Journal of Vocational Behavior 81 (2012) 354363

Contents lists available at SciVerse ScienceDirect

Journal of Vocational Behavior

journal homepage: www.elsevier.com/locate/jvb

A prole approach to self-determination theory motivations at work

Christina M. Moran a,, James M. Diefendorff b, Tae-Yeol Kim c, Zhi-Qiang Liu d
a b c d

PRADCO, 178 E. Washington St., Chagrin Falls, OH 44022, USA University of Akron, Department of Psychology, Akron, OH 44325-4301, USA Management Department, China Europe International Business School, 699 Hongfeng Road, Pudong, Shanghai 201206, China Management School, Huazhong University of Science and Technology, Wuhan, Hubei Province, China

a r t i c l e

i n f o

a b s t r a c t
Self-determination theory (SDT) posits the existence of distinct types of motivation (i.e., external, introjected, identified, integrated, and intrinsic). Research on these different types of motivation has typically adopted a variable-centered approach that seeks to understand how each motivation in isolation relates to employee outcomes. We extend this work by adopting cluster analysis in a person-centered approach to understanding how different combinations or patterns of motivations relate to organizational factors. Results revealed five distinct clusters of motivation (i.e., low introjection, moderately motivated, low autonomy, self-determined, and motivated) and that these clusters were differentially related to need satisfaction, job performance, and work environment perceptions. Specifically, the self-determined (i.e., high autonomous motivation, low external motivation) and motivated (i.e., high on all types of motivation) clusters had the most favorable levels of correlates; whereas the low autonomy (i.e., least self-determined) cluster had the least favorable levels of these variables. Published by Elsevier Inc.

Article history: Received 17 April 2012 Available online 27 September 2012 Keywords: Cluster analysis Profile analysis Self-determination theory Motivation

1. Introduction Across the many theories of work motivation it is quite common to conceptualize motivation as varying primarily in quantity rather than in quality or type (Gagn & Deci, 2005). As an exception, self-determination theory (SDT; Deci & Ryan, 2000; Ryan & Deci, 2000) suggests that there are two primary types of motivation that can guide individual behavior: extrinsic motivation (i.e., to attain a reward or consequence separable from an activity itself) and intrinsic motivation (i.e., to do something because of an inherent inclination or interest; Gagn & Deci, 2005). Further, extrinsic motivation can be divided into four types ranging from least to most autonomous: external (i.e., for reward or praise), introjected (i.e., to avoid guilt or anxiety), identified (i.e., because the person sees value in the activity), and integrated (i.e., because the person has internalized the reasons for engaging in the behavior; Gagn & Deci, 2005; Koestner & Losier, 2002; Ntoumanis, 2002; Wang & Biddle, 2001). Research from a variety of domains has linked these distinct forms of motivation to situational characteristics (e.g., Bono & Judge, 2003), well-being outcomes (e.g., Ryan, Deci, & Grolnick, 1995), and effective functioning (e.g., high effort expenditure, better learning; Grolnick & Ryan, 1987; Ryan & Connell, 1989). However, most of this research has adopted a variable-centered approach in which the focus is on testing the relationships of each type of motivation with other variables (cf. Aldenderfer & Blashfield, 1984). Although such an approach provides valuable information about the direct and unique links of each motivation with other variables, it ignores the possibility that (a) distinct constellations of motivational profiles exist in the population and (b) these SDT motivation profiles may correspond to differences in other variables. This perspective is consistent with taking a person-centered approach to conceptualizing SDT motivations, arguing that distinct motivational profiles might exist and that investigating these motivation types might reveal unique insights into the ways in which SDT motivations tend to co-occur and the effects of these profiles on other variables.
Corresponding author. E-mail addresses: cms116@zips.uakron.edu (C.M. Moran), jdiefen@uakron.edu (J.M. Diefendorff), tykim@ceibs.edu (T.-Y. Kim), zqliu@mail.hust.edu.cn (Z.-Q. Liu). 0001-8791/$ see front matter. Published by Elsevier Inc. http://dx.doi.org/10.1016/j.jvb.2012.09.002

