Even though consensus exists among scholars about the importance of a committed workforce for the performance of an organisation, there is disagreement about its conceptualisation (Klein, Molloy & Brinsfield 2012; Meyer 2014). Roodt (2004) argues that the lack of conceptual clarity could be attributed to the haphazard way in which research on commitment evolved. Disagreement exists as to whether the experience of commitment can be described as cognitive (positive expectations of goals’ attractiveness and attainability), affective (positive emotions such as satisfaction with goals), volitional (wilful dedication to goals), conative (intentions to work hard to implement goals), or inclusive of all four psychological constructs (Klein et al. 2012; Roodt 2004). A study conducted by Van Lill, Roodt and De Bruin (2020) provides evidence that goal commitment could be viewed as a hierarchical model with a strong unidimensional factor, which consists of cognitive, emotional, volitional and conative dimensions. The general factor of commitment is theorised to be an intrinsically motivated bond with a goal (Klein et al. 2012), which influences the direction, intensity and persistence of subordinates’ efforts at work (Meyer 2014). Following prior studies, goal commitment was used as the benchmark in the present study to investigate the effectiveness with which different goal-setting styles motivate subordinates to pursue and implement goals (Piccolo & Buengler 2013). However, in contrast to prior studies, goal commitment was measured as a hierarchical factor which is inclusive of a cognitive, affective, volitional and conative factor.

House and Mitchell’s (1975) path-goal theory is the only framework that explicitly addresses the different styles that managers adopt during goal setting to motivate subordinates to pursue goals (Piccolo & Buengler 2013). Few, if any, alternative frameworks of managers’ goal-setting styles, with the necessary empirical support, since 1975 could be identified. An alternative framework that accounts for key evidence in the preceding 50 years of goal-setting literature might provide a new perspective and advance scientific literature (Locke & Latham 2019).

Two constructs measured by Latham, Erez and Locke (1988), namely participative goal setting and goal rationales, provide axes around which an alternative and coherent theory of managers’ goal-setting styles can be formulated. However, as portrayed in Figure 1, the constructs are reformulated in the present study to reflect universal dimensions on which subordinates form social impressions of managers (Cuddy et al. 2011; Rosenberg, Nelson & Vivekananthan 1968). Subsequently, goal rationale is reformulated as intellectual competence, which reflects managers’ ability to persuade others to pursue collective goals by using convincing rationales in support of the importance and practicality of the proposed goals (Cuddy et al. 2011; Latham et al. 1988; Rosenberg et al. 1968; Yukl, Kim & Chavez 1999). Participative goal setting, in turn, is reformulated as warmth, which refers to the degree to which managers are concerned about others’ well-being and their respectfulness towards alternative views or concerns raised while setting goals (Falbe & Yukl 1992; House & Mitchell 1975; Latham et al. 1988; Rooney & Gottlieb 2007). Intellectual carelessness and coldness anchor the intellectual competence and warmth dimensions on the opposite ends of the axes. Intellectual carelessness occurs when managers set goals impulsively, with little thought given to the importance or practical implications of goals (De Vries 2018; Shaw, Erickson & Harvey 2011). A perception of coldness is created by managers who are overly critical and discourage others from participating in goal setting (De Vries 2018; Rooney & Gottlieb 2007).

Four goal-setting styles are inferred from the two theoretical axes, which are argued to have weak or strong relationships with intellectual competence, warmth, intellectual carelessness and coldness depending on the constructs’ relative locations in Figure 1. The four styles inferred from the two axes are: deliberative, directive, complaisant and hostile.

Deliberative managers are perceived as persons in positions of authority who encourage democratic forms of goal setting, by thoughtfully engaging subordinates in critical discussions, to set goals that are meaningful and practical (Gastil 1994; Van Lill, Roodt & De Bruin 2019). The view that managers are intellectually competent could emerge when managers provide new insight about the meaningfulness or practical relevance of goals when deliberating with subordinates (Schneiderhan & Khan 2008; Van Lill et al. 2019). The impression that deliberative managers exchange relevant information, and invite and carefully consider alternative views, could simultaneously increase the sense that managers are warm (Redeker et al. 2014; Schneiderhan & Khan 2008; Van Lill et al. 2019).

Directive managers are perceived by subordinates as forceful persons in positions of authority who set goals with assertiveness, conviction and a determination to achieve set standards and follow procedures (Redeker et al. 2014; Schriesheim & Kerr 1974). Anderson and Kilduff (2009) posit that dominance is strongly associated with indicators of intellectual competence, such as the ability to persuade others. However, subordinates who perceive their managers as decisive and unbending in convictions about assigned goals might also believe that subordinates’ views and concerns are disregarded (Gagné & Deci 2005; Rooney & Gottlieb 2007), reducing managers’ perceived warmth (Cuddy et al. 2011; Redeker et al. 2014).

