Abstract
The aim of the present study is to provide a deeper understanding of how autonomy-supportive leadership affects volunteer motivation by taking volunteers’ individual differences into account. For this purpose, self-determination theory (SDT) was utilized because this conceptual framework considers both the social context (i.e., autonomy-supportive leadership) and individual differences (i.e., causality orientations) as antecedents of motivation. Causality orientations alter the way individuals perceive their social context as either autonomy-supportive or controlling (i.e., autonomy orientation or control orientation, respectively). Therefore, it is hypothesized that both types of causality orientations serve as moderators of the relationship between autonomy-supportive leadership and volunteer motivation. The hypotheses were tested on N = 1,979 volunteers. The results revealed that the relationship between autonomy-supportive leadership and volunteer motivation varied as a function of the strength of autonomy and control orientation. The importance of the moderating role of individual differences on volunteer motivation is discussed.
Keywords
A European Union study on volunteering revealed that 92 to 94 million adults are involved in volunteering. This figure indicates that about 22% to 23% of the European population (aged above 15 years) engages in volunteering (Study on Volunteering in the European Union, 2010). In most European countries, the number of active volunteers has increased over the past decade. Voluntary work is of great importance because of its impacts on and benefits for local communities, the society in general, and for volunteers themselves. To maintain these services for the community and society, it is of considerable practical interest for volunteering organizations to know how to retain their volunteers for long-term engagements, especially as volunteer turnover can be extremely costly and time consuming in terms of decreased organizational morale, volunteer replacement costs (recruitment and training), and decreased ability to serve the public (McElroy, Morrow, & Rude, 2001; Musick & Wilson, 2008). This knowledge is likely to become even more pertinent because of the shift from traditional volunteering, with a lifelong commitment, to a new type of sporadic and episodic volunteering, with a decreased commitment to any particular organization (Vantilborgh et al., 2011; Wilson, 2012). Therefore, it has become even more challenging to motivate volunteers to stay within an organization for continuing and long-term engagements.
When nonprofit organizations try to establish long-term relationships with their volunteers, they should be aware that feelings of pressure, deadlines for work completion, or external control have been detected as harmful for a continuing engagement (cf. Amabile, DeJong, & Lepper, 1976; Grube & Piliavin, 2000; Musick & Wilson, 2008; Stukas, Snyder, & Clary, 1999). The study of Stukas et al. (1999) on “mandatory volunteerism” programs for students illustrated these findings. The requirement to volunteer to graduate—a strong perception of external control—decreased future intentions to volunteer. Consequently, volunteers’ experience of autonomy—as opposed to control and pressure—appeared to be especially beneficial for a long-term engagement (Gagné, 2003; Sobus, 1995).
As volunteering usually takes place within an organization (Grube & Piliavin, 2000), the organizational context could well play an important role in inducing a feeling of autonomy through giving volunteers choice and encouragement for personal initiative (Gagné, 2003). Volunteer coordinators, as part of the organizational context, play a major role in inducing feelings of autonomy or control, as they are directly involved in activities to attract, orient, retain, and organize the volunteers in their charge (Studer & von Schnurbein, 2012). Volunteers stated that coordinators who enable them to act autonomously (e.g., offering opportunities for personal initiative) and who were not too controlling (e.g., establishing tight rules on work schedules) greatly increased their motivation to volunteer (Leonard, Onyx, & Hayward-Brown, 2004; Musick & Wilson, 2008; Vantilborgh et al., 2012). Research has shown initial evidence that inducing a feeling of acting autonomously through volunteer management has a positive effect on volunteering, whereas a feeling of being controlled or pressured by the volunteer coordinator (i.e., a feeling of being micromanaged) has the opposite effect, although further consideration on this topic is needed (Vantilborgh et al., 2011). Consequently, the aim of the present article is to provide expanded insight into the impact of autonomy-supportive management practices on volunteer motivation.
To study volunteer motivation, the present study utilizes self-determination theory (SDT; Deci & Ryan, 2000; Deci & Ryan, 2008b) as a theoretical framework. The focus of SDT lies in the autonomous versus controlled “quality” of motivation (Deci & Ryan, 2000). Another popular theoretical approach dealing with the question of volunteer motivation is the functional approach (cf. Clary et al., 1998), which focuses on motives underlying volunteering (e.g., social, career). SDT was chosen for this study as its distinction between autonomous and controlled motivation offers the opportunity to gain a deeper understanding of the quality of volunteer motivation and the occurrence of feelings of autonomy or control through management practices. Additionally, SDT proposes the social environment (i.e., managerial autonomy support) as a main antecedent of motivation—making it possible to have a closer look at the relationship between management practices and volunteer motivation within an established theoretical framework.
