks to examine the potential mechanisms that might account for the effects of participation, considering how far they are attributable to the intrinsic importance of selfdetermination or to the benefits of participation for improvements to other aspects of the work situation-in particular the opportunities for high-quality training, the ability for employees' to make full use of their skills and the level of job security.

POSITIVE AND CRITICAL PERSPECTIVES ON DIRECT PARTICIPATION
There has been a long tradition of research on the quality of work that has argued for the importance of greater participation in decisions about work tasks for employee wellbeing (Hackman and Lawler, 1971;Hackman and Oldham, 1980;Warr, 2007). It has emphasised its benefits either for job satisfaction (an indicator of evaluative well-being), or for affective psychological well-being (an indicator of hedonic well-being). While to some degree related concepts, these have been shown to capture distinctive aspects of well-being (for a discussion, see Bryson et al., 2014). From the 1980s, such arguments were extended by psychologists to the importance of task control for psychological health, in part due to its role in reducing the effects of high levels of work pressure (Johnson and Johansson, 1991;Karasek and Theorell, 1990;Theorell, 2007).
From the 1980s, there was a growing interest in the potential benefits of direct participation for employee motivation. An influential new strand of management theory argued that in an advanced economy, improvements in performance required a shift in the dominant paradigm 'from control to commitment' and that this could only be achieved through greater decentralisation of decision-making to the workforce (Lawler, 1986;Walton, 1985aWalton, , 1985b. While also highlighting the importance of task discretion, its advocates adopted a broader conception of participation, emphasising in addition the importance of wider organisational voice. This view inspired extensive research on the determinants and consequences of 'organisational commitment', in particular affective commitment involving the individual's emotional attachment to and identification with their organisation Allen, 1991, 1997;Meyer et al., 1990). The degree of decision-making autonomy was found in a number of studies to be an important predictor of such commitment (Boyd et al., 2011;Meyer and Allen, 1997). Similar ideas also came to inform a growing literature advocating 'high involvement' or 'high performance' management practices to encourage employees to use more of their skills and initiative (Applebaum et al., 2000;Felstead et al., 2010;Inanc et al., 2015;Lawler et al., 1995). Some researchers have proposed a possible synergy between participatory processes and other aspects of organisational design, in particular an emphasis on skill formation and individual performance incentives. But empirical analysis of the associations of these different dimensions of organisational structure indicates that the linkages are loose and interaction effects difficult to discern (Wood et al., 2015). In our succeeding analysis, we focus on the specific effects of the different types of direct participation, rather than their effects as part of a wider bundle of workplace characteristics.
There has also been, however, a critical literature of the assumptions underlying theories of participation. This has pointed to the potential for management to use participatory involvement to heighten its control of employees, by undermining traditional forms of union representation and inculcating a vision of the employment relationship favourable to management objectives (Barker, 1993;Legge, 1995;Purcell, 1993;Ramsay, 1977;Ramsay et al., 2000;Willmott, 1993). Participation, it is argued, exposes employees more systematically to management's definition of values, leading them to internalise the priority of performance objectives with respect to quality, flexibility and productivity. Consequently, while participative practices may be successful in increasing employees' commitment to the organisation, they may not necessarily improve the quality of job tasks. Employers may be able to take advantage of employees' acceptance of managerial priorities to increase work intensity resulting in lower employee well-being. While scepticism about participatory practices has focused on both task discretion and organisational participation, its arguments would appear to be particularly germane to organisational participation where employees are most likely to be exposed to managerial arguments about the pressures facing the organisation and the need to give priority to performance objectives.
Overall, then, despite substantial empirical research, existing evidence for the effects of participation on employee well-being remains problematic. In particular, there have been conflicting results on the issue of whether it could have negative effects on well-being as a result of work intensification. Some studies appear to confirm that high involvement work practices have negative effects on well-being, while others have found no relationship with measures of work intensity and overall positive effects for employee well-being (Boxall and Macky, 2014;Macky and Boxall, 2008;White et al., 2003).

Direct, mediating, or artefactual effects?
A problem highlighted by such conflicting views of the effects of participation is that there has been a scarcity of research on the mechanisms that link participation to different outcomes-whether with respect to organisational commitment or individual well-being. In principle, three broad possibilities can be distinguished. The first relates to the intrinsic effects of participation, whereby involvement in decision-making leads directly to changes in organisational commitment, job satisfaction and affective psychological well-being. In the second, the implications of participation derive from the fact that it influences other features of the work context, which in turn affect these employee outcomes. In the third, the relationship between direct participation and employee attitudes is not a causal one but reflects unobserved prior differences between individuals.