Subordinates who perceive managers as complaisant might believe that managers are willing to compromise organisational goals to accommodate subordinates (Einarsen, Aasland & Skogstad 2007) and, consequently, the style is associated with greater warmth (Cuddy et al. 2011; Redeker et al. 2014). A perception of managers as complaisant might, however, be associated with an ineffectiveness to rise above individual concerns and set broader strategic goals (Einarsen et al. 2007). The perception that managers are less attentive to broader strategic matters might lead subordinates to believe that managers are intellectually careless (Einarsen et al. 2007; Van Lill et al. 2019).

Hostile managers are, in subordinates’ view, destructive persons in positions of authority who actively undermine subordinates’ attempts to reason and misrepresent facts to support assigned goals (Duffy, Ganster & Pagon 2002; Rooney & Gottlieb 2007; Shaw et al. 2011). Perception of managers as individuals who ridicule subordinates’ views while misrepresenting facts to persuade others might cause subordinates to believe that these managers are higher in intellectual carelessness and coldness (Cuddy et al. 2011; De Vries 2018; Rooney & Gottlieb 2007).

Each goal-setting style is hypothesised to have a different relationship with goal commitment depending on its relative location in Figure 1. However, this relationship is argued to be dependent on an additional social layer of interpretation that subordinates use to make sense of their exchanges with managers when setting goals (Cropanzano et al. 2017), namely supervisor-focused interactional justice. Supervisor-focused interactional justice refers to the perceived fairness of interpersonal treatment received from managers (Bies 2015), which in the current study is argued to be influenced by subordinates’ perceptions of managers’ relative standing on intellectual competence and warmth during goal setting. Greenberg (2009) differentiates between two concepts of interactional justice, namely interpersonal and informational justice. When managers provide clear rationales for goals (intellectual competence), subordinates are theorised to experience a deeper sense of being treated with sincerity and honesty and, therefore, attribute supervisor-focused informational justice to the goal-setting style (Bies 2015; Greenberg 2009). When managers are perceived to be considerate of alternative views or concerns about goals (warmth), subordinates are theorised to experience a greater sense of being treated with dignity and respect (Bies 2015; Greenberg 2009) and, therefore, attribute supervisor-focused interpersonal justice to the goal-setting style.

The respondents were 451 working adults aged 18 and older who had at least a Grade 12 education, a satisfactory English reading ability and who reported to a manager in their work setting. To achieve generalisability of results, 17 organisations across various sectors and industries were targeted for participation in an online or paper-based survey. Respondents were drawn from six participating organisations that represented the industries of agriculture, finance, manufacturing, science, human health and transport. A total of 77% of the respondents were non-managerial employees, while 23% were managers; 36% had a university qualification and 64% held Grade 12 (matric) or a diploma.

A five-point intensity rating scale was employed for all the items (Zikmund et al. 2010). Verbal anchors were used to define extreme points at both ends of the scale (Zikmund et al. 2010).

Access to organisations in Gauteng was obtained by approaching executives or human resource practitioners within the organisations. The questionnaires were administered using an online platform or a paper-and-pencil version of the survey. To prevent non-responses, which posed a threat to the external validity of the survey administrated using the online platform, frequent reminders were sent to encourage participation. To prevent missing responses, which posed a threat to the external validity of the paper-and-pencil survey, respondents’ questionnaires were inspected upon submission.

The study was low in risk, but precautions were taken to ensure that participation was voluntary and anonymous, that no harm was caused, that the questions were completed truthfully and that informed consent had been given to use the results for purposes of research. Ethical approval for the research project was obtained from the Research Ethics Committee, College of Business and Economics, University of Johannesburg. FOM2016-IPPM02, 14 April 2016.

As a preliminary step in investigating each research hypothesis, the multivariate skewness and kurtosis, inter-factor correlations, factor regressions, mean, standard deviation and inter-item reliabilities of each scale were inspected. Inter-factor correlations and factor regressions were inspected by conducting structural equation modelling (SEM) with robust maximum likelihood estimation (MLM) (Li 2016). Fit, based on absolute indices, was considered suitable if the root mean square error of approximation (RMSEA) and standardised root mean residual (SRMR) were 0.08 or lower (Brown 2006; Browne & Cudeck 1992). Fit, based on comparative indices, was considered suitable if the comparative fit index (CFI) and Tucker–Lewis index (TLI) were 0.90 or higher (Brown 2006; Hu & Bentler 1999).