C.M. Moran et al. / Journal of Vocational Behavior 81 (2012) 354363

355

Studies conducted in the educational and sport realms have made some progress in looking at motivation profiles (e.g., Ntoumanis, 2002; Ratelle, Guay, Vallerand, Larose, & Sencal, 2007; Wang & Biddle, 2001), but to our knowledge no organizational research has examined this issue. The present study aims to bridge this gap in the literature by (a) assessing each of the primary motivations described by SDT in an organizational setting, (b) identifying motivation profiles in our sample, and (c) linking the motivation profiles to correlates at work. 2. Motivation from a self-determination theory perspective Intrinsic motivation is present when individuals do something for pleasure or enjoyment, whereas extrinsic motivation occurs when individuals do something because of external forces (Deci & Ryan, 2000). Although authors have long recognized a distinction between extrinsic and intrinsic motivation (e.g., Porter & Lawler, 1968), SDT is unique in that it further divides extrinsic motivation into four types (i.e., external, introjected, identified, integrated) that vary in the degree to which motivation has been internalized (Gagn & Deci, 2005). External regulation is the most extrinsic form of motivation as it represents motivation due to explicit external control (Ntoumanis, 2002). Introjected motivation is the second-most extrinsic form of motivation, reflecting compulsion, avoidance of guilt or anxiety, and a sense that one should or ought to complete the behavior (Koestner & Losier, 2002; Ntoumanis, 2002; Wang & Biddle, 2001). These two types of extrinsic motivation are described as controlled forms of motivation because they place the impetus for action solely with external factors (Sheldon & Elliot, 1998). Identified motivation describes regulation of behavior for reasons more consistent with one's goals and identity, with individuals seeing the actions as personally important (Koestner & Losier, 2002). People acting based on an identified motive do so because they want to as opposed to feeling that they ought to, as in introjected motivation (Wang & Biddle, 2001). Finally, integrated motivation is the most internalized form of extrinsic motivation (Gagn & Deci, 2005) in which the person values and accepts the reasons for the behavior, though he/she may still not consider it to be inherently fun or interesting (Gagn & Deci, 2005). For instance, a nurse may fully identify with actions aimed at alleviating patient suffering, though he/she may not consider the actions to be enjoyable. Identified and integrated extrinsic motivations, along with intrinsic motivation, are autonomous forms of motivation (Sheldon & Elliot, 1998). Research has tended to examine the relationships of each motivation with other variables, either by itself or in the presence of the other motivations. For example, Reeve (2002) noted the following correlates of autonomous (i.e., internalized) motivation in students: academic achievement (Miserandino, 1996), perceived competence and self-worth (Ryan & Grolnick, 1986), positive emotionality (Ryan & Connell, 1989), creativity (Amabile, 1985), and retention (Vallerand & Bissonette, 1992). Other authors have suggested that intrinsic, integrated, and even identified motivation are related to increased achievement, positive affect, persistence, effort, and well-being (e.g., Frederick-Recascino, 2002; Koestner & Losier, 2002). Research on SDT in work contexts has generally found that autonomous motivations (e.g., intrinsic, integrated, identified) and factors known to enhance autonomous motivation (e.g., autonomy-supportive environments) lead to better well-being and effectiveness than controlled motivations (e.g., external, introjected; Baard, Deci, & Ryan, 2004; Bono & Judge, 2003; Deci et al., 2001; Greguras & Diefendorff, 2010; Ilardi, Leone, Kasser, & Ryan, 1993; Kasser, Davey, & Ryan, 1992). However, there is some debate about the prevalence and effects of intrinsic and extrinsic motivations at work (e.g., Sansone & Harackiewicz, 2000). For example, some authors suggest that intrinsic motivation is less likely to occur in the work context than in other realms (e.g., hobby, sport) because of the inherent focus on compensation and recognition at work (e.g., Baard, 2002). However, other authors note that extrinsic rewards issued independent of task engagement, as in the case of salaried positions, do not necessarily undermine intrinsic motivation (e.g., Deci, Koestner, & Ryan, 1999). 3. Person-oriented versus variable-oriented approaches Person-centered research has been referred to as a holistic, interactionistic view in which the individual is seen as an organized whole, functioning and developing as a totality (Bergman & Magnusson, 1997, p. 291). This approach is in contrast to the variable-centered perspective, which aims to assess relations between variables across individuals (Aldenderfer & Blashfield, 1984). Thus, a point of distinction between the two approaches is that the person-centered perspective begins by identifying individuals with common attributes and then aims to describe how groups of homogeneous individuals function, whereas the variable-centered perspective starts by identifying variables of interest and then aims to describe how these variables function across individuals (Aldenderfer & Blashfield, 1984). In addition to answering different research questions, the two approaches are also associated with different analytic tools, with the variable-centered approach emphasizing correlation and regression analyses and the person-centered approach relying on cluster analysis and related profiling techniques. In the case of SDT, most of the research to date has been conducted from a variable-centered perspective (e.g., Baard et al., 2004; Ilardi et al., 1993; Kasser et al., 1992; Ryan et al., 1995). Such research is useful for understanding how particular motivations uniquely relate to outcomes, but it does not tell us much about how an individual's standing on multiple motivations might shape outcomes or whether some constellations of SDT motivations are more common at work than others. 4. The present investigation In the current investigation, we measured the five motivations from SDT in a sample of employees from China. We then performed cluster analysis on the data to identify distinct motivation profiles in the sample. Cluster analysis is particularly