Direct effects
One version of the argument that personal control is intrinsically important draws on theories of basic needs. For instance, in 'Self-Determination Theory' autonomy is one of three basic psychological needs-the others being competence and relatedness (Ryan and Deci, 2000). As work is one of the key life arenas in which such needs are expressed, managerial support for autonomy is held to relate directly both to employees' motivation and psychological adjustment (Baard et al., 2004;Deci et al., 1989). Such theories make demanding assumptions about the universality of needs. An argument allowing for greater individual variation views the importance of self-determination as an outcome of processes of value formation. For instance, Argyris (1964) has suggested that needs for self-determination and growth are partly a consequence of socialisation within the educational systems of advanced societies.

Mediated effects
An alternative view is that higher employee participation may affect employee attitudes primarily through its effects on other aspects of the work situation. The critical perspective on participation has paid particular attention to its negative implications for work intensification. But there is also some evidence that employees' capacity to influence decisions may have positive effects with respect to a number of other factors that have been shown to be important for well-being-in particular opportunities for learning through better training provision, greater scope for the use of skill and greater job security (Boxall et al., 2015;Heyes et al., 2017;Leach et al., 2003;Morrison et al., 2005).
By giving employees greater involvement in decision-making, participation may improve not only the availability but also the quality of training, because the needs of the workforce are more likely to be taken into account. It is likely that wider organisational voice will be particularly important for training provision, because it will require a decision about resource allocation that is usually decided at a relatively high organisational level. Direct participation is also likely to increase employees' ability to make full use of their specific skills and knowledge, because they will have greater choice about how the job is to be done (Morrison et al., 2005). Skill utilisation in turn has been shown to be a significant factor affecting higher levels of satisfaction at work (Kornhauser, 1965;O 0 Brien, 1982;Warr, 2007).
The effects of direct participation may also be mediated by its implications for perceived job insecurity. There is now substantial evidence that job insecurity can have strong negative effects for psychological well-being (Burchell, 2011;Cheng and Chan, 2008;Keim et al., 2014;Sverke et al., 2006). Some studies have indicated that the psychological distress resulting from job insecurity may be reduced where employees can participate in decision-making (Bussing, 1999;Gallie et al., 2017;Probst, 2005). This may be because employees feel that they will have greater influence over the eventual outcome, or, in terms of the organisational justice theory (Greenberg, 1987(Greenberg, , 1990, because participation increases procedural legitimacy and people will be more ready to accept personally disadvantageous decisions if they regard procedures as fair.

Artefactual effects
Finally, both of the approaches discussed previously-whether emphasising direct or mediated mechanisms of direct participation-assume that there is a genuine causal relationship. However, a third possibility is that the relationship is artefactual, reflecting unobserved variations in individual dispositions that influence both the propensity of employees to participate and the likelihood that they will be committed to their organisations and have specific levels of well-being. This places the emphasis on prior differences in the values, motivational disposition and personality characteristics of employees. This possibility has been difficult to address in the greater part of previous empirical studies because of their cross-sectional nature. Relatively few studies have used longitudinal data that is representative of the wider workforce, thereby allowing more rigorous testing of causal assumptions. Where a longitudinal design has been adopted, the focus has been primarily on the implications of task discretion for psychological well-being. These have indeed underlined the importance of employees' everyday influence over the work task for the risks of psychological distress, both independently and in combination with high levels of work pressure (Bentley et al., 2015;Stansfeld, 2006;Theorell et al., 2015). But, to our knowledge, there are no longitudinal studies of a representative type that have examined both task and organisational level participation to assess their relative importance or distinctive effects.