A parallel multiple mediation model was inspected by conducting SEM using version 0.6-4 of the lavaan package in R (Rosseel 2012; Rosseel & Jorgensen 2018) to examine H_1A, H_2A, H_3A, H_4A, H_1B, H_2B, H_3B and H_4B. Model parameters were obtained by conducting SEM with ML and Bollen-Stine bootstrapping with 5000 bootstrap resamples (Bollen & Stine 1992; Finney & DiStefano 2006; Hayes 2018). Hayes’s (2018) guidelines were followed to specify the paths between the latent variables, and to calculate the direct, indirect, indirect total and total effects. As recommended by Baron and Kenny (1986), the products of paths relevant to H_1B to H_4B (a*b) were calculated to inspect the hypothesised multiple mediator model. The indirect effects were interpreted in line with the squared values in Cohen’s (1988) guidelines (Kenny 2018; Shrout & Bolger 2002). An effect size of 0.01 to 0.09 was interpreted as small, 0.09 to 0.25 was interpreted as moderate, and 0.25 to 1 was interpreted as large (Cohen 1988; Kenny 2018; Shrout & Bolger 2002).

Mardia’s multivariate skewness and kurtosis coefficients for the entire set of 72 items (manifest variables) were 99 332.88 (p < 0.05) and 117.73 (p < 0.05), which provided justification for the use of more robust estimation techniques. The standardised inter-factor correlations are reported in Table 1, with the mean, standard deviation and internal consistency reliability of each variable in the last four rows. The models specified to determine the inter-factor correlations yielded satisfactory fit (χ² [df] = 1567.44 [835]; CFI = 0.94; TLI = 0.94; SRMR = 0.06; RMSEA = 0.05).

The internal consistency reliability of each of the scales in Table 1 was satisfactory (α and ω ≥ 0.71). The size of the correlations between the factors pointed towards the order in which the styles were expected to be related to informational justice, interpersonal justice and goal commitment. The deliberative style had the highest positive correlation with informational justice, interpersonal justice and goal commitment. The directive style also had a positive relationship with informational justice, interpersonal justice and goal commitment, albeit smaller than the deliberative style. Even though the complaisant style had positive relationships with informational justice, interpersonal justice and goal commitment, the size of these relationships was small. Per the predictive hypotheses, the hostile style had negative relationships with informational justice, interpersonal justice and goal commitment. The inter-factor correlations merited a further investigation into the predictive validity of different goal-setting styles for goal commitment.

The central aim of the current study is to investigate the relationship between managers’ goal-setting styles and subordinates’ goal commitment. Evidence in support of the cognitive schema that subordinates use to make sense of different goal-setting styles, as proposed in Figure 1, assists in validating the framework used to explain the relationships between goal-setting styles and goal commitment. The relative standing of goal-setting styles was tested by regressing the four goal-setting styles on four different social perceptions, namely intellectual competence, warmth, intellectual carelessness and coldness, simultaneously. The standardised factor regressions are reported in Table 2, with the mean and standard deviation of the social perceptions in the last two rows. The model specified to determine the factor regressions (χ² [df] = 1602.29 [961]; CFI = 0.96; TLI = 0.95; SRMR = 0.05; RMSEA = 0.05) yielded satisfactory fit. The internal consistency reliability coefficients of the social perception scales are also provided in the last two rows of Table 2.

The internal consistency reliability of each of the scales in Table 2 was satisfactory (α and ω ≥ 0.93). The factor regressions indicate that the deliberative style had the strongest positive relationship with intellectual competence and warmth. The directive style had a small positive relationship with intellectual competence but was not positively related to coldness. Evidence, at this juncture, suggests that directive managers might simply be unsuccessful at increasing perceptions of warmth among subordinates. A small positive relationship was found between the complaisant style and intellectual carelessness and warmth. The hostile style had a positive relationship with intellectual carelessness and coldness. Most of the styles’ associations with social perceptions, apart from the directive style’s relationship with coldness, confirmed the relative location of goal-setting styles in Figure 1. Given the evidence in support of the location of goal-setting styles in Figure 1, the next step was to test the predictive hypotheses.