356

C.M. Moran et al. / Journal of Vocational Behavior 81 (2012) 354363

well-suited when taking an exploratory approach to identifying clusters, which is appropriate in this situation given the lack of theory and research on SDT profiles at work (which is in contrast to other techniques, such as latent class analysis, that adopt a confirmatory approach to supporting cluster solutions; Marsh, Ldtke, Trautwein, & Morin, 2009). Once motivation profiles were identified, we examined whether these profiles were associated with different levels of work environment perceptions, psychological need satisfaction, and job performance. Generally, profiles may differ in terms of the quantity (i.e., all motivations are low, medium, or high) or quality (e.g., some motivations are high and others are low; Marsh et al., 2009) of the variables upon which the profiles are derived. Although we refrain from making formal hypotheses regarding which motivation profiles might emerge at work, a simplified matrix of possible motivation profiles focusing on the more general autonomous and controlled motivations is presented in Table 1. This table shows how the motivations might combine to produce profiles that differ in quantity and quality. Profile analyses with students in academic contexts (i.e., Boich, Sarrazin, Grouzet, Pelletier, & Chanal, 2008; Hayenga & Corpus, 2010; Liu, Wang, Tan, Koh, & Ee, 2009; Ntoumanis, 2002; Ratelle et al., 2007) have typically found three- or four-cluster solutions, with the clusters corresponding to the motivated, moderately motivated, amotivated, external, and internal profiles in Table 1. Further, these studies generally report that internally-oriented profiles perform best. Examining motivation profiles of adolescents in sport contexts across two studies, Gillet, Vallerand, and Rosnet (2009) supported a four-cluster solution (i.e., corresponding to motivated, moderate internal, internal, and external in Table 1) in their first study, with the external profile having the lowest performance. In their second study, Gillet et al. (2009) found a three-cluster solution (i.e., moderate internal, moderately motivated, and motivated in Table 1) and showed that the motivated profile performed better than the moderately motivated profile. Finally, Wang and Biddle (2001), in a study of junior college athletes identified five clusters in their data, roughly corresponding to internal, motivated, poorly motivated, amotivated, and moderate external in Table 1. Although these past studies show some consistency (e.g., identifying 35 clusters, linking internal clusters to better outcomes than external clusters), there is almost no consistency across the studies in terms of the variables used as input to form clusters. For example, Liu et al. (2009) and Ntoumanis (2002) included antecedents and outcomes (e.g., incremental and entity beliefs, perceived sport competence, physical activity, well-being, need satisfaction, effort, enjoyment) along with the motivation variables in forming clusters. Aldenderfer and Blashfield (1984) advised against this procedure as it confounds cluster membership with relationships to external criteria. Another limitation of past work is studies typically did not include the full set of extrinsic motivations, with no study measuring integrated motivation. Finally, the profiles that have been identified in past research focus on educational and sport contexts, which may not produce the same clusters that are important in a work context (Baard, 2002). For instance, external pressure may be very salient in educational settings, whereas intrinsic pleasure may be prominent in sport contexts. The work setting may differ from both of these by providing more of an equal emphasis on both ends of the SDT motivation spectrum (e.g., Ratelle et al., 2007). As a result, we might expect that distinct clusters will emerge in the work context. Further, it is possible that the clusters that do emerge will have different effects at work compared to at school or sport contexts. A person-centered approach to SDT motivations in the workplace can shed light on these issues. 5. Method 5.1. Participants and procedure We selected organizations to sample with the goal of obtaining a broad cross-section of employees who differ in the type of work performed. To do this, we selected organizations located in different areas of the country and working in different industries. This approach resulted in participants coming from 12 large organizations (1 hospital, 2 financial securities companies, 3 manufacturing companies, 1 service company, 2 real estate companies, 1 energy company, 2 government agencies) distributed in three regions of China: Wuhan (a central city), Shanghai (an eastern city), and Guangzhou (a southern city). Further, each company employed more than 400 individuals. Seventy-one of 87 managers contacted agreed to participate in the study under the condition that they received copies of the results. Managers were asked to complete the supervisor survey, then to give another survey to three or four of their immediate subordinates for completion during work hours. All surveys were returned directly to the researchers in sealed envelopes. Brislin's (1986) back-translation procedure, in which all translators were blind to the study's hypotheses, was used to translate