Hypotheses
In the light of the literature, we focus on three general hypotheses about the implications of participation: Hypothesis 1a: Higher direct participation in terms of both task discretion and organisational participation leads to higher organisational commitment, job satisfaction and affective psychological well-being.
Hypothesis 1b: Organisational participation increases organisational commitment but is detrimental to job satisfaction and affective psychological well-being.
Hypothesis 1c: The apparent effects of participation can be accounted for by unobserved prior individual differences that affect both the propensity to participate and organisational commitment, job satisfaction and affective psychological well-being.
From the discussion with respect to mediating processes, there are also a number of contrasting hypotheses that can be drawn: Hypothesis 2a: Participation has positive effects for well-being because of its intrinsic importance in meeting employee needs or values.
Hypothesis 2b: Participation has positive effects for well-being because of its implications for other aspects of organisational context such as training provision, opportunities for the use of skills and lower job insecurity.
Hypothesis 2c: Participation has negative effects on well-being because of its role in increasing work intensity.
The remainder of the article is structured as follows. In the next section, we present the survey data upon which the analysis is based, the indicators of participation and the different measures of work attitudes and well-being, the potential mediating variables and the modelling strategy. In the third section, we examine the relationship between participation and the hypothesised mediating variables. We then turn in the fourth section to consider both the gross effects of participation on organisational commitment, job satisfaction and affective psychological well-being and the effects once account has been taken of the mediating variables. In the final section, we discuss the implications of the analyses for our initial hypotheses.

The survey data
In seeking to address these issues, we draw on nationally representative longitudinal data of employees in Britain, which are part of the British Skills and Employment Survey series. The initial sample, interviewed in 2012, consisted of 3,200 people aged 20 to 65 in paid employment, with a response rate of 49 per cent (see appendix in Felstead et al., 2015). In 2014, a random subsample of 1,108 of the original respondents were reinterviewed, with a response rate of 71 per cent (GFK, 2014). Interviews at both waves were conducted face-to-face, but the interview length for the second follow-up survey was substantially shorter (30 minutes compared with approximately one hour).

Indicators
3.2.1 Direct participation: task discretion and organisational participation The measure of individual task discretion was derived from four questions designed to assess how much personal influence people had over specific aspects of their work. They asked, 'How much influence do you personally have on … how hard you work; deciding what tasks you are to do; deciding how you are to do the task; deciding the quality standards to which you work?' There was a four-point response scale: 'a great deal (of influence)', 'a fair amount', 'not much' or 'none at all'. Response points were scored from 0 for 'none at all' to 3 for 'a great deal'. As the four items had a Cronbach's alpha of 0.77, a summary index was created by averaging across the four items.
Our measure of organisational participation is constructed from a combination of two sets of questions-the first captures the scope (or range) of decisions over which organisational participation takes places; the second focuses on the degree of influence that employees can exercise through channels for organisational participation.
To assess the scope of organisational participation, respondents were asked initially, 'At your workplace, does management hold meetings in which you can express your views about what is happening in the organization?' For those who could express their views, additional questions were asked about six issues on which consultation took place. The objective was to construct a measure that reflected the relative strategic importance of the issues on which employees could express their views. Issues were grouped into those relating to more immediate work activity (working practices, health and safety and training), those concerned with decisions about products (planned changes in products or services) and those involving longer-term financial issues (the financial position of the organisation and investment plans). A four-point scale of organisation participation was constructed from these categories running from 0 for no participation, 1 for participation limited to work activity issues, 2 for participation limited to work activity and product decisions and 3 for participation that included strategic issues (investment plans and financial situation). Hence, the higher the score, the more involvement employees have in strategic decisions. The hierarchical nature of the measure is reflected in the fact that the average number of issue areas on which employees could express their views increased across the successive levels. We take the two highest categories (decisions about products and longer-term financial issues) as representing high levels of scope.
A second measure sought to capture the influence of organisational participation by asking, 'Suppose there was going to be some decision made at your place of work that changed the way you did your job. Do you think that you personally would have any say in the decision about the change or not?' Those who reported that they would have a say were asked whether that influence would be a great deal, quite a lot or just a little. Those who thought they could exercise a great deal or quite a lot of influence were categorised as having a high degree of influence. 1 A dummy variable of 'high organisational participation' was constructed by combining the two measures of organisational influence. The variable takes the value of one if individuals reported both high levels of scope (able to influence decisions about products and longer-term financial issues) and high levels of influence (able to exercise a great deal or quite a lot of influence over organisational decisions that affect their work).

Organisational commitment
The measure of organisational commitment, reflecting identification and attachment to the current employer, is drawn from scales developed by Mowday et al., (1979). The index of organisational commitment is composed of seven items (Cronbach's alpha = 0.85), asking people how much they agreed or disagreed (on a four-point scale) with the following statements about their feeling towards the organisation they worked for: • I am willing to work harder than I have to in order to help this organisation to succeed. • I feel little loyalty to this organisation (scores reversed).
• I find that my values and the organisation's values are very similar.
• This organisation really inspires the very best in me in the way of job performance. • I am proud to be working for this organisation.
• I would take almost any job to keep working for this organisation.
• I would turn down another job with more pay in order to stay with this organisation.

