1Management control not only shapes decision-making processes through tools such as budgeting and reporting systems, but also strongly influences managers’ behavior (Grafton et al. 2010). This second dimension has generated a significant stream of research,
specifically in the Anglo-Saxon literature (Naro 1998) in the field of RAPM (Reliance on Accounting Performance Measures). A number of studies, for example, examine how results-based control systems (setting of objectives, performance measurement, and reward allocation)??”one of the most frequently used types of control in the West (Merchant and Otley 2006)??”influence managers’ attitudes and behavior and, indirectly, their performance. These studies have analyzed both the negative
effects of results-based control systems, such as the adoption of dysfunctional behavior or job-related tension, as well as its positive effects, for instance on job satisfaction and performance. As stated by Hartmann (2000, p. 451) in an RAPM-focused article of the Accounting Organization and Society journal, Anglo-Saxon researchers have noted that “RAPM constitutes the only organized critical mass of empirical work in management accounting at present.” However, despite the
significant number of studies making up this research trend, very few have examined the effects of results-based control systems on the key variable of managers’ goal commitment. This situation is all the more surprising given that organizational behavior literature, which has inspired research on results-based control systems, has in fact recognized the importance of goal-commitment. 2Goal theory, one of the main organizational behavior theories, postulates that specific, difficult-but-attainable goals lead to stronger individual performance than ill-defined and easy goals (Locke and Latham 1990). This relationship is, however, contingent on goal commitment (Locke and Latham 2002). Locke et al. (1988, p. 23) contend that “it is
virtually axiomatic that if there is no commitment to goals, the goal-setting does not work.” Numerous studies in the fields of organizational behavior (Locke et al. 1988; Locke and Latham 1990; Klein et al. 1999) and management accounting (Kren 1990; Chong and Chong 2002; Wentzel 2002; Maiga 2005; Chong and Johnson 2007), demonstrate that individuals perform better when they accept, and commit to attaining, a particular goal. Commitment to a goal mobilizes effort and
increases persistence, and is thus a direct determinant of performance (Kren 1990). In their meta-analysis, Klein et al. (1999) find a positive correlation between goal commitment and performance, although they observe that the relatively few studies in the field have explored the effects of goal commitment rather than its antecedents. They conclude that “the sparse attention given to … antecedent variables [of commitment] suggests that this research has not been sufficiently
systematic” (Klein et al. 1999, p. 893). 3Management control systems, in particular when they lead to results-based control, may well play a key role in enhancing or reducing goal commitment. As such, management control systems are potentially important antecedents of goal commitment. By gaining a better understanding of how management control systems
affect managers’ goal commitment, researchers and management control practitioners may identify ways of improving their effectiveness. However, very few studies have attempted to determine which management control system dimensions increase goal commitment and how this relationship operates. 4The purpose of our study is to expand our knowledge of goal commitment
antecedents by examining the role of two variables that are directly related to the concept of objectives, namely participation in goal setting and quality of performance feedback. We would expect participation in goal setting to lead to goal acceptance since individuals are more willing to accept objectives when they have been involved in their definition. However, the link between participation and goal commitment has yet to be clearly established. Previous studies have generated mixed results
(Locke et al. 1988). Certain recent studies have confirmed a positive relationship between participation and goal commitment, but on the basis of small samples and/or specific contexts. Concerning performance feedback, very few studies have empirically examined the extent to which providing good quality feedback on performance increases managers’ goal commitment. Providing managers with regular feedback throughout the year enables them to gauge their performance with regards to
their goals, which should in turn increase their goal commitment. 5Since the existing results are scarce and somewhat mixed, further investigation is needed to better understand the relationship between participation in goal setting and the quality of feedback on the one hand and goal commitment on the other. In addition, the effect of mediation variables also needs
to be considered. Managers’ objectives are set via a process in which managers interact with their supervisors. Supervisors may, or may not, facilitate managers’ participation in this goal-setting process. Similarly, supervisors are important sources of performance feedback, although their level of involvement in analyzing results and understanding the causes of variances will vary. The supervisor’s attitude when establishing objectives and assessing performance probably has a strong impact on
the relationship that forms between the supervisor and his or her managers. In particular, we would expect the supervisor’s attitude to influence managers’ level of trust in their supervisors. If managers trust their supervisors, they will be more inclined to commit to their goals. Very few studies have examined the role of trust in the relationship between participation, feedback, and goal commitment, despite organizational behavior research showing that trust plays a key role.