Table 1 Matrix of possible motivation profiles. Autonomous motivation Low Controlled motivation Low Moderate High Amotivated Moderate externals Externals Moderate Moderate internals Moderately motivated Motivated Externals High Internals Motivated internals Motivated

Note. Controlled motivation encompasses external and introjected motivations, whereas autonomous motivation encompasses the three types of motivation that are more internalizedidentified, integrated, and intrinsic motivations (Sheldon & Elliot, 1998). Because the matrix of possible profiles is actually five-dimensional in the present studyin contrast to the two displayed in this tablethe number of possible profiles is much greater than suggested above. However, this table provides an idea of the ways in which distinct types of motivation might combine to form profiles of motivation in a concise, two-dimensional format.

C.M. Moran et al. / Journal of Vocational Behavior 81 (2012) 354363

357

the surveys from English to Chinese. Two bilingual individuals independently translated the survey from English to Chinese and a third bilingual individual translated the survey back to English. This procedure resulted in 15 words or phrases that did not retain their meaning across translations, and these wordings were revised based on agreement among the translators. A total of 226 employee questionnaires were returned and matched respectively with 62 supervisor questionnaires. Nine of 71 managers failed to distribute subordinate surveys (a response rate of 87.3%). Of the 226 subordinates, 56.8% were male. Subordinates' average age was 31.44 years (SD = 7.83); their average job tenure was 9.84 years (SD = 7.81). Of the 62 supervisors, 79.4% were male, with an average age of 38.12 years (SD = 6.01) and average job tenure of 16.63 years (SD = 7.18). One subordinate's data was excluded because of missing data. Thus, all measures used in the present study were completed by the final sample of 225 subordinates, with the exception of in-role performance which was rated by supervisors (n = 62). 5.2. Measures 5.2.1. Social support This measure contained three items (Campion, Medsker, & Higgs, 1993). A sample item is, My team increases my opportunities for positive social interaction. Responses were made on a 5-point scale from Strongly Disagree to Strongly Agree. All scale reliabilities (coefficient alpha) are presented in Table 2. 5.2.2. Job characteristics The job characteristics of skill variety, task identity, task significance, and feedback were measured using 12 items (Job Diagnostic Survey; Hackman & Oldham, 1975). Each construct was measured with three items; for example, a skill variety item was, My job gives me the opportunity to use many new technologies. Job autonomy was measured with Morgeson and Humphrey's (2006) nine-item scale assessing autonomy in work methods, work scheduling, and decision making. A sample item is, My job allows me to plan how I do my work. Responses were made on a 5-point scale from Strongly Disagree to Strongly Agree. 