Furthermore, SDT suggests that not only the social environment must be taken into account when examining motivation, but also the way in which people differ in their interpretations of the social environment. Therefore, SDT considers individual differences (i.e., causality orientations) as a secondary antecedent of motivation (Gagné & Deci, 2005). The conceptualization of individual differences within SDT is especially interesting as it reflects a personality trait that alters the way individuals perceive their social context—and thus their supervisor—as autonomy supportive or controlling (Deci & Ryan, 2000). In more detail, the present study investigates the main and moderating role of autonomy-supportive leadership and volunteers’ individual differences on motivation.
SDT
SDT is a meta-theory on human motivation (Deci & Ryan, 2000). Central to the conceptualization of SDT is the distinction between autonomous motivation and controlled motivation (Deci & Ryan, 2008a). Autonomously motivated behavior is experienced as “out of free will,” with a full sense of volition and choice. Autonomously motivated volunteers, who act based on interest and fun or because of an underlying personal value, reported higher job satisfaction (Millette & Gagné, 2008) and greater work effort (Bidee et al., 2013). In contrast, controlled motivated behavior involves a feeling of external pressure and of having to engage in an activity or to avoid punishment or feelings of guilt. Volunteering is negatively affected when individuals feel obliged by external contingencies and thus act based on controlled motivation (Gagné, 2003). Both types of motivation consist of a combination of behavioral regulations that differ in the degree of internalization—along a continuum from intrinsic motivation to external regulation (Deci & Ryan, 2000). Autonomous motivation consists of intrinsic motivation, integrated regulation, and identified regulation. Intrinsically motivated volunteers act based on self-determination and engage in an activity because they find it interesting and enjoyable. Identified regulated volunteers act based on identification with the value of a specific volunteering activity. For example, these volunteers identified with the importance of visiting lonesome people in a hospital. Integrated regulated volunteers also identified with the importance of visiting lonesome people and additionally integrated those identifications with other aspects of the self (Deci & Ryan, 2000). In the organizational context, integrated regulation is sometimes excluded from autonomous motivation as it is psychometrically difficult to separate from identified regulation (Gagné, Chemolli, Forest, & Koestner, 2008; Vallerand et al., 1992).
Controlled motivation includes external regulation and introjected regulation. Externally regulated individuals volunteer fully based on external contingencies (e.g., to attain desired consequences or to avoid punishment; Deci & Ryan, 2000). To illustrate, volunteers who act based on external regulation would visit lonesome people in hospital merely to gain recognition from their friends, family, or supervisor. Introjection represents a partial internalization of regulations that have not yet become part of the self (e.g., engaging to protect self-worth or due to threats of guilt and shame; Deci & Ryan, 2000). That is, volunteers who act based on introjected regulation volunteer because otherwise they would feel guilty not having visited the elderly woman in hospital whom they regularly visit once a week.
Autonomous and controlled motivation both reflect a person’s intention to act; whereas amotivation reflects a lack of intention to act (Deci & Ryan, 2008a). Similar to previous studies in the volunteering context (Bidee et al., 2013; Gagné et al., 2010; Haivas, Hofmans, & Pepermans, 2012; Millette & Gagné, 2008), this study omitted amotivation because of its focus on active and intentional types of motivation.
Autonomy-Supportive Leadership
In the organizational context, SDT proposes managerial autonomy support as main antecedent of motivation (Gagné & Deci, 2005). Autonomy-supportive leadership can be characterized as understanding and acknowledging subordinates’ perspectives, giving opportunities for choice, supporting individuals’ competences, and encouraging personal initiative (Deci et al., 2001; Gagné, 2003). In the volunteering context, autonomy-supportive leadership refers to the interpersonal climate created between a supervisor and a volunteer (Gagné & Deci, 2005). Autonomy-supportive leadership affects both types of motivation (Gagné, 2003; Gagné & Deci, 2005). Broad empirical evidence exists for the positive impact of autonomy-supportive leadership on autonomous motivation (Black & Deci, 2000; Deci & Ryan, 1987; Gagné & Deci, 2005; Haivas et al., 2012; Williams & Deci, 1996). However, the results concerning the impact of autonomy-supportive leadership on controlled motivation have been less consistent. SDT presumes a negative relationship between managerial autonomy support and controlled forms of motivation because an autonomy-supportive context promotes internalization of extrinsic motivation (Gagné & Deci, 2005). Some studies have reported this theoretically expected negative relationship between autonomy support and controlled motivation. For example, negative correlations were found in a study examining the impact of therapists’ autonomy support on therapy-motivation (Pelletier, Tuson, & Haddad, 1997) and in a study of young athletes’ controlled motivation through parental autonomy support (Gagné, Ryan, & Bargmann, 2003). However, most of the studies to date have reported correlations close to zero for the relationship between autonomy support and controlled motivation or controlled forms (regulations) of extrinsic motivation (cf. Balaguer, Castillo, & Duda, 2008; Chan, Lonsdale, Ho, Yung, & Chan, 2009; Hagger, Chatzisarantis, Culverhouse, & Biddle, 2003; Williams, Grow, Freedman, Ryan, & Deci, 1996; Wilson & Rodgers, 2004; Zeldman, Ryan, & Fiscella, 2004). For example, a study by Chan et al. (2009) examined physiotherapists’ autonomy-supportive behavior on patient motivation and a study by Williams et al. (1996) investigated autonomy support as a motivational predictor of weight loss. In addition, Zeldman et al. (2004) tested the role of autonomy support and motivation in a methadone maintenance treatment. None of these studies found a significant correlation between autonomy support and controlled motivation. A recent study in the volunteering context even found a positive relationship between autonomy-supportive leadership and controlled motivation (Haivas et al., 2012).