For instance, trust encourages employees to do their best to contribute to achieving organizational goals and improves the exchange of information between employees (Kramer, 1999, p. 583). 6Our study therefore examines the mediating effect of trust in the supervisor. Finally, as both participation in goal setting and performance feedback are expected to affect the
level of trust, this study also considers their combined effect on trust. 7Our study has three objectives. Firstly, it seeks to identify which dimensions of results-based control systems enhance managers’ goal commitment. More precisely, our study focuses on two dimensions, namely participation in goal setting and the quality of feedback provided to managers during
the year and at year end. The effects of these two variables on goal commitment are at present little-known; however, as companies frequently use them in their control processes, they merit further examination. 8Secondly, our study seeks to better understand the supervisor’s role in the relationship between management control systems and managers’ goal commitment. In
particular, we aim to assess the extent to which participation in goal setting and quality of feedback help to increase trust in the supervisor which, in turn, fosters managers’ commitment to their goals. 9Finally, our study examines whether participation and quality of feedback act independently or whether their effects are greater when combined. From a practical
point of view, this question is important since it enables companies to determine whether it is relevant and/or effective to improve participation and feedback separately or concomitantly. 10To summarize, the objective of our study is to examine the direct and interactive effects of participation in goal setting and performance feedback on managers’ goal
commitment, introducing trust in the supervisor as a mediating variable. 11This paper is organized as follows. In section 1 we review the literature and formulate our hypotheses. In section 2 we present our research method, followed by our results in section 3. Finally, in section 4 we discuss our findings and set out the limitations of our study. 12In section 1.1, we explore the concept of goal commitment and examine its relationship with participation (1.1.1) and feedback (1.1.2). In the following section, we analyze the role of trust (1.2), and explore its relationship with goal commitment (1.2.1), with participation (1.2.2), and with feedback
(1.2.3), as well as its mediating effect (1.2.4). Lastly, we examine the combined effect of participation and feedback on trust (1.2.5). 13Goal commitment is a long-standing variable in the literature on organizational behavior and motivation. A central postulate in goal-setting theory is that specific, difficult-but-attainable goals improve
performance (Locke et al. 1981; Locke and Latham 1990); however, this is only true if individuals are committed to achieving these goals. From the outset, goal-setting theory recognized goal commitment as a key mediating variable in the relationship between setting objectives and performance (Locke 1968), before going on to specifically examine its role (Locke et al. 1988; Hollenbeck et al. 1989). 14Previous studies have suggested several definitions and measures for goal commitment. Goal commitment is generally defined as an individual’s determination to extend effort towards a goal over time in order to achieve it (Locke et al. 1981; Locke and Latham 1990). 15Klein et al. (2012) recently developed a clarified and generalized concept of commitment, which can be applied to any type of commitment. According to these authors, commitment is a “volitional psychological bond reflecting dedication to and responsibility for a particular target” (p. 137). This definition of commitment can be applied to any target within the
workplace: the target may be the organization, leading to the notion of organizational commitment, or a specific objective leading to the concept of goal commitment, which is the focus of our study. The definition put forward by Klein et al. (2012) is consistent with the above-mentioned definitions. It encompasses the notion of effort deployed and maintained over time to achieve the objective, and takes into account every type of objective, whether it is imposed, self-defined, or
negotiated. 16Based on this definition and on the literature, Klein et al. (2012) build a model describing the effects and the determinants of commitment. The latter include characteristics of the individual involved and of the target, as well as a number of interpersonal, social, and organizational factors, such as human resources practices. Klein
et al. (2012) do not specifically mention control systems, although the literature has repeatedly demonstrated that these systems influence attitudes and behavior, such as job satisfaction or job-related tension. One can therefore infer that control systems also influence goal commitment, even if relatively few studies have empirically examined this relationship. 17In addition, according to the Klein et al. (2012) model, these antecedents will influence the way in which an individual perceives the objective and his or her commitment. This model suggests that commitment will be stronger if the objective is salient, viewed positively, perceived as attainable, and if it exists in an environment of trust. This last aspect is the
focus of our study. 18The literature on goal commitment addresses all types of individual. In this paper, we will focus on managers and on their commitment to objectives, which are often set by results-based control systems. 19Managers’ participation in setting their goals is one of the most explored issues in management control (Chong and Johnson 2007), often referred to as budget participation. 20Brownell (1982, p. 124) defines participation as “an organizational process
whereby individuals are involved in, and have influence on, decisions that have direct effects on those individuals.” Goal setting is an example of such a decision. 21A significant body of research has been published on the links between participation and managerial performance, but has not confirmed the existence of a positive relationship (see Bonache
et al. 2012 for a meta-analysis). One possible explanation for these mitigated results is that the relationship between participation and performance might be mediated or moderated by a number of other variables (Derfuss 2009). Goal commitment is one such variable (Chong and Chong 2002; Wentzel 2002; Maiga 2005; Chong and Johnson 2007). 22The first studies conducted in the field of management control suggested that participation should encourage managers to accept their goals and feel committed to them (e.g., Argyris 1952; Hofstede 1967; Dunbar 1971). 23Several arguments have been put
forward in the literature to account for the link between participation and goal commitment. Firstly, participation allows objectives to be clarified or justified. Participation enables subordinates to obtain information and explanations from their supervisors, leading to a greater understanding of the supervisors’ expectations, of the allocation of resources, and of the reason for the objectives. This improved understanding reduces uncertainty and ambiguity and helps managers to focus their
efforts on the goals to be attained (Shields and Shields 1998; Chong and Chong 2002; Li and Butler 2004; Maiga 2005; Chong and Johnson 2007). Secondly, participation enables subordinates to influence their objectives. According to Maiga (2005), participation encompasses two dimensions: communication and influence. Participation enables subordinates to influence their objectives, thus making them more accurate and realistic. In addition, because of this participation, managers believe they have
greater control over their objectives (Hollenbeck et al. 1989; Klein et al. 2012). A final argument relates to the effect of participation on self-esteem. Participation strengthens organizational identification because it is rewarding and thus increases self-esteem: “One may thus categorize oneself more easily as a significant member of an in-group” (Maiga 2005, p. 216). Participation should therefore lead managers to better accept and commit to their objectives
(Argyris 1952; Hofstede 1967; Locke and Latham 2002; Maiga 2005). 24Certain management control studies have empirically examined the relationship between participation in goal setting and goal commitment, with two of these studies showing a direct link. Chong and Chong (2002) found a significantly positive relationship between participation and goal commitment in a
sample of 79 middle managers in Australian industrial companies. Maiga (2005) examined two dimensions of participation: 1) the extent to which information is communicated back and forth between headquarters and sub-unit managers in setting or revising budget objectives, and 2) the extent to which sub-unit managers influence these objectives. Based on a sample of 173 American investment responsibility centers, Maiga (2005) found that these two dimensions of participation were positively
and significantly related to managers’ goal commitment. 25Other management control studies have shown that the relationship between participation and goal commitment is indirect since it is mediated by one or several variables: perceived justice (Wentzel 2002; Maiga and Jacobs 2007; Sholihin et al. 2011), information relevance, the difficulty and
acceptance levels of the goal (Chong and Johnson 2007), or trust (Maiga and Jacobs 2007; Sholihin et al. 2011). 26Most of these studies use small size samples and/or examine specific professional environments. In order to generalize their results, the following hypothesis will be tested: 27H1: Participation in goal setting is positively related to managers’ goal commitment. 28Feedback is defined as information provided to an individual regarding the relevance
and the effectiveness of their decisions and/or behavior in relation to their objectives (Ilgen et al. 1979; Renn 2003). Several sources of feedback co-exist: for instance, feedback can be internal, that is to say generated by the individuals themselves, or it may come from an external source (Herold and Fedor 2003). External feedback may result from exchanges with co-workers or be received from supervisors (Steelman et al. 2004). In this study, we
consider this second type of feedback, namely information provided by supervisors at different moments during the year, including at year end. This type of feedback plays a key role in results-based control systems since it helps managers to make appropriate decisions to improve their performance (Hartmann and Slapni?ar 2009). Feedback may concern either results or behavior (Kim 1984), and can be formal or informal (Pitkänen and Lukka 2011). Formal feedback provides managers with structured
information based on comparisons between actual performance and pre-set goals or standards. Informal feedback is provided through discussions between the manager and his or her supervisors. Distinguishing between formal and informal feedback practices is, however, somewhat ambiguous, as they are strongly interlinked (Pitkänen and Lukka 2011). Nevertheless, this distinction can help us to identify how feedback may affect commitment and trust. 29Two theories can be used to explain the effect of feedback on motivation and performance. According to the cybernetic control model, feedback makes it possible to verify whether the state of a system is consistent with a predefined standard; if it is not, the system triggers an adjustment process to close the gap. According to this approach, feedback encourages individuals to take
steps to reduce the gap between actual performance and objectives (Klein 1989; Renn 2003). Cognitive motivation theory, on the other hand, suggests that feedback improves performance because it facilitates the detection of mistakes, learning, and the development of strategies aiming to improve task efficiency (Ashford and Cummings 1983). Having information on their own performance may thus make individuals more motivated in their work (Fried and Ferris 1987; Renn 2003). 30These two theories suggest a link between feedback and goal commitment. If individuals receive good quality feedback on a regular basis, they will have information enabling them to gauge how close they are to their objectives. Such information is likely to foster their goal commitment for two reasons. On the one hand, this information reminds the manager of his or her objectives.