5.2.3. Motivation A measure of the five motivations (three items each) was created based on the theoretical work of Deci and Ryan (2000; Ryan & Deci, 2000), with specific items for the external, introjected, identified, and intrinsic scales adapted from one or more of the following sources: Grant (2008), Guay, Vallerand, and Blanchard (2000), Levesque, Sell, and Zimmerman (2006), and Ryan and Connell (1989). Because none of these past studies assessed integrated motivation, items were written to tap this construct based on the underlying conceptual definition. Prior to administering the items, they were reviewed for conceptual clarity by a team of four experts trained in this area of study, with small revisions to item wordings made. Participants were asked to respond to the question Why are you motivated to do your work? by indicating the extent to which they agreed with each of the 15 items using a 5-point scale from Strongly Disagree to Strongly Agree. Confirmatory factor analysis of these items was conducted using Mplus version 6.11 (Muthn & Muthn, 19982011). The a priori, five-factor structure fit the data well ( 2 = 180.18, df = 80, root-mean square error of approximation [RMSEA] = .08, standardized root mean square residual [SRMR] = .05, TuckerLewis index [TLI] = .92, comparative fit index [CFI] = .94) according to established guides for acceptable fit (i.e., RMSEA b .08, SRMR b .10, CFI > .90, TLI > .90; Vandenberg & Lance, 2000). Several alternative, more parsimonious models were tested, all of which fit the data worse. The next best fitting model combined the identified and integrated motivations ( 2 = 259.57, df = 84, RMSEA = .10, SRMR = .07, TLI = .86, CFI = .89), but it still fit worse than the a priori model ( 2 = 79.39, df = 4, p b .001). Factor loadings are presented in Table 3.

Table 2 Means, standard deviations, and zero-order correlations among study variables. Mean 1. External motivation 2. Introjected motivation 3. Identified motivation 4. Integrated motivation 5. Intrinsic motivation 6. Social support 7. Job autonomy 8. Skill variety 9. Task identity 10. Task significance 11. Feedback 12. Need satisfaction 13. In-role performance 3.60 3.96 4.07 3.87 3.62 3.99 3.26 3.56 3.32 3.59 3.57 5.12 3.82 SD 0.80 0.86 0.75 0.90 0.95 0.90 1.01 0.94 1.07 1.00 0.81 0.81 0.55 1 .63 .09 .12 .15 .10 .01 .01 .03 .10 .00 .10 .11 .11 2 .86 .42 .25 .30 .43 .22 .12 .09 .22 .22 .42 .14 3 4 5 6 7 8 9 10 11 12 13

.72 .53 .51 .50 .28 .39 .28 .49 .26 .56 .22

.86 .65 .45 .28 .39 .30 .30 .22 .50 .26

.88 .46 .29 .27 .22 .28 .16 .52 .18

.79 .33 .35 .31 .38 .25 .61 .12

.91 .35 .40 .30 .22 .43 .13

.73 .34 .43 .21 .36 .22

.82 .36 .37 .36 .16

.76 .39 .37 .13

.70 .32 .05

.91 .23

.78

Note. SD = standard deviation. Internal consistency reliabilities are along the diagonal. N = 225. p b .05. p b .01.