Employee well-being
Given the distinction in the literature between evaluative and hedonic well-being, we have taken two measures-an indicator of job satisfaction, which relates to people's general evaluation of their job in the light of their reference groups and past experience, and an indicator of affective psychological well-being, which captures their more immediate emotional response to their work. Job satisfaction is measured through a single question: 'All in all, how satisfied are you with your job?', with a seven-point response scale from completely satisfied to completely dissatisfied (scores were reversed).
The measure of affective psychological well-being is drawn from the enthusiasmdepression subscale proposed by Warr (1990). People were initially asked 'Thinking of the past few weeks, how much of the time has your job made you feel each of the following...?' There followed a series of adjectives, some positive some negative. To tap enthusiasm-depression, the adjectives were 'depressed', 'gloomy', 'miserable', 'cheerful', 'enthusiastic' and 'optimistic'. Responses were made against a standard five-point frequency scale ranging from 'never' to 'all of the time'. A reliability analysis gave a Cronbach alpha of 0.87 for the six items. A scale was constructed by averaging responses to the items, with the negative items reversed. Higher scores and therefore positive coefficients indicate better psychological well-being.

Potential mediating variables
We focus on four potential mediating factors: the quality of training provision, opportunities for skill use, work intensity and job insecurity. The measures of these were constructed as follows: 3.2.4.1 Training quality. To assess training quality, employees were asked whether they had received any training that helped them to improve the way they did their job, improved their skills, led to a qualification or a credit towards a qualification, and provided them with skills that were useful for another employer either in the same industry or a different industry. As the five items had a scale alpha of 0.85, we have standardised the individual items and averaged them into an additive scale. Those who had not received any training at all were assigned a score of zero.
3.2.4.2 Opportunities for skill use. Opportunities for skill use were measured with an item asking how much the person agreed or disagreed that 'In my current job I have enough opportunity to use the knowledge and skills that I have'. Scores ran from 1 'strongly disagree' to 4 'strongly agree'.

Work intensity.
Work intensity was measured by five items asking people to assess the extent to which their job required them to work very hard and under a great deal of tension and also how often their work involved working at very high speed and to tight deadlines and having to put in extra time over and above the formal hours of their job. A factor analysis showed one underlying dimension, and the items have a Cronbach's alpha of 0.68.

Job insecurity.
Job insecurity was assessed through two questions. The first asked: 'Do you think there is any chance at all of you losing your job and becoming unemployed in the next twelve months?' Those who thought they might lose their job were then asked about the likelihood of this happening, with response options of very likely, quite likely, evens, quite unlikely and very unlikely.

Modelling strategy
Given that one of our central aims was to assess the possibility that the relationship between direct participation and employee outcomes was due to unobserved individual heterogeneity, we use fixed effects models to control for time-invariant personal fixed effects. The main advantage of applying fixed effect modelling to longitudinal data is that it allows us to control for individual characteristics that may confound the effect of participation on work attitudes and well-being. For instance, if unobserved individual characteristics (such as personality dispositions) influence both one's job satisfaction and propensity to engage in participatory practices, the estimated effect of participation on job satisfaction based on crosssectional data can be spurious. Fixed effect models focus on within-individual patterns of change, which effectively filter out the influences of time-invariant individual-level factors that could influence both dependent and independent variables.
We start by examining the effects of the different forms of direct participation on the potential mediating variables to see whether there are grounds for thinking that they might at least partially channel the effects of participation. Then we turn to estimate the relationship between participation and the three dependent variables: organisational commitment, job satisfaction and affective psychological well-being. We look first at the gross effects of task discretion and organisational participation to see whether they are significant once account is taken of individual heterogeneity. Next, we examine the effects with time-varying individual and work status controls and finally consider the effects once the potential mediating variables have also been taken into account. If significant effects of direct participation on well-being persist even when mediating variables are introduced, we conclude that the effects of participation are at least partially because of its intrinsic importance for employees. If significant initial effects of participation disappear once the mediating variables have been introduced, we conclude that direct participation has an influence predominantly through its effects on other aspects of the work environment that are important for worker attitudes.