Consequently, the research to date suggests that autonomy support can have increasing as well as decreasing effects on controlled motivation. This leads to the assumption that a moderator variable influences the relationship between autonomy-supportive leadership and controlled motivation. As SDT considers individual differences reflecting personality traits that alter the way individuals perceive their supervisor, examining the effect of autonomy-supportive leadership in relative isolation from individual differences could provide an incomplete picture of volunteer motivation. Therefore, the present study examines the moderating role of individual differences in the relationship between autonomy-supportive leadership and volunteer motivation.
Causality Orientations
According to SDT, individuals differ in their general orientation toward the regulation of their behavior (Gagné & Deci, 2005). This so-called locus of causality (Deci & Ryan, 1985) is defined as a stable and trait-like tendency to generally experience social contexts as autonomy supportive (termed as autonomy orientation) or controlling (termed as control orientation; Gagné & Deci, 2005). As the term “locus of causality” is frequently confused with “locus of control” (Rotter, 1966), it is important to note that the terms differ: “Locus of control refers to whether people believe that outcomes are controllable,” whereas locus of causality “refers to the perceived source of initiation and regulation of behavior” (Deci & Ryan, 1985, p. 113).
Individuals with an autonomy-oriented personality tend to interpret social contexts as autonomy supportive and external events (e.g., rewards) as informational and supportive of their self-determination (Deci & Ryan, 1985). Therefore, the behavior of individuals with an autonomy orientation is regulated based on interests and self-endorsed values (Deci & Ryan, 2000). An autonomy-oriented personality has been found to be positively related to autonomous motivation (Kwan, Hooper, Magnan, & Bryan, 2011; Williams & Deci, 1996).
In contrast, individuals with a control-oriented personality tend to perceive external events as pressuring and controlling (Deci & Ryan, 1985); they interpret events as policing (Soenens, Berzonsky, Vansteenkiste, Beyers, & Goossens, 2005) and chronically perceive pressure from the environment (Rost, Weibel, & Osterloh, 2010). Control orientation has been found to be related to controlled forms of motivation (Kwan et al., 2011; Williams & Deci, 1996), and a feeling of being driven by others’ expectations (Deponte, 2004).
Both causality orientations are seen as conceptually distinct traits and not as a continuum from autonomy orientation to control orientation (Deci & Ryan, 1985). The General Causality Orientation Scale (GCOS) was constructed to measure the level of each orientation within a person and not to classify a person as having one or the other of the orientations (Deci & Ryan, 1985). Empirical work has revealed this conceptual independence, as autonomy orientation is unrelated to control orientation (r = .03, not significant; Deci & Ryan, 1985). Therefore, individuals who score low in control orientation are not automatically ranked high in autonomy orientation. In addition to these two types of causality orientations, a third orientation (impersonal orientation) reflects the general tendency to be amotivated (Gagné & Deci, 2005) and not to engage in any action (Gagné, 2003). However, as the focus of this study was not on amotivation, the impact of the impersonal orientation on volunteer motivation was not examined.
The Moderating Role of Causality Orientations
To date, most studies have examined the effect of autonomy-supportive leadership and individual differences in relative isolation from each other (Gagné & Deci, 2005; Vansteenkiste, Niemiec, & Soenens, 2010). Following the call for the simultaneous examination of both social context and personality (Gagné & Deci, 2005; Lam & Gurland, 2008; Vansteenkiste et al., 2010), the present study investigates the moderating role of volunteers’ personality on the relationship between autonomy-supportive leadership and volunteer motivation. Thereby, the proposition of SDT that causality orientations alter the way individuals perceive the characteristics of the organizational context is followed. Some researchers (cf. Vansteenkiste et al., 2010) have interpreted causality orientations differently, arguing that autonomy-oriented individuals would especially benefit from an autonomy-supportive context, whereas control-oriented individuals would benefit from a controlling context. This approach is called the “match-perspective” (Vansteenkiste et al., 2010). However, Deci and Ryan’s (1985) conceptualization of causality orientations is different in that causality orientations are understood as trait-like tendencies to interpret a social context as autonomy supportive or controlling. The present study and the following moderation hypotheses explicitly rely on this conceptualization.