If no feedback is provided, there is a risk that the manager will focus on the everyday management of his or her entity without keeping in mind the objectives that were set at the beginning of the year. On the other hand, the information provided through good quality feedback alerts managers to the variances between actual and expected results. We have seen above that commitment involves the maintenance of effort over time. Feedback provides information about negative variances and is therefore
a useful warning to managers, encouraging them to maintain or increase effort so as to reach their objectives. 31The literature suggests that good feedback provides useful information to subordinates, namely, accurate information communicated with the appropriate frequency and timeliness by a reliable source (Ilgen et al. 1979; Steelman et al.
2004; Hartmann and Slapni?ar 2009; Pitkänen and Lukka 2011). This good feedback will retain the attention of the manager, as suggested by Ashford and Northcraft (2003) who state that individuals with multiple tasks give greater attention and priority to a particular task when they receive specific, favorable, and frequent feedback thereon. 32Empirical research
provides evidence that effective feedback affects managers’ attitudes, behaviors, or work performance (Kenis 1979; Steelman et al. 2004; Steelman and Rutkowski 2004; Hartmann and Slapni?ar 2009; Kuvaas 2011). Kuvaas (2011), for instance, conducted a survey based on 803 questionnaires collected from three Norwegian organizations (a bank, a governmental agency, and a pharmaceutical company). Results show that well perceived feedback is positively related to
organizational commitment. Moreover, when feedback is perceived to be useful, performance improves in function of the frequency thereof. 33However, the relationship between feedback and performance is not always significant; several moderating variables can interfere in this relationship (Kluger and DeNisi 1996; Steelman et al. 2004). According to
Renn (2003), commitment is required for the relationship between feedback and performance to exist since commitment increases the effectiveness in acquiring, processing, and using the information provided during the feedback. It is therefore important to examine the relationship between feedback and managers’ commitment. Chong and Johnson (2007) found that managers who receive job-relevant information are more committed to their goals. In reality, their study refers to a different variable since
it examines the relevance of all types of information provided to managers and not only feedback on their actual performance. Martocchio and Dulebohn (1994), however, did not reach the same conclusion. They examined the specific case of feedback provided to 86 university employees during a software program training course. Their results showed no significant relationship between the feedback communicated to participants and their commitment, in an, albeit, very specific context. 34Empirical evidence concerning the link between feedback and commitment is scarce and somewhat mixed, indicating that further investigation is required. The above arguments suggest that good feedback should increase commitment. We therefore propose the hypothesis that when managers receive performance feedback, their goal commitment will increase. 35H2: Performance feedback is positively related to managers’ goal commitment. 36“Trust is a psychological state comprising the intention to accept vulnerability based on
positive expectations of the intentions or the behavior of another” (Rousseau et al. 1998, p. 395), without “the former having the possibility of surveillance over the latter” (Mayer et al. 1995, p. 712). This definition covers different parties and relationships between these parties (for example, the parties to a contract). In this study, we focus on trust in the supervisor. Subordinates who trust their supervisors expect them to act supportively and benevolently,
protecting and promoting their interests and showing consideration and sensitivity to their needs (Read 1962; Whitener et al. 1998). We are particularly interested in how trust may influence the way in which managers think they will be treated by their supervisor in terms of goal setting, performance evaluation, and rewards. Management control research was quick to take up the topic, and continues to examine this particular form of trust (Hopwood 1972; Otley 1978; Sholihin
et al. 2011). 37Trust in the supervisor will develop and remain if the supervisor displays a number of “trustworthy” characteristics. According to the model developed by Mayer et al. (1995), which has been frequently used in subsequent research, the trustee needs to possess three essential characteristics. The first is ability, namely a
group of skills, competencies, and characteristics that enable the superior to have influence within some specific domain. The second characteristic is benevolence. Benevolence is the extent to which a trustee is believed to want to do good to the trustor, acting in a way that protects the trustor’s interests. The third dimension is integrity, which involves the trustor’s perception that the trustee adheres to a set of principles that the trustor finds acceptable. Such issues as
the consistency of the supervisor’s past actions, credible information about the supervisor from other parties, belief that the supervisor has a strong sense of justice, and the extent to which the supervisor’s actions are congruent with their words, all affect the degree to which the supervisor is judged to have integrity. As the relationship between a manager and their supervisor develops, the manager forms an opinion about the trustworthiness of their supervisor. 38The literature suggests that a manager’s goal commitment is influenced by the manager’s level of trust in his or her supervisor. This relationship can be explained by goal theory (Locke et al. 1981; Locke et al. 1988; Locke and Latham 1990), and by social exchange theory (Homans 1958; Blau 1964; Cropanzano and
Mitchell 2005). 39Goal theory suggests that trust is an important determining factor of managers’ commitment. In a situation of trust, managers will tend to think that their supervisors are not seeking either to manipulate or go against their interests when setting goals. Managers will think that goals have been set with consideration and even benevolence, and will
therefore be more inclined to accept and appropriate the objectives. 40Social exchange theory (Homans 1958; Blau 1964; Cropanzano and Mitchell 2005) also provides an explanation for the effect of trust on managerial commitment. According to this theory, organizations are transaction forums that generate the obligation to reciprocate. In a social exchange, one
individual voluntarily provides a benefit to another, invoking an obligation of the other party to reciprocate by providing some benefit in return. According to some researchers, trust in the supervisor follows the principles of this social exchange theory (Whitener et al. 1998; Dirks and Ferrin 2002). Individuals consider that their relationship with their supervisors, and more globally with their organization, cannot be reduced to a short-term economic exchange, but incorporates
a long-term relationship and reciprocal obligations. If they believe their supervisors will show consideration and benevolence and will act in their best interests, they will reciprocate by adopting the behavior required by their supervisors (Dirks and Ferrin 2002). When managers trust their supervisors, they will be more willing to do what their supervisors and the company expect from them and in particular, to commit to their goals. Conversely, if individuals do not trust their supervisors,