Effects of direct participation on training quality, opportunities for skill use, work intensification and job insecurity
The first stage of the analysis turns to the assessment of the relationship between participation and the potential mediating variables to establish whether the latter are indeed affected by participation in the way theoretically assumed. Table 1 presents the effects of participation on training quality, opportunities for skill use, work intensity and job insecurity.
Task discretion has a highly significant effect on opportunities for skill use at work, but no significant effects on training quality, work intensity or job insecurity. Organisational participation has a positive effect on both training quality and on opportunities for skill use, and a negative effect on job insecurity. The analysis then confirms that direct participation influences three of the potential mediating variables -training quality, skill use and job insecurity-although the different forms of participation have distinctive effects. But there is no evidence that either task discretion or organisational participation have a significant effect on work intensity. We therefore exclude work intensity from the set of mediating variables, although retaining it as a control variable in the event that it captures something distinctive about the type of organisation that may confound the relationship between participation and worker attitudes.

Effects of participation on organisational commitment and employee well-being
In the next stage of the analysis, we turn to examine the effects of participation on the outcome variables (organisational commitment, job satisfaction and affective psychological well-being) in a fixed effects model that takes account of the potential confounding influence of unobserved individual heterogeneity. This leads to a more differentiated picture of the effects of participation than appears from a random effects model, which does not take account rigorously of unobserved differences in individual dispositions. In a random effects model (Appendix Table A1) all dimensions of participation are significantly related to both organisational commitment and the two measures of well-being. While Table 2, which takes account of personal fixed effects, confirms that there are participation effects for each of the outcome variables, it can be seen that the coefficients are considerably reduced. The effect of task discretion is relatively small for organisational commitment compared with its effects on both aspects of well-being. Moreover the association between organisational participation and affective well-being is no longer significant. Table 3 shows the results of further analyses to examine first whether the previous gross effects of participation can be accounted for in terms of changes in the control variables-age, hours of work, work intensity and contract status-and second how far the effects are reduced by the introduction of the mediating variables. The first column for each dependent variable gives the patterns with controls, the second the results when the mediator variables also have been added.  Note: Fixed effects models. Samples restricted to respondents on all control variables and to employees who remained in the organisation. *** p < 0.01. ** p < 0.05. * p < 0.10. It can be seen that the effects of task discretion on job satisfaction and affective psychological well-being are only slightly reduced by the controls and remain statistically significant. However, the association between task discretion and organisational commitment is now no longer significant. Turning to organisational participation, the introduction of the controls again does not change substantially its gross effects on either organisational commitment or job satisfaction, while organisational participation is now significant for affective psychological wellbeing. Even taking account of controls, the effect of both indicators of participation, taken together, is quite substantial in the case of job satisfaction, and somewhat smaller though still significant in the case of psychological wellbeing and organisational commitment. Thus, a standard deviation increase in task discretion and organisational participation would lead to an increase of 0.23 of a standard deviation with respect to job satisfaction, of 0.13 of a standard deviation in affective psychological well-being and of 0.12 with respect to organisational commitment.
The final step was the introduction of the mediator variables. As expected, they had a strong effect on all of the measures-training quality and opportunities for skill use increased both organisational commitment and well-being, while job insecurity reduced them. Their introduction, however, leads to little change in the coefficients for task discretion and has no effect on the significance levels. This suggests that the effects of task discretion on employee well-being are largely direct and attributable to the intrinsic importance to employees of being able to take decisions about their immediate jobs.
The pattern for organisational participation is more complex. The effect of organisational influence on job satisfaction remains significant when the mediators are taken into account, although there is some decrease in the strength of the coefficient. But the effect of organisational participation is no longer significant for organisational commitment or affective psychological well-being. Overall, then, in contrast to task discretion, the effects of organisational participation pass to a greater extent through their implications for other aspects of the work environment.