Moderating Role of Autonomy Orientation
Prior to stating the moderation hypothesis concerning autonomy orientation, a summary of what is understood at this point from previous research is presented. First, autonomy-supportive leadership has a positive impact on volunteers’ autonomous motivation (cf. Gagné & Deci, 2005; Haivas et al., 2012). Second, an autonomy-oriented personality—and the tendency to perceive autonomy support in a social context—also has positive impacts on autonomous motivation (Kwan et al., 2011; Williams & Deci, 1996). Concerning the moderation hypothesis used in this study, it is expected that the relationship between autonomy-supportive leadership and volunteers’ autonomous motivation depends on the strength of volunteers’ autonomy orientation (personality). Individuals with a “strong” autonomy orientation generally tend to derive autonomy support from social contexts—regardless of whether the context actually is autonomy supportive or not—resulting in high autonomous motivation. Accordingly, the relationship between autonomy-supportive leadership and autonomous motivation is expected to be close to zero for volunteers with a strong autonomy-oriented personality. In contrast, individuals with a “weak” autonomy orientation do not generally derive autonomy support from social contexts. Consequently, the relationship between autonomy-supportive leadership and autonomous motivation is expected to be positive, as the more autonomy supported they feel, the more autonomously motivated they are.
Therefore, the following hypotheses present the expected main and moderating effects of autonomy orientation on the relationship between autonomy-supportive leadership and autonomous motivation:
Hypothesis 1a (H1a): Autonomy-supportive leadership has a positive main effect on autonomous motivation.
Hypothesis 2a (H2a): An autonomy-oriented personality has a positive main effect on autonomous motivation.
Hypothesis 3a (H3a): Autonomy orientation moderates the relationship between autonomy-supportive leadership and autonomous motivation. The relationship between autonomy-supportive leadership and autonomous motivation is close to zero for volunteers with a strong autonomy orientation and positive for volunteers with a weak autonomy orientation.
Moderating Role of Control Orientation
The moderation hypothesis on control orientation is based on the following concluding points from previous research and theoretical considerations. First, based on the theoretical assumptions of SDT, a negative relationship between autonomy-supportive leadership and controlled motivation is expected. Higher autonomy-supportive leadership should result in less controlled motivation (Gagné & Deci, 2005). Second, individuals with a control-oriented personality tend to perceive their environment as controlling, which leads to higher controlled motivation. Analogous to autonomy orientation, it is expected that the relationship between autonomy-supportive leadership and volunteers’ controlled motivation depends on the strength of volunteers’ control orientation (personality). Individuals with a “strong” control orientation tend to perceive external pressure and control out of any social context (Deci & Ryan, 1987)—even from autonomy-supportive leadership—resulting in high controlled motivation. Accordingly, the relationship between autonomy-supportive leadership and controlled motivation is expected to be close to zero for volunteers with a strong control-oriented personality. In contrast, individuals with a “weak” control orientation tend to perceive external pressure to a much lesser extent. Consequently, the relationship between autonomy-supportive leadership and controlled motivation is expected to be negative, as the more autonomy supported they feel, the less controlled motivated they are.
Therefore, the following hypotheses present the expected main and moderating effects of control orientation on the relationship between autonomy-supportive leadership and controlled motivation:
Hypothesis 1b (H1b): Autonomy-supportive leadership has a negative but weak main effect on controlled motivation.
Hypothesis 2b (H2b): A control-oriented personality has a positive main effect on controlled motivation.
Hypothesis 3b (H3b): Control orientation moderates the relationship between autonomy-supportive leadership and controlled motivation. The relationship between autonomy-supportive leadership and controlled motivation is close to zero for volunteers with a strong control orientation and negative for volunteers with a weak control orientation.
To conclude, the present study investigates the main and moderating roles of autonomy-supportive leadership and volunteers’ causality orientations on autonomous and controlled motivation, which has not been examined in the volunteering context to date.
Method
Participants and Procedure
Participants were recruited through charitable volunteering organizations in the German-speaking part of Switzerland. A questionnaire was sent to each volunteer via the organizations’ mail distribution with a link printed on the front page to offer the possibility to take part online. As an incentive, a small donation was given to the organizations for each completed questionnaire. Participation was completely voluntary. Prior to data collection, semistructured interviews with volunteer coordinators were conducted to ensure that all volunteers had a supervisor with whom they had contact several times a year. Additionally, participants were instructed to leave out questions that were not appropriate to their volunteering situation or tasks. Overall, 2,017 volunteers responded to the survey (response rate: 41%). Participants either filled out the paper–pencil questionnaire or the web-based questionnaire. The questionnaires were identical in content and structure. Five participants who filled out the online questionnaire twice and a further 33 participants who indicated that they do not volunteer regularly were eliminated. The final sample consisted of 1,979 volunteers (59.5% females; 3.3% did not indicate gender); the age varied between 17 and 87 years (M = 64.08, SD = 11.76). On average, the participants volunteered for 3.28 hr per week (SD = 2.43) and for a length of M = 5.62 years (SD = 4.95). Participants were engaged in 26 different charitable projects (e.g., visitation services, driving services, trust services, and migrant support).