they will try to protect themselves instead of focusing their effort on improving their performance (Mayer et al. 1995). More precisely, “believing that a leader is not honest, does not have integrity, and may take advantage of a follower is likely to make one unwilling to commit to the goals set by a leader, for fear of putting oneself at risk” (Dirks and Ferrin 2002, p. 614). 41In the above-mentioned model, Klein et al. (2012) consider trust to be a component of the cognitive and affective processes that affect the individual’s perception of the objectives and their context. According to these authors, trust favors two central dimensions of commitment, namely dedication and responsibility for a target (Klein et al. 2012, p.
140). In fact, a minimum level of trust is required for individuals to perceive objectives positively and feel motivated to achieve them. 42Empirical research confirming the relationship between trust and goal commitment is scarce. An early study by Oldham (1975) found that when individuals trust their supervisors, they are more likely to
appropriate the objectives set by these supervisors, with appropriation being one of the determinants of goal acceptance. More recently, the two aforementioned studies (Maiga and Jacobs 2007; Sholihin et al. 2011) found that trust in the supervisor affects goal commitment. 43These theoretical considerations and the paucity of empirical evidence lead
us to test the following hypothesis: 44H3: Trust in the supervisor is positively related to managers’ goal commitment. 45Goal setting is a process that
generally involves managers and their supervisors. The latter play a key role since they decide whether to allow managers to take part in discussions and the extent to which they will take the managers’ points of view into consideration. In a participative process, supervisors provide managers with explanations. If managers are able to participate in setting their objectives, they will regard their supervisors as showing respect and benevolence, and therefore as being trustworthy. It is
therefore likely that participation in goal setting will improve managers’ trust in their supervisors. 46Research has confirmed the relationship between participation and trust, both theoretically (Shields and Shields 1998; Whitener et al. 1998; Dirks and Ferrin 2002) and empirically (Lau and Buckland 2001; Lau and Tan 2006; Maiga and Jacobs 2007;
Sholihin et al. 2011). When managers participate in setting their goals, they feel that their supervisors are sharing and delegating control. According to the social exchange theory set out above, managers perceive the opportunity to participate as a sign of respect and benevolence on the part of their superior. This, in turn, increases their trust in the superior (Whitener et al., 1998). In addition, several studies have empirically examined the link between
participation and trust. Lau and Buckland (2001) examined the effects of emphasizing goal attainment on job-related tension based on a sample of 160 managers from Norwegian industrial companies. Their results showed that participation and trust, which operate as mediating variables in their model, were significantly and positively related. Similarly, Lau and Tan (2006) conducted a survey on 152 functional managers in Singapore-based companies and found a significant link between these
two variables, both direct, and indirect through procedural perceived justice. This finding is also confirmed by Sholihin et al.’s study (2011), mentioned above, although other studies mitigate these results. For instance, Bijlsma and Van de Bunt (2003) used 76 interviews and 1,053 questionnaires in a hospital to identify the antecedents of trust in the supervisor. They concluded that participation was not a significant determinant of trust. 47We will therefore test this relationship, and formulate the following hypothesis: 48H4: Participation in goal setting is positively related to trust in the supervisor. 49Feedback, in particular informal feedback, helps promote discussion between managers and their supervisors (Malo and Mathé 1998, p. 125 ; Pitkänen and Lukka 2011). The exchange that takes place during this feedback offers a valuable opportunity for managers to assess the ability, benevolence, and integrity of their supervisors, and hence to
observe whether the supervisors show characteristics that make them trustworthy (Mayer et al. 1995). In fact, during this exchange the superiors analyze their subordinates’ results. This allows superiors to demonstrate their competence by helping their subordinates to identify the causes of potential negative variances or by examining possible corrective action plans with them. In addition, the superior’s involvement is likely to be perceived as proof of benevolence since it
reflects the superior’s willingness to help the manager reach his or her objectives. 50Research on the relationship between feedback and trust has mainly examined the extent to which pre-existing trust helps to improve how feedback is received and used. Trust seems to be particularly important when feedback is unfavorable. According to Audia and Locke (2003),
objective use and assessment of unfavorable feedback is possible only if the source of the feedback is considered to be trustworthy due to his or her ability, integrity, and benevolence. Similarly, O’Leary-Kelly and Newman (2003) demonstrated theoretically that unfavorable feedback may explain antisocial or even aggressive behaviors when the former comes from sources believed to be untrustworthy. Based on a sample of 405 managers working in two industrial companies, Steelman and Rutkowski
(2004) showed that individuals who receive unfavorable feedback are motivated to improve their performance only if the feedback is provided by a person believed to be trustworthy, if it is delivered with consideration, and if the quality of feedback is good (i.e. consistent and useful). In a different context, Peterson and Behfar (2003) relied on a longitudinal study conducted over a year on 67 teams of four MBA students to show that the reception of unfavorable feedback increased
intra-group conflicts in the teams in which the level of trust between the members was weak. 51Conversely, few studies have attempted to examine the opposite relationship, namely the effects of feedback on the development and the retention of trust. In the above-mentioned study by Bijlsma and Van de Bunt (2003), the interviews highlighted the importance of
feedback. Results demonstrated that feedback is positively and significantly related to trust. Interviews showed that employees who do not trust their supervisors complain about their supervisors’ distance, lack of interest, and consequent lack of knowledge about their subordinates’ performance. On the other hand, supervisors who give fair feedback to their subordinates and help them improve their performance foster their subordinates’ trust in them. The qualitative study was complemented by
quantitative tests that confirmed a significant positive link between feedback and trust. Hartmann and Slapni?ar (2009) conducted a survey on 160 Slovenian bank employees and found that managers who receive good quality feedback from their supervisors have more trust in them. 52Research that empirically demonstrates the effect of feedback on trust is thus