Before the relationships in H₁ to H₄ could be inspected, the fit statistics of the parallel multiple mediator model had to be evaluated. The fit statistics for the model (χ² [df] = 2252.66 [836]; CFI = 0.92; TLI = 0.92; SRMR = 0.07; RMSEA = 0.06) were satisfactory. Two alternative models to Figure 2 were specified, to inspect the viability of other explanations for the relationship between the observed and latent variables. Firstly, due to the high inter-factor correlations between the deliberative style, the directive style, interpersonal justice and informational justice (see Table 1), the measurement variables of all the factors were loaded on a unidimensional factor called the constructive approach to goal setting. The direct relationship between the constructive approach, hostile style, as well as the complaisant style, and goal commitment was, in turn, determined. However, the fit statistics for the model (χ² [df] = 4508.16 [850]; CFI = 0.80; TLI = 0.79; SRMR = 0.08; RMSEA = 0.10) were unsatisfactory. Secondly, the mediation of interactional justice (a unidimensional factor containing items from both informational and interpersonal justice) on the relationship between goal-setting styles and goal commitment was specified. This model also did not yield satisfactory fit with the observed data (χ² [df] = 3015.43 [841]; CFI = 0.88; TLI = 0.87; SRMR = 0.07; RMSEA = 0.08). After establishing model fit, path coefficients were inspected to obtain a detailed understanding of the coefficients on which direct and indirect relationships between goal-setting styles and goal commitment were calculated, which are reported in Figure 2.

The results of the study suggest that subordinates’ commitment to goals is influenced by their unique social perceptions of their managers, based on two dimensions of social cognition, namely the degree of warmth (increasing self-determination, positive emotions about and an internal dedication to goals) and intellectual competence (increasing positive expectations and intentions to implement goals) associated with managers’ goal-setting styles (Cuddy et al. 2011; Van Lill et al. 2019). As a result, each style’s relationship with goal commitment depends on its relative location to intellectual competence, warmth, intellectual carelessness and coldness in Figure 1. New insights on the social dynamics between managers and subordinates when setting goals is, therefore, obtained by integrating the theory of goal setting with workplace deliberative democracy and social cognition.

The present study further provides evidence that perceptions of managers’ styles as informationally fair, more so than interpersonally fair, has important implications for subordinates’ commitment to goals. The findings shed a unique light on an additional social perceptual layer that could be incorporated into future studies to make sense of democratic goal setting, namely supervisor-focused informational justice.

The dimensions outlined in this section could be used by managers to actively seek structured feedback from subordinates about their goal-setting styles. If managers can identify early signs of the negative impact of their goal-setting styles by asking subordinates what they think of and feel about goals set, including taking note of expressions of subordinates’ dedication to and intentions to implement goals, managers could engage their teams by using more effective goal-setting styles that are likely to ensure subordinates’ sustained effort on important and practically relevant collective goals. However, managers should also be cognisant that subordinates evaluate managers’ goal-setting styles through the lenses of supervisor-focused informational justice in the workplace. Goal-setting styles that are more strongly associated with the rules of informational justice (honesty and thoroughness) may have a greater motivational impact on subordinates’ commitment to goals set.

A self-report survey was administered to respondents, to ensure an efficient measure of the relationships between subjective variables at a given time, which may be perceived as a weakness in terms of common method variance (Spector 2019). However, given the centrality of subordinates’ social perceptions in the present study, it would not have made sense to infer their social cognition from other stakeholders’ (managers, peers, subordinates or clients) points of view (Podsakoff et al. 2012; Spector 2019). Furthermore, a measurement of subordinates’ impressions of managers’ goal-setting styles and goal commitment at different time intervals might have introduced extraneous variables into the design (Spector 2019). To reduce method bias in the study under discussion, apprehension was reduced by protecting respondents’ anonymity (cf. Podsakoff et al. 2003). The ambiguity of items was further investigated by conducting an item-sort exercise with subject matter experts (cf. Podsakoff et al. 2012; Van Lill 2019). Items with low substantive validity were subsequently adjusted or removed from the measures (Howard & Melloy 2016; Van Lill 2019).

An investigation of the consistency of the effects of goal-setting styles on goal commitment across goal conditions might provide invaluable information about situational variables that could influence the relationships reported. For example, the SEM could be nested in different types of goals (i.e. learning or performance goals) or managerial levels. Learning goals might elicit more interest and enjoyment than performance goals from subordinates and, as a result, lead to greater goal commitment, irrespective of managers’ goal-setting styles. The importance of goals set at higher levels of management might accentuate the need for greater deliberation to get commitment when compared to lower levels of management. To investigate the effects of situational variables, a multilevel SEM model that nests the effects in different goal conditions could be investigated (RabeHesketh, Skrondal & Zheng 2012). According to Stone-Romero and Rosopa (2008), erroneous inferences can be made about causal effects based on non-experimental research. The study under discussion relied on theory, as well as the inspection of alternative SEM models, to determine if supervisor-focused informational and interpersonal justice mediate the relationship between goal-setting styles and goal commitment. However, it is strongly recommended that future experimental studies be designed to determine whether the findings of this study are replicable.