Measures
Work Climate Questionnaire (WCQ)
Autonomy support by one’s supervisor was measured using the six-item short version of the WCQ (Baard, Deci, & Ryan, 2004). The scale was originally designed to measure managers’ autonomy support in paid work settings. The items were adapted to the volunteering setting by replacing the word “manager” by “supervisor,” as proposed by Baard et al. (2004). The scale has been used in the volunteering context before, to measure autonomy-supportive leadership (cf. Gagné, 2003; Haivas et al., 2012). All items refer to explicit behaviors of a supervisor. An item example is: “I feel that my supervisor provides me choices and options.” Participants rated each item on a scale from 1 (strongly disagree) to 5 (strongly agree). The internal consistency was good (α = .88).
GCOS
Causality orientations are typically measured using the GCOS (Deci & Ryan, 1985). Respondents read 12 vignettes describing hypothetical situations (e.g., “You have a school-age daughter. On parents’ night the teacher tells you that your daughter is doing poorly and doesn’t seem involved in the work. You are likely to . . . ”), each followed by two possible reactions. One item reflects the autonomy orientation (e.g., “Talk it over with your daughter to understand further what the problem is”) and one reflects the control orientation (e.g., “Make sure she does the assignments, because she should be working harder”). Participants rated the likelihood of having each of two possible reactions depicting autonomy and control orientation to each scenario on a scale from 1 (very unlikely) to 5 (very likely). The third possible reaction that represents impersonal orientation was excluded as it reflects a general tendency to be amotivated, which was not examined in the present study. The internal consistency was acceptable, with α = .68 for autonomy orientation, and α = .74 for control orientation. The internal consistency of both GCOS scales in this study was comparable to the internal consistency in previous studies (Deci & Ryan, 1985; Scherhorn, Hass, Hellenthal, & Seibold, 2010).
Motivation at Work Scale (MAWS)
Respondents’ motivation for volunteering was measured using the revised MAWS (Gagné et al., 2012). As it is very difficult to psychometrically separate integrated regulation from identified regulation in the organizational context (Gagné et al., 2008), the final version of the revised MAWS does not contain items for integrated regulation; consequently, the questionnaire contains eight items for autonomous motivation (intrinsic and identified regulation: “I volunteer because I have fun doing my task”) and eight items for controlled motivation (introjected and external regulation: “I volunteer because I want to avoid disappointing others, for example, my supervisor, clients, family”). Participants indicated on a scale from 1 (strongly disagree) to 5 (strongly agree) why they engage in their volunteering activity. Autonomous and controlled motivation had good internal consistency, with α = .86 and α = .88, respectively.
Results
Descriptive statistics and bivariate correlations for all variables are presented in Table 1. The correlations between volunteers’ causality orientations and motivation showed that autonomy orientation and autonomous motivation (r = .22, p < .001) were positively related, as well as control orientation and controlled motivation (r = .28, p < .001). Consistent with previous results (Kwan et al., 2011; Williams & Deci, 1996), no significant correlation was detected between autonomy orientation and controlled motivation (r = .00, p = .97), or between control orientation and autonomous motivation (r = .02, p = .30). In line with the theoretical assumptions (Deci & Ryan, 1985), both causality orientations could be seen as distinct (r = .09, p < .001). It could also be observed that the average values of volunteers’ autonomy orientation and autonomous motivation were higher than their control orientation and controlled motivation.
Means, Standard Deviations, and Correlations of All Variables.
Note: WCQ = Work Climate Questionnaire measuring autonomy-supportive leadership. GCOS = General Causality Orientation Scale measuring causality orientations (personality). Correlations higher than ±.07 are significant at p < .05.
Hierarchical regression analyses were used to test the assumed main effects and both moderation hypotheses. All predictor variables were centered prior to calculating the interaction term to reduce potential multicollinearity (Aiken & West, 1991). Significant interactions were further analyzed, calculating simple-slope tests and regions of significance (Preacher, Curran, & Bauer, 2006).