scarce, in particular in the management control literature. This therefore leads us to test the following hypothesis: 53H5: Performance feedback is positively related to trust in the supervisor. 54Considering our previous discussion and hypothesis, it can be assumed that trust is a mediating variable in the relationship between participation and goal commitment on the one hand and feedback and goal commitment on the other. In other words, it is because participation and feedback foster trust that they increase goal commitment. The exchanges that take place
between managers and their supervisors during participative goal-setting processes and during performance feedback appear to increase managers’ trust in their supervisors and, indirectly, their commitment to achieving their goals. More precisely, it is a question of knowing whether trust is the only variable explaining the effects of participation and feedback on goal commitment (total mediation), or if the direct effects of participation and feedback on goal commitment remain significant after
adding trust to the model (partial mediation). 55Two studies have highlighted the mediating effect of trust in the relationship between participation in goal setting and goal commitment (Maiga and Jacobs 2007; Sholihin et al. 2011). Maiga and Jacobs (2007) found, in a sample of 163 questionnaires collected in industrial divisions, that
participation is related to budget goal commitment, through the mediating role of perceived justice and trust. Sholihin et al. (2011) found the same type of relationship in a sample of 54 managers in a British financial institution. These two studies were conducted on small samples and in specific contexts. In addition, no studies have examined the mediating role that trust can play in the relationship between feedback and goal commitment. 56We will therefore test the following hypotheses: 57H6: Trust mediates the relationship between participation in goal setting and goal commitment. 58H7: Trust mediates the relationship between performance feedback and goal commitment. 59If, as we suppose,
participation and feedback affect trust, we also need to examine their combined effect. Are these effects independent or do they mutually reinforce each other? It appears likely that trust can only be created, and above all preserved, if the interactions between managers and their superiors take place regularly throughout the year. In the long term, it is the sequence of participative goal-setting stages (for instance, at the beginning of each year), and of regular feedback during the year and
at year end, that enable trust to develop. If this is the case, the interaction between participation and feedback would be a significant predictor of trust, to a greater extent than each of these variables considered separately. 60We therefore test the following hypothesis: 61H8: The interaction between participation and feedback is positively related to trust. 62Figure 1 shows the full theoretical model tested in this research. Figure 1 Theoretical model Theoretical model2 – Method63This section presents our data collection method and the sample description (2.1), the measurement instruments employed (2.2), and lastly the statistical method used (2.3). 2.1 – Data collection and sample64In order to test our hypotheses, we used a quantitative approach based on a survey, which was the most suitable method of collecting perceptions anonymously (Hartmann and Slapni?ar 2009). We targeted managers whose level of responsibility made it likely that their performance would be evaluated via the setting and monitoring of objectives. 65Our purpose was also to target managers from different sectors, companies, and functions so as to improve the external validity of our results. We collected the survey data from managers attending 18 different training seminars in three business schools. Most of these seminars were conducted by professors other than the authors of the present article. Questionnaires were handed out during breaks or before the seminars started, and took about 10 minutes to complete. A short oral presentation was delivered informing participants that the research did not pertain to the topics covered in the training seminars and that completing the questionnaire was not compulsory. We distributed 381 questionnaires and collected 370 responses. Of the 370 respondents who returned their questionnaires, 45 declared that objectives were not used to evaluate their performance, and that they could therefore not answer the main questions in this study. This left 325 questionnaires, representing an 85% response rate. A missing value analysis showed no variables with 5% or more missing values; however, four cases with more than 5% missing data and one outlier were removed from the sample. Our data analyses were therefore performed on a total of 320 questionnaires. 66The companies in the sample are relatively large, in terms of both headcount (< 500: 37%; 500-2000: 15%; > 2000: 48%) and sales revenue (< €1.5 million: 5%; €1.5-50 million: 27%; > €50 million: 68%). The main sectors are industry (44%) and non-financial services (16%), and 52% are listed on the stock market, or belong to a group that is listed. Respondents’ tenure in their company suggests that most of them know their company well (< 2 years: 8%; 2-5yr: 33%; > 5yr: 59%). They have managerial positions in many different functions (e.g., sales, finance, manufacturing, R&D, HR, supply chain). 2.2 – Measurement instruments67All instruments were taken from the existing literature. The 14 questions used to measure the four constructs asked participants to indicate their level of agreement with a statement on a five-point scale. The length and clarity of the questionnaire were first tested on 19 managers, with this pre-test leading to some small wording adjustments. Table 1 shows the 14 questions (i.e., items) and their four corresponding constructs. We present below the origins of each instrument. Table 1 List of items and descriptive statisticsList of items and descriptive statisticsN = 320 for all items. Answers range from 1 to 5 for all items. 68We measured participation in setting objectives (PART) with four items taken from the widely used Kenis (1979) instrument. These four items take into account two dimensions of the participation process: the consultation by the supervisor when objectives are set (items P2 and P3, see Table 1) and the influence that the manager has on those objectives (items P1 and P4). 69As for performance feedback (FDBK), we combined two items from the instruments proposed by Kenis (1979) and one from Steelman (2004). We selected these items because they measure the manager’s perception of the amount (items F1 and F2) and usefulness (F3) of the feedback he or she receives. 70We used Hartmann and Slapni?ar ‘s (2009) three-item instrument, itself taken from Read’s (1962) instrument, to measure trust in the supervisor. We asked managers whether 1) they believe that their supervisor will always act in managers’ interests if he or she has the chance (T1); 2) they think that their supervisor keeps them fully and honestly informed about what is important to managers (T2); and 3) they are convinced that if their supervisor makes a decision contrary to their interests, this decision is justified by other reasons (T3). We chose to use this instrument because it clearly corresponds to the concept of trust as being the willingness of an individual to be vulnerable to another based on positive expectations regarding this other person’s intentions and behavior. 