SUMMARY AND DISCUSSION
In the light of the conflicting arguments in the literature, our objective has been to use newly available representative longitudinal data for Great Britain to examine more rigorously the relationships between direct participation and organisational commitment on the one hand and employee well-being on the other. There also has been substantive disagreement about the implications of direct participation. Some have seen it as enhancing the welfare of both management and employees by increasing employees' commitment to their organisations, while improving employees' well-being. But others have argued that it may increase employees' sense of commitment to their organisation but at the cost of a deterioration in their conditions of work. To assess these arguments, we examined the effects of direct participation on organisational commitment, job satisfaction and affective psychological well-being.
Past research on the effects of direct participation on well-being has focused mainly on one specific type of direct participation (task discretion) and has been primarily based on cross-sectional data, leaving open the possibility that apparent relationships are artefactual, resulting from unobserved differences between individuals. In addressing these issues, our data have several distinctive strengths. They enable us to look at direct participation both in terms of control over the job task and in terms of wider participation over organisational decisions. A longitudinal design makes it possible to take account of unobserved individual heterogeneity by examining the implications of changes in participation for changes in organisational commitment and well-being within a given individual. Further, the data provide indicators of several possible mediators of the effects of participation. It is therefore possible to examine whether its implications for well-being are direct, relating to its intrinsic importance for employees, or indirect, resulting from the way it affects other characteristics of the work environment that have important influences on wellbeing. One limitation to our study is that there are only two waves in the data, just two years apart, and this short period might lead to some underestimation of the effects of participation, if such effects were to come about with a considerable lag. However, despite this limitation, our analysis has uncovered significant effects of both types of direct participation on organisational commitment and well-being, which suggests that an authentic relationship is likely to exist even with conservative tests.
The analysis leads to four main conclusions. First, in so far as there are significant overall effects of participation, they are positive rather than negative for both organisational commitment and employee well-being. This is the case both for task discretion and for organisational participation. With respect to the contrasting arguments in the literature, the results provide support for the more optimistic view about the implications of organisational participation (see Hypothesis 1a), implying that it has benefits for both employers and employees. If the sceptical view (see Hypothesis 1b) were correct, organisational participation could be expected to encourage commitment, but be detrimental to job satisfaction and affective psychological well-being, because it would lead to deterioration in the quality of jobs as a result of increased work intensity. This was not, however, the pattern that emerged. It was the case that organisational participation increased organisational commitment, but neither organisational participation nor task discretion were associated with higher work intensity. Moreover, both types of direct participation were associated with higher job satisfaction.
Second, the results show that the positive relationship between direct participation and organisational commitment and well-being cannot be adequately accounted for by unobserved individual heterogeneity. Our estimates are based on the effects of within-individual changes over time, and thus control for personal fixed effects such as individual differences in motivation and disposition. While unobserved heterogeneity would appear to underlie some of the associations that are found if personal fixed effects are not controlled for (for instance in the relationship between task discretion and organisational commitment), there are still significant participation effects for each of the measures of well-being. Our results then are not consistent with the view (see Hypothesis 1c) that the relationships are purely artefactual, reflecting the fact that dispositions that encourage individuals to participate also lead to higher levels of well-being.
Third, we have found that the effects of direct participation reflected its intrinsic desirability as well as its implications for other aspects of the work environment (see Hypotheses 2a and 2b). We hypothesised (see Hypothesis 2b) that it might influence worker attitudes indirectly through its effects on the quality of training provision, skills use and job security. Introducing the mediating variables attenuated the strength of the effects of participation in a way consistent with the view that indirect effects are important. There was no longer evidence that direct participation had any direct effect on organisational commitment, indicating that its positive effect was due to improvements it produced in other aspects of the work environment. However, it is notable that both task discretion and organisational participation still had direct effects on well-being even after the mediating variables had been taken into account. This was true for both task discretion and organisational participation with respect to job satisfaction, although only task discretion had direct effects on affective psychological well-being. Overall, the effects of task discretion appeared to be primarily direct, with little change following the introduction of potential mediating variables. It is likely that this reflects the inherent value to employees of being able to exercise control over their working lives, either because of basic needs for autonomy or salient cultural values of selfdetermination.
Finally, it should be noted that each type of participation-whether task discretion or organisational participation-was shown to have positive effects even when considered together. Moreover, although both were positive for job satisfaction, they also had distinctive effects: only task discretion had benefits for affective psychological well-being, while only organisational participation was important for organisational commitment. They are then not substitutable but rather complementary forms of direct participation. This points to the need for a broad strategy in enhancing opportunities for direct participation, increasing opportunities for employee influence at different organisational levels. Direct participation, in this broader form, is likely to be a particularly important aspect of any policy initiative to improve the quality of work, because it is not only intrinsically important to people, helping to meet their needs for self-determination, but also contributes to a wider improvement in the work environment.