Autonomy Orientation as Moderator
The first hierarchical regression analysis tested whether volunteers’ autonomy orientation moderated the relationship between autonomy-supportive leadership and autonomous motivation. In step 1, to examine the main effects, autonomous motivation was regressed on autonomy support and autonomy orientation. The main effect of autonomy-supportive leadership was positive, β = .29, t(1,727) = 12.67, p < .001, as well as the main effect of autonomy orientation (personality), β = .16, t(1,727) = 6.93, p < .001 (see Table 2), showing that both autonomy support and autonomy orientation positively predicted autonomous motivation. Therefore, Hypotheses H1a and H1b were supported. In step 2, the interaction term was entered to test for moderation. Adding the interaction term resulted in a significant increase in explained variance, ΔR2 = .004, p < .01. The full regression model accounted for 13.3% of the variance of autonomous motivation. The interaction between autonomy support and autonomy orientation is depicted in Figure 1. As suggested by Aiken and West (1991), values of −1 versus +1 SD below or above the mean of autonomy orientation were used to indicate low versus high values of autonomy orientation. As the simple-slope analysis showed, the relationship between autonomy-supportive leadership and autonomous motivation was positive for volunteers with a weak autonomy orientation (–1 SD), with a slope of b = 0.25, t(1,727) = 11.01, p < .001. For volunteers with a strong autonomy orientation (+1 SD), the relationship between autonomy-supportive leadership and autonomous motivation was closer to zero, b = 0.16, t(1,727) = 6.67, p < .001 (Figure 1). Thus, H1c was supported. Further analyses showed that the region of significance for autonomy orientation ranged from 1.14 to 7.72, indicating that any simple slope outside this region was significant. The centered values for autonomy orientation ranged from −3.45 to 0.55. The region of significance was above the possible range of the values for autonomy orientation, indicating that for any value of the “autonomy orientation” moderator the simple slope of autonomous motivation on autonomy-supportive leadership was significantly different from zero. Therefore, even individuals with a very strong autonomy orientation showed a slight decrease in autonomous motivation under less autonomy-supportive leadership.
Hierarchical Multiple Regression Analyses Predicting Autonomous Motivation (Hypothesis 1c) and Controlled Motivation (Hypothesis 2c) from Autonomy-Supportive Leadership and Causality Orientations.
Note: WCQ = Work Climate Questionnaire measuring autonomy-supportive leadership. GCOS = General Causality Orientation Scale measuring causality orientations (personality).
p < .05. **p < .01. ***p < .001.
Autonomy orientation moderating the relationship between autonomy-supportive leadership and autonomous motivation. GCOS aut = autonomy orientation.
Control Orientation as Moderator
The second hierarchical regression analysis tested whether volunteers’ control orientation moderated the relationship between autonomy-supportive leadership and controlled motivation. In step 1, to examine the main effects, controlled motivation was regressed on autonomy support and control orientation. The main effect of autonomy-supportive leadership was positive, β = .13, t(1,638) = 5.60, p < .001, indicating that higher autonomy-supportive leadership led to higher controlled motivation. Therefore, H2a was not supported. The main effect of control orientation (personality), β = .28, t(1,638) = 11.87, p < .001, was positive, supporting H2b. In step 2, the interaction term was entered to test for moderation. Adding the interaction term resulted in a significant increase in explained variance, ΔR2 = .003, p = .02 (Table 2). The full regression model accounted for 9.9% of the variance of controlled motivation. The pattern of the significant interaction is depicted in Figure 2. As previously, values of −1 versus +1 SD below or above the mean of control orientation were used to indicate low versus high values of control orientation (Aiken & West, 1991). The simple-slope analysis further clarified the moderating influence of control orientation. The simple slope of autonomy support on controlled motivation was positive for individuals with a strong control orientation (+1 SD), b = 0.22, t(1,638) = 5.68, p < .001, and close to zero for individuals with a weak control orientation (–1 SD), b = 0.08, t(1,638) = 2.06, p = .04. This supported the assumption of an interaction, but not exactly as specified in H2c, as the main effect of autonomy-supportive leadership was positive instead of negative. Further analyses showed that the region of significance for control orientation ranges from −8.99 to −0.77, indicating that any simple slope outside this region is significant. The centered values for control orientation range from −1.92 to 2.08. Therefore, the simple slope of autonomy-supportive leadership on controlled motivation was not significant for volunteers with a level of control orientation >1.01 SD below the mean (corresponding to 31.0% of the volunteers). For volunteers with higher levels of control orientation, the relationship between autonomy-supportive leadership and controlled motivation became stronger (and more positive) as the level of control orientation increased.
Control orientation moderating the relationship between autonomy-supportive leadership and controlled motivation. GCOS contr = control orientation.
Discussion
The aim of the present study was to have a closer look at the impact of autonomy-supportive leadership on volunteer motivation within the framework of SDT. In particular, the moderating role of volunteers’ autonomy and control orientation (personality) on the impact of autonomy-supportive leadership on volunteer motivation was studied.