71Prior studies have used many different instruments to measure goal commitment (GOAC). To allow for comparisons with studies in the control area, we used Wentzel’s (2002) three-item instrument (G1 to G3). This measurement instrument has been used by other authors in the control field (Maiga 2005; Maiga and Jacobs 2007), and is very similar to the instrument used by Sholihin et al. (2011). Wentzel’s instrument does not, however, include the determination dimension of goal commitment, namely the idea that the individual maintains the intention of reaching the goal over time. In order to account for this dimension, we therefore added a fourth item (G4) taken from Klein et al. (2001). 72Table 1 shows the 14 items, as well as their univariate statistics. It also shows the mean and standard deviation of each construct using the average of their corresponding observed items. 2.3 – Statistical method73To test our hypotheses and model (Figure 1), which in particular assumes that participation and feedback affect goal commitment through the mediating effect of trust, we used structural equation modeling (SEM). SEM is an appropriate method for testing models with mediating effects and is often used in management accounting research (Bisbe et al. 2007; Henri 2007). However, several approaches are possible depending on, among other things, the research purpose, the sample size, and the distribution of items (Hair et al. 2011). Because several items did not follow a normal distribution, we chose a partial least squares (PLS-SEM) approach, using SmartPLS 2.0 software (Ringle et al. 2005). The objective of PLS-SEM is to maximize the explained variance of the dependent (latent) variables. Hence, it is mainly a predictive approach. On the other hand, unlike covariance-based SEM approaches (e.g., AMOS or LISREL), PLS-SEM does not calculate goodness-of-fit indices, but, instead generates indicators that assess the predictive abilities of the overall model (coefficient of determination (R2 value) and Stone-Geisser’s Q2 value) and of each explanatory variable (standardized path coefficients and f2 effect sizes). A bootstrapping procedure is used to test path coefficients’ statistical significance. [1] 74SEM literature usually recommends using a two-step approach to analyze latent variable models (e.g., Bisbe et al. 2007; Henri 2007; Hair et al. 2011; Hair et al. 2014). In the first step, the approach assesses the quality of the measurement model, in other words, the constructs’ validity and reliability. In the second step, the method analyzes the structural model and the hypotheses. We follow this two-step approach in the presentation of our results below. We then assess whether control variables affect the model. 3 – Results3.1 – Assessment of the quality of the measurement model75This first step consists in analyzing the reliability of each indicator (i.e., measured items), and the convergent and discriminant validity of each construct (i.e., latent variables). 76The reliability of each indicator can be assessed based on its factor loadings. To have an acceptable reliability level, each indicator must have a factor loading that is higher than 0.70 on its respective construct. Table 2 shows that the lowest factor loading is 0.765. Therefore, all indicators have more than 50% of their variance explained by their respective construct and have a satisfactory reliability. Table 2 Factor loadings and reliability of indicatorsFactor loadings and reliability of indicatorsAs all factor loadings are greater than 0.70, indicator reliability is considered to be satisfactory. 77Convergent validity is adequate when indicators of a given construct share more variance with this construct than with their measurement error terms. It can be assessed using the average variance extracted (AVE), which must be greater than 0.50, or using the Cronbach’s alpha or the composite reliability ratio (CR), which must both be greater than 0.70 (Hair et al. 2006). Table 3 shows that these conditions are fulfilled and that convergent validity is therefore acceptable for the four constructs of our analysis. Table 3 Analysis of convergent validity and discriminant validity of constructsAnalysis of convergent validity and discriminant validity of constructsCR: composite reliability; Alphas: Cronbach’s alphas; AVE: average variance extracted. 78Finally, discriminant validity measures the degree to which a construct is distinct from the others. It is ensured when a construct shares more variance with its indicators than with other constructs (Hair et al. 2006). In other words, the square-root of the AVE of each construct must be greater than the correlations between this construct and each of the three others. The correlation matrix in the right-hand panel of Table 3 shows that this condition is met: the diagonal bold figures represent the square roots of the AVEs. We can observe that they are all greater than the between-construct correlations located in the other cells. Discriminant validity is thus acceptable for each of the four constructs of this study. 3.2 – Analysis of the structural model and test of the hypotheses79The second step in the SEM approach assesses the structural model and the relationships between latent variables (i.e., constructs). We firstly observe that there is no collinearity problem between the explanatory latent variables of each part of the model (i.e., firstly, between participation and feedback and, secondly, between participation, feedback, and trust). Indeed, all VIF (Value Inflation Factors) have a value lower than 5 (the highest being 1.6). 80In order to test our hypotheses and the mediating and interactive effects, we used a multi-step approach (Hartmann and Slapni?ar 2009; Sholihin et al. 2011). Firstly, we tested whether participation and feedback affected goal commitment directly (H1 and H2). Secondly, we tested whether trust was related to goal commitment (H3), to participation (H4), and to feedback (H5). Thirdly, we tested whether trust mediated the relationships between participation and goal commitment, and between feedback and goal commitment, checking whether these mediations were total or partial (H6 and H7). Finally, we tested the interactive effect of participation and feedback on trust (H8). Table 4 presents our results. Table 4 Analysis of the structural model and test of hypotheses***,**Panel 1: Model without mediationPanel 1: Model without mediationPanel 2: Model with mediator: TRUSTPanel 2: Model with mediator: TRUST