358

C.M. Moran et al. / Journal of Vocational Behavior 81 (2012) 354363

5.2.4. Psychological need satisfaction at work The Basic Need Satisfaction at Work Scale (Baard et al., 2004; Deci et al., 2001; Ilardi et al., 1993; Kasser et al., 1992) consists of 21 items to which participants responded on a 7-point scale from Not at all true to Very true. An example item is, I have been able to learn interesting new skills on my job. In some cases (e.g., Baard et al., 2004; Deci et al., 2001), this scale has been broken out into three subscales pertaining to the satisfaction of autonomy, relatedness, and competence needs. Results are presented for one overarching need satisfaction construct and three individual subscales. 5.2.5. In-role performance In-role performance was measured with seven items from Williams and Anderson (1991). Supervisors rated the degree to which subordinates adequately complete assigned duties and meet formal performance requirements of the job, for example. Responses were made on a 5-point scale from Strongly Disagree to Strongly Agree. 6. Results Means, standard deviations, reliabilities, and correlations among variables are in Table 2. One outlier was removed from the analyses because it was more than three standard deviations from the mean (Liu et al., 2009). Cluster analysis was conducted in two steps. The first step consisted of subjecting the five motivation variables (left unstandardized given that all responses were made on the same 5-point scale and the belief that relative differences on the scale could be meaningful) to agglomerative (i.e., hierarchical) cluster analysis using Ward's (1963) method and the squared Euclidean distance measure of similarity. Ward's (1963) method was chosen to optimize the minimum variance within clusters (Aldenderfer & Blashfield, 1984). Inspection of the dendogram, agglomeration schedule, and cluster means suggested that a five-cluster solution should be retained. The dendogram indicated that two to five clusters could be meaningfully discerned from the data, but any more than five would result in some cases being treated as their own clusters due to the very minute (i.e., indistinguishable) differences in clusters. Inspection of the agglomeration scheduleby plotting changes in agglomeration values each time a larger cluster is separated into two smaller clusters, similar to the use of the scree plot to determine the number of factors to retain in exploratory factor analysesindicated that between five and seven clusters should be retained, since changes in agglomeration values appeared to lessen after five clusters, and after seven clusters changes appeared negligible. Since both methods of determining the appropriate amount of clusters suggested that it would be appropriate to retain five clusters, cluster means were inspected to ensure that each profile was distinct from (i.e., not redundant with) the others. All of this information converged to support a five-cluster solution. The second step consisted of applying k-means cluster analysis to the same five motivation variables using the cluster centers from the hierarchical analysis as initial seed points. This step has been recommended (e.g., Aldenderfer & Blashfield, 1984) on the basis that hierarchical methods only go through the data one time, such that the initial data order can unduly influence the cluster solution (Gower, 1967), in contrast to iterative methods such as k-means cluster analysis. Sixteen percent (n = 37) of participants were categorized into the first cluster, which can be described as a low introjection cluster. Participants in this cluster had
Table 3 Items and standardized factor loadings resulting from confirmatory factor analysis of the motivation questionnaire. Factors and items Why are you motivated to do your work? External motivation Because my boss wants me to do it. Because the situation demands it. Because I get paid to do it. Introjected motivation Because I would feel guilty if I did not do well. Because I would feel ashamed if I did poorly. Because I would feel bad about myself if I did not do a good job. Identified motivation Because I believe my work is valuable. Because my work is important. Because I value the work. Integrated motivation Because my work goals and personal goals are integrated. Because my work is a big part of who I am. Because my work helps to define me. Intrinsic motivation Because I find the work interesting. Because the work is fun. Because I find the work engaging. Factor loading

.77 .60 .47

.89 .79 .78

.81 .64 .60

.82 .84 .81

.85 .88 .80

C.M. Moran et al. / Journal of Vocational Behavior 81 (2012) 354363

359

moderate scores on all types of motivation (see Fig. 1) except introjected motivation for which they had particularly low scores. In comparison to our a priori profile matrix in Table 1, this profile represents a particular subtype of the moderate internal profile, though not one that could have been anticipated based on past SDT profile research. Cluster 2 (n = 68; 30.2% of participants) scored moderately on all types of motivation, and thus corresponds to the moderately motivated profile in Table 1. Twelve percent (n = 27) of participants were grouped into Cluster 3, a cluster that can be described as low autonomy due to its particularly low means on integrated and intrinsic motivations. This cluster roughly corresponds to the moderate external profile in Table 1, except that identified motivationan autonomous form of motivationis at a moderate level that is similar to the controlled motivations. Cluster 4, containing 15.1% of participants (n = 34), could be described as self-determined as it is high on all types of motivation except external. This motivation corresponds most closely to the internal profile in Table 1, except that introjected motivation is high and at about the same level as the autonomous forms of motivation. Finally, Cluster 5 (n = 59; 26.2%) corresponds to the motivated cluster in Table 1 as individuals report high levels of all five motivations. Interestingly, we did not observe profiles that corresponded to several a priori profiles in Table 1 (i.e., amotivated, external, motivated internal, motivated external) and only two (i.e., moderately motivated, motivated) of the five observed profiles corresponded precisely to the a priori profiles described in Table 1. 6.1. Validation of cluster solution: links to correlates Aldenderfer and Blashfield (1984) noted that the preferred method of validating a cluster solution is to conduct significance tests on links to criteria of interest. Thus, we examined whether job characteristics (i.e., job autonomy, skill variety, task identity, task significance, and feedback), social support, in-role performance, and need satisfaction relate to a person's motivational profile. To test these links, we conducted several univariate ANOVAs with cluster membership as the independent variable and the correlates listed above as dependent variables (we repeated these analyses using multilevel modeling to account for any nesting effects associated with team membership and the substantive findings and interpretation did not change; these results are available from the first author). We also examined whether the profiles differed in terms of gender composition, job tenure, or age and did not find any differences. As shown in Table 4, the ANOVAs were significant for all of the correlates. To understand which clusters differed from one another on the correlates, we performed follow-up post-hoc tests using Tukey's (1953) honestly significant difference (HSD) correction for family-wise error. As shown in Table 4, social support and job characteristic perceptions were generally highest for the self-determined and motivated clusters (which did not differ from one another) and lowest for the low autonomy and low introjection clusters. For social support, the self-determined (i.e., Cluster 4) and motivated (i.e., Cluster 5) clusters had the highest scores, followed by the moderate cluster (i.e., Cluster 2), and both the low introjection (i.e., Cluster 1) and low autonomy (i.e., Cluster 3) clusters (see Table 4). The self-determined and motivated clusters reported more job autonomy than the low introjection and low autonomy clusters, with the self-determined cluster also being significantly different from the moderate cluster. The self-determined and motivated clusters reported the most skill variety and task identity, followed by the low introjection and moderate clusters, with the low autonomy

4.5

Mean

3.5

2.5

Low introjection

Moderate

Low autonomy

Self-determined

Motivated

Profile
External Introjected Identified Integrated Intrinsic

Fig. 1. Final cluster solution resulting from k-means cluster analysis.

360

C.M. Moran et al. / Journal of Vocational Behavior 81 (2012) 354363

Table 4 ANOVA results linking cluster membership to correlates. DV Cluster means 1 Low introjection Social support F4,220 = 24.67 R2 = .31 Job autonomy F4,219 = 6.77 R2 = .11 Skill variety F4,219 = 10.61 R2 = .16 Task identity F4,219 = 6.41 R2 = .11 Task significance F4,219 = 8.76 R2 = .14 Feedback F4,219 = 4.04 R2 = .07 Need satisfaction F4,220 = 35.82 R2 = .39 Autonomy NS F4,220 = 23.89 R2 = .30 Relatedness NS F4,220 = 15.33 R2 = .22 Competence NS F4,220 = 22.36 R2 = .29 Performance F4,220 = 4.39 R2 = .07 N % of total N 3.41
2,4,5