First, the assumption that volunteers’ autonomy orientation would moderate the relationship between autonomy-supportive leadership and autonomous motivation was tested. In support of H1c, volunteers with a strong autonomy orientation showed higher autonomous motivation under less autonomy-supportive conditions compared to volunteers with a weak autonomy orientation. Thus, volunteers with a strong autonomy orientation seemed more likely to be able to perceive autonomy from a not-so-supportive social context than did volunteers with a weak autonomy orientation. Under autonomy-supportive conditions, all volunteers showed nearly as much autonomous motivation, independent from their autonomy orientation. This result may be interpreted in a way that a strong autonomy orientation could serve as partial protection from low autonomy-supportive social contexts, as similarly found by Hagger and Chatzisarantis (2011). The positive main effect of autonomy support on autonomous motivation (H1a) was in line with previous results (Gagné & Deci, 2005); the same can be said for the positive main effect of autonomy orientation on autonomous motivation (H1b; Kwan et al., 2011; Williams & Deci, 1996).
Second, the assumption that volunteers’ control orientation would moderate the relationship between autonomy-supportive leadership and controlled motivation was tested. Based on the theoretical assumptions of SDT (Deci & Ryan, 2000) and previous studies (Gagné et al., 2003; Pelletier et al., 1997; Williams et al., 1996), a negative main effect of autonomy-supportive leadership on controlled motivation was assumed (H2a). Instead, a slightly positive main effect, contradictive to the theoretical assumption, was found. Haivas et al. (2012) also found a positive relationship—with an effect size similar to this study—between autonomy support and controlled forms of regulation. This effect is somewhat surprising but not fully inexplicable. They stated that the positive main effect could be partly explained by the fact that introjected regulation (i.e., the more internalized form of controlled motivation) is substantially correlated with autonomous motivation. This suggests that introjected regulation might be closer to autonomous motivation than theoretically assumed. From a practical point of view, the positive correlation between autonomy-supportive leadership and controlled motivation might indicate that a nice, autonomy-supportive supervisor not only induces a positive feeling of autonomy, but also a certain feeling of pressure or obligation. This interpretation is strengthened by the fact that autonomy-supportive leadership is substantially correlated to introjected regulation that is defined as behaving due to threats of guilt and shame (Deci & Ryan, 2008a).
However, the data of the present study support the assumption that control orientation moderates the relationship between autonomy-supportive leadership and controlled motivation (H2c). This relationship is positive for volunteers with a medium to strong control orientation, whereas no correlation exists between autonomy-supportive leadership and controlled motivation for volunteers with a weak control orientation (i.e., >1.01 SD below mean levels of control orientation). Therefore, volunteers with medium to high levels of control orientation seem to perceive a substantial feeling of pressure and obligation from autonomy-supportive leadership. It is assumed that the more these volunteers are autonomy supported, the more pressure and obligation they perceive, and therefore, the more likely they are to volunteer based on controlled motivation. Volunteers with a weak control orientation do not seem to perceive the same feeling of pressure and obligation from autonomy-supportive leadership.
Limitations and Further Research
Inevitably, this study was also subject to limitations. First, the design of the study only allowed for investigating the impact of more or less autonomy-supportive leadership and not the impact of controlling leadership. For future research, it would be interesting to measure controlling interpersonal behavior—as opposed to autonomy support—to investigate the moderating role of causality orientations on the relationship between controlling leadership and volunteer motivation. Although the quality of a standardized laboratory setting was not quite reached, the advantage of the measurement of autonomy support used in this study was that it made it possible to have a closer look at the complexity of the relationship between volunteers and their supervisors. As autonomy-supportive leadership did positively correlate with autonomous and controlled motivation, autonomy support seems to contain a broad range of different supporting, controlling, informing, encouraging or limiting facets.
Second, the interaction effects were rather small in effect size. As data were collected using survey research, common method variance (CMV; Podsakoff, MacKenzie, Lee, & Podsakoff, 2003) might be a concern. In the case of regression analyses testing for moderation effects, Siemsen, Roth, and Oliveira (2010) demonstrated that CMV causes a deflation in the estimated interaction effects, but cannot create an artificial interaction effect. Based on their analyses, Siemsen et al. (2010; p. 470) stated that “finding interaction effects despite the influence of CMV in the data set should be taken as strong evidence that an interaction effect exists”. Additionally, even though tasks in which volunteers had regular contact with their supervisors were chosen, there may have been a few volunteers with sparse supervisor contact who completed the WCQ, even though they found it difficult to rate the autonomy-supportiveness of their supervisors. As a consequence, measurement errors were possibly higher in the data than expected, serving to additionally reduce the power and effect size of the main and interaction effects. Nevertheless, both assumed moderation hypotheses were confirmed, providing evidence that causality orientations moderate the relationship between autonomy-supportive leadership and volunteer motivation.