Panel 3: Model with moderator: PART x FDBKPanel 3: Model with moderator: PART x FDBKAnalysis of the structural model and test of hypotheses***,**t: Student’s t 81Panel 1 in Table 4 presents the results regarding the direct effects of participation (PART) and feedback (FDBK) on goal commitment (GOAC), without including trust (TRUST) in the model. It can be seen that both path coefficients are positive and significant (p < 0.01). Participation and feedback are therefore both significantly related to goal commitment. Hypotheses H1 and H2 are supported. The positive value of Q2 (Q2 = 0.152) shows that participation and feedback are relevant predictive variables of goal commitment. Conversely, it can be observed that the coefficient of determination is small (R2 = 0.235), meaning that, taken together, participation and feedback explain less than 25% of the variance of goal commitment. The comparison of path coefficients (0.380 vs. 0.232) indicates that participation has a stronger effect on goal commitment than feedback. The sizes of these effects remain moderate for participation (f2 = 0.158) and low for feedback (f2 = 0.069). 82Panel 2 in Table 4 introduces trust (TRUST) as a mediator. Participation (PART) and feedback (FDBK) are seen to be positively and significantly related to trust (TRUST), with path coefficients of 0.259 (p < 0.01) and 0.501 (p < 0.01), respectively. This supports hypotheses H4 and H5. The two variables are relevant predictors of trust (Q2 = 0.288) and, together, explain 37% of the variance of trust (TRUST R2 = 0.374). We note, in particular, the high value of the feedback path coefficient, which has a high effect on trust (f2 = 0.379) compared with that of participation, which has a low to moderate effect (f2 = 0.102). In order to increase trust in supervisors, companies should focus on improving the quality of feedback rather than increasing the level of participation. 83Will this increase commitment? We verify that trust has a positive and significant effect on goal commitment (0.209; p < 0.01), which supports hypothesis H3. However, its effect size is low (f2 = 0.032). Taken together, PART, FDBK, and TRUST are relevant predictors of goal commitment (Q2 = 0.169) and explain 26% of its variance (GOAC R2 = 0.259), which remains moderate. 84Because of its low effect on commitment, is trust a mediating variable? To answer this question, we use Baron and Kenny’s (1986) three-step approach. Based on our results, we confirm that participation (PART) and feedback (FDBK) are both significantly related to goal commitment (GOAC) (Step 1) and to trust (TRUST) (Step 2), and that trust, in turn, has a significant effect on goal commitment (Step 3). We can thus conclude that trust mediates the relationships between participation and goal commitment and between feedback and goal commitment, which supports hypotheses H6 and H7. In addition, when comparing panels 1 and 2 in Table 4, we note that the path coefficient between participation and goal commitment decreases from 0.380 (p < 0.01) to 0.322, but remains strongly significant (p < 0.01). The mediating effect of trust in the participation/goal commitment relationship is, thus, only partial (Baron and Kenny 1986). The path coefficient between feedback and commitment remains significant when trust is introduced, although to a lesser extent, moving from 0.232 (p < 0.01) to 0.126 (p < 0.05). While the mediating effect remains partial for feedback, it is stronger than for participation. Overall, we note that the participation effect on commitment is mainly direct: out of a total effect coefficient of 0.377, the direct effect coefficient (0.322, i.e., 85% of the total effect) is much higher than the indirect effect coefficient (0.055 = 0.259 x 0.209, i.e., 15% of the total effect). Conversely, feedback mainly affects trust (coeff. = 0.501). As a result, its total effect on commitment (coeff. = 0.231) comes almost as much from its indirect effect through trust (coeff. = 0.105 = 0.501 x 0.209, i.e., 45% of the total effect) as from its direct effect (coeff. = 0.126, i.e., 55% of the total effect). As a whole, the total effect of participation on commitment remains higher than the total effect of feedback. 85Finally, do the respective effects of participation and feedback on trust reinforce each other? Panel 3 in Table 4 shows the results of this interactive effect for hypothesis H8. We created a latent variable to measure the PARTxFDBK interaction. More specifically, this variable’s twelve indicators are the product of the four participation items multiplied by the three feedback items (12 = 4 x 3) (Henseler and Fassott 2010). Panel 3 in Table 4 shows that the path coefficient of the relationship between this interaction variable and trust is very low and not significant (0.017; t = 0.087; p = 94%). Hypothesis H8 is not supported. Participation (PART) and feedback (FDBK) are independently related to trust (TRUST). 3.3 – Analysis of control variable effects86To complete our analysis, we tested the effects of nine control variables included in our questionnaire. We introduced each of these variables [2] into the model in turn, in order to test each variable’s effect on trust and, separately, on commitment. Of the 18 models thus obtained, only two relationships were significant. The first, between the number of employees and trust (coeff. = 0.089; t = 2.135; p < 0.05), led to a small increase in the TRUST item’s adjusted R2 from 0.370 to 0.376. The second relationship, between the hierarchical level and commitment (coeff. = 0.132; t = 2.549; p < 0.05), increased the adjusted R2 from 0.254 to 0.265. We then examined more thoroughly whether these control variables might influence the overall model’s results via two multi-group analyses (PLS-MGA) (Hair et al. 2014). Regarding the number of employees, we created two samples corresponding to the two extreme modalities (out of three) of the variable. The first sample (< 500 employees) comprised 116 cases and the second sample (> 2,000 employees) 155 cases. For the hierarchical level, we used cases on both sides of the median modality, obtaining two sub-samples of 107 cases each: the first corresponding to hierarchical levels above (but not including) “N-2”, where N-2 refers to two reporting levels below the highest-ranking supervisor; and the second corresponding to hierarchical levels equal to or below “N-3”, where N-3 represents three reporting levels below the highest-ranking supervisor. These PLS-MGA analyses did not identify any significant differences (at p< 0.10) in the model’s five relationships. Finally, to test for possible unobserved heterogeneity, we performed a FIMIX-PLS analysis (Rigdon et al. 2010). The lowest segmentation criteria (AIC, BIC, and CAIC) were obtained with two segments, increasing strongly beyond. Since the entropy criterion was 0.39 at this two-segment level, we could not conclude that unobserved heterogeneity existed in our study. Overall, we did not find any contingent effect that would have led us to modify our model and/or to qualify our results. Figure 2 shows the final model. Figure 2 Final model***,**Final model***,***** Path coefficient significant at p < 0.01 4 – Discussion and conclusions87The purpose of this study was to test whether participation in target setting and reception of performance feedback increase managers’ trust in their supervisors and, as a consequence, their goal commitment. Based on a sample of 320 managers, our results confirm that participation and feedback are both significantly related to trust and goal commitment. Trust is indeed a mediating variable, but its mediating effect is only partial, the direct effects of participation and feedback on goal commitment remain significant after adding trust to the model. 88Regarding participation, our results confirm the findings of previous research showing a direct relationship between participation and goal commitment (Chong and Chong 2002; Maiga 2005), between participation and trust (Lau and Buckland 2001; Lau and Tan 2006; Sholihin et al. 2011), and a mediating effect of trust in the relationship between participation and goal commitment (Maiga and Jacobs 2007; Sholihin et al. 2011). In particular, like Sholihin et al. (2011), we observe that the direct effect of participation on goal commitment is much stronger than its indirect effect through trust. 89Turning to feedback, our results contribute to a better understanding of a field in which research is scarce. They confirm the findings of the few studies that have demonstrated a direct link between feedback and goal commitment (Chong and Johnson 2007) or a direct link between feedback and trust (Bijlsma and de Bunt 2003; Hartmann and Slapni?ar 2009). Our results also show that trust is a partial mediating variable in the relationship between feedback and goal commitment, a finding that had not previously been established. 90The mediating role of trust suggests that we should consider the importance of interpersonal relationships when implementing results-based control systems. This method of control encourages interaction along the reporting line between each manager and his or her direct supervisor. The functioning of results-based control systems and, more specifically, as shown in this study, managers’ level of participation in setting their goals, and, to an even greater extent, the quality of feedback they receive from their supervisors, significantly influence managers’ perceptions of their supervisor on one important point: the level of trust they can place in them. This is an important point because, in the end, trust will contribute, at least in part, to increasing managers’ goal commitment. 91In order to understand how control systems affect managers’ attitudes and behaviors, in particular their goal commitment, it is thus important to examine the way these systems structure and affect relationships between managers and their direct supervisors. The «trust in supervisor» variable enables us to better understand the relationships along the hierarchical reporting line and, more specifically, how the introduction of results-based control systems affects these relationships. 92Conversely, contrary to our expectations, our study fails to identify a combined effect of participation and feedback on trust. This finding, not previously explored, is interesting in itself since it shows that participation and feedback significantly affect trust independently of one another. In other words, it seems that each of those control devices, in isolation, contributes to improving the quality of the relationship between a manager and his or her supervisor. If managers are able to develop a constructive dialogue with their supervisor at the target-setting stage, their trust in their supervisor will be strengthened even if they do not receive high-quality feedback at a later stage. Similarly, if supervisors provide managers with regular and high-quality feedback, a relationship of trust may develop even if managers are not able to significantly participate in setting their goals. 93In terms of practical consequences, our results show that managers’ goal commitment is highly influenced by the control procedures put in place, especially in terms of setting managers’ goals and providing them with feedback. Using a participative approach to set managers’ goals increases their goal commitment directly, as well as indirectly through the trust that develops between managers and their supervisors. In addition, providing managers with regular and good quality feedback on their performance strongly increases supervisory trust, which also contributes to increasing goal commitment. Companies should therefore use these two control tools and encourage the development of trust. This can be achieved in two ways. On the one hand, they must implement clear rules to ensure a genuine exchange between managers and their supervisors when setting managers’ objectives. Similarly, they should develop feedback systems providing managers with regular updates on their performance. On the other hand, companies can also work with supervisors, through communication and training, to encourage them to involve their subordinates in setting their objectives. In addition, they can raise supervisors’ awareness of the importance of the feedback they give to their subordinates to encourage them to take time to analyze results with their subordinates and to establish a fruitful dialogue. 94This study has a number of limitations that would benefit from further research. Firstly, questionnaires were collected from executives participating in training programs. This provided a high response rate and a large and diversified sample, which was important given our desire to generalize our research. However, this data collection process may have biased the responses, despite the fact that we neither selected the participants nor conducted research pertaining to the topics covered in the training seminars. Secondly, the cross-sectional survey method used in this study does not allow us to interpret correlations as causal effects. Our interpretations are theory-driven but other rationales could have been advanced. For instance, it may be that managers who trust their supervisors are more inclined to feel that they are involved in setting their goals and/or that they receive good quality feedback. Similarly, highly involved managers may request feedback from their supervisors. Thirdly, as with every survey, variables are based on perceptions. The use of perceptions is probably consistent and relevant for measuring attitudes such as trust, for instance; conversely, more objective measures could be used for participation or feedback. This limitation should be qualified, however, since managers’ commitment may be influenced by the perception that they participate rather than their actual participation. Fourthly, we did not attempt to control possible endogeneity problems with our explanatory variables. Participation and feedback may themselves be explained by other variables that are not included in our model but that could in fact be the real causes of our dependent variables. This problem, well known in econometrics, is receiving increasing attention in our field, despite the recommended solutions being more difficult to implement (Chenhall and Moers 2007; Larcker and Rusticus 2010). Finally, other variables could be introduced into the study, not only explanatory variables related to control systems such as the level of goal difficulty, but also mediating or moderating variables such as perceived justice or level of uncertainty. 95Despite these limitations, our study provides interesting evidence regarding the way in which managers’ participation in setting their objectives and the feedback that they receive strengthen their goal commitment, and, presumably, their resulting performance. AcknowledgementsThe authors would like to thank the CCA reviewers, as well as Didier Lemaître (Congrès AFC 2013), for their helpful suggestions, which helped improve this paper. Notes
References
What are characteristics of motivating goals according to Goal Setting Theory?Goal-setting theory is one of the most influential theories of motivation. In order to motivate employees, goals should be SMART (specific, measurable, aggressive, realistic, and time-bound).
Which is the best summary of goal setting theory?Which is the best summary of Goal-Setting Theory? Specific goals increase performance. Goals make no difference in performance.
What is considered the most integrative theory of motivation?The most well-known process theory of motivation is the reinforcement theory, which focused on the consequences of human behavior as a motivating factor.
What theories of motivation would be characterized as content perspectives quizlet?content perspectives include four theories: Maslow's hierarchy of needs theory. McClelland's acquired needs theory. Deci and Ryan's self-determination theory.
|
The use of feedback and participation in deciding how to achieve goals is part of which theory?
zusammenhängende Posts
Urheberrechte © © 2024 decemle Inc.