2 Moderate 3.95
1,3,4,5

3 Low autonomy 3.12

2,4,5

4 Self-determined 4.57
1,2,3

5 Motivated 4.441,2,3

2.974,5

3.223,4

2.612,4,5

3.691,2,3

3.531,3

3.403,4,5

3.513,4,5

2.701,2,4,5

3.951,2,3

3.881,2,3

3.213,4,5

3.124,5

2.691,4,5

3.731,2,3

3.681,2,3

3.303,4,5

3.593,5

2.821,2,4,5

3.981,3

3.921,2,3

3.344,5

3.583

3.162,4,5

3.741,3

3.791,3

4.582,4,5

5.061,3,4,5

4.232,4,5

5.681,2,3

5.621,2,3

4.152,3,4,5

4.811,3,4,5

3.471,2,4,5

5.381,2,3

5.341,2,3

4.942,4,5

5.381,4,5

4.934,5

6.041,2,3

5.761,2,3

4.614,5

4.913,4,5

4.182,4,5

5.561,2,3

5.761,2,3

3.695

3.82

3.555

3.81

4.021,3

37 16.4%

68 30.2%

27 12.0%

34 15.1%

59 26.2%

Note. Post-hoc tests are corrected for family-wise error using Tukey's Honestly Significantly Different correction. DV = dependent variable. NS = need satisfaction. Total N = 225. Superscripts indicate means in a given row that are statistically different at p b .05 from the cluster identified by the superscript. p b .01. p b .001.

cluster reporting the lowest skill variety (and task identity on par with the moderate cluster). The self-determined and motivated clusters reported the most task significance, followed by the low introjection and moderate clusters, and then the low autonomy cluster. Finally, the self-determined, motivated, and moderate clusters reported the most feedback, followed by the low introjection and low autonomy clusters. Thus, the self-determined and motivated clusters generally experienced more enriched jobs and social support, whereas the low autonomy and low introjection clusters tended to perceive lower levels of these work characteristics. Results for psychological need satisfaction (overall and for specific dimensions) followed a similar pattern. Specifically, the self-determined and motivated clusters experienced the most need satisfaction, followed by the moderate cluster, and lastly by the low introjection and low autonomy clusters (see Table 4). Results for in-role performance were somewhat different, as the motivated cluster had the highest supervisor ratings of performance, followed by the moderate and self-determined clusters. Performance for the motivated cluster was rated significantly higher than performance of the low introjection and low autonomy clusters. These findings suggest that having high levels of the autonomous motivations is important for need satisfaction, irrespective of the level of external motivation. In contrast, a profile with high levels of all motivations resulted in better supervisor ratings of performance than did a profile with high motivation on all dimensions except external. Comparison of the profile results to standard regression analyses (i.e., variable-centered approach with the SDT motivations as predictors; see Table 5) is informative. Note that in each case, the variable-centered approach produced models that accounted for more variance in the correlates than did the person-centered approach. This result is not surprising (Marsh et al., 2009) given that the variance present in the continuously measured variables is necessarily reduced when forming groups. However, it is also instructive to examine the specific coefficients in the regression models. Although for social support all five motivations uniquely predicted the dependent variable, for job autonomy and feedback none of the motivations uniquely predicted the dependent variable. One model had four significant coefficients (overall need satisfaction), two models had two significant coefficients (skill variety, task identity), and two models had one significant coefficient (task significance, job performance). In almost every

C.M. Moran et al. / Journal of Vocational Behavior 81 (2012) 354363

361

comparison, the profile analysis approach yielded an interpretation that goes beyond the regression analysis approach (e.g., regression analysis showed that integrated motivation uniquely predicted performance, but the comparison of profiles produced several significant differences). 7. Discussion Findings from cluster analysis and ANOVA suggest several conclusions. First, individual motivation on the job from an SDT perspective can be described concisely as falling into one of five patterns: low introjection, moderately motivated, low autonomy, self-determined, and motivated. Some of these patterns are distinct from patterns observed in sport and educational contexts and do not directly correspond to what might be deduced from a priori