Third, antecedents and consequences of volunteer activities change over time (Finkelstein, 2008). Therefore, cross-sectional studies such as this can only provide partial insight into the complex process of determinants of volunteer motivation. Nonetheless, the present study was based on a large and diverse sample of volunteers engaged in a broad range of different charitable volunteering tasks, as suggested by Lam and Gurland (2008). Therefore, the results of this study may well be generalizable to other volunteering contexts, taking into consideration that the impact of leadership might be larger in certain volunteering tasks (i.e., very close cooperation between volunteers and supervisors) than in others (i.e., volunteering tasks with a clear end like in rotating systems or event volunteering). A large and diverse sample might also cause some biasing effects of demographic variables. However, it is assumed that they may have an additional main effect on volunteer motivation and would not bias the moderating effects examined in this study. Future studies, most likely employing a longitudinal design, would be needed to examine questions of causality. For example, it could be interesting and useful to study possible effects of age on volunteer motivation.
Fourth, the results of this study could also be of interest in the context of paid work or in other organizational settings. Prior to generalizing the results of the present study, however, at least the influence of personality in different organizational settings should be taken into account. According to Snyder and Ickes (1985), personality is more predictive in case of “weak” settings (i.e., allowing individuals to act on their own impulses because strong norms and rules are missing), whereas traits are less predictive of behavior in “strong” settings (i.e., contexts providing strong norms and rules). In general, it is assumed that the interplay of volunteers and their supervisors corresponds more to weak settings than to strong settings (not excluding that there might also be some volunteering contexts with strong norms and rules; e.g., volunteering in hospitals). Thus, the results of this study as such might not be transferable to the paid work context, as the personality of paid workers might play a minor role compared to the volunteering context.
Fifth, Haivas et al. (2012) also found a positive relationship between autonomy support and controlled motivation in their study on volunteering. Therefore, the direction of this relationship may—in addition to the moderating effect of personality—also depend on the research context. As the feeling of acting autonomously is highly valued by volunteers (Leonard et al., 2004; Vantilborgh et al., 2012), it might be concluded that volunteers are perhaps more sensitive to perceiving even small signs of control, especially when they are control-oriented. Therefore they may—contrary to other contexts—perceive control out of autonomy-supportive leadership, which in turn results in a positive relationship between autonomy-supportive leadership and controlled motivation. However, as there are only very few studies examining the relationship between autonomy-supportive leadership and controlled motivation in the volunteering context, further studies on this assumption are needed.
Conclusions and Practical Implications
In conclusion, the results of this study substantiate the assumption that individual differences in causality orientations moderate the relationship between autonomy-supportive leadership and autonomous or controlled volunteer motivation. In the latter case, control orientation (personality) may explain the variation in the relationship between autonomy support and controlled motivation in previous studies (cf., Gagné et al., 2003; Haivas et al., 2012; Williams et al., 1996), even though no indicators for a negative correlation were found. From an SDT point of view, the findings of this research study indicate that examining the impact of the social context and personality in relative isolation from each other could indeed lead to an incomplete picture of (volunteer) motivation.
Nevertheless, despite its limitations, this study showed that autonomy-supportive leadership is, in general, something positive and beneficial for volunteers. It was demonstrated that volunteers with a weak autonomy orientation benefit most of all from autonomy support. Volunteers with a strong autonomy orientation also benefit from autonomy-supportive leadership, while at the same time their personality seams to partially protect them from contingencies that may reduce autonomous motivation. On the contrary, the positive relationship between autonomy-supportive leadership and controlled motivation indicates that nice, kind, and supportive volunteer coordinators may, at the same time, produce a certain feeling of pressure or obligation for certain volunteers. This could occur because volunteers would not like to disappoint a nice and supporting supervisor by, for example, terminating the voluntary engagement, substantially reducing the amount of volunteering or making mistakes. The stronger control-oriented the volunteers are, the more likely this feeling of obligation is perceived. Additionally, volunteers with a strong control orientation also explicitly look for opportunities to please someone or to meet others’ expectations (e.g., their volunteering coordinator, family or friends). Based on these findings, a practical implication may be that volunteers need to have the opportunity to stop volunteering or substantially reduce their engagement without feeling guilty towards their supervisors. Therefore, the end of a voluntary engagement should always be taken into account and be handled as a natural occurrence (Musick & Wilson, 2008) and ideally, organizations should openly address these points with their volunteers (e.g., when starting the engagement or in feedback meetings). Furthermore, it could be useful for practitioners to gain insight on how to detect control orientation within volunteers. It is suggested that volunteers with a control orientation might have difficulty with setting boundaries towards volunteering and refusing requests from their supervisors on additional volunteering tasks. Although supervisors might be tempted to take advantage of this willingness, they should be aware that acknowledging this willingness may serve to strengthen a feeling of guilt, control, and obligation within volunteers, and caution should be exercised.
Footnotes
Declaration of Conflicting Interests
The authors declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
This work was supported by the Swiss National Science Foundation (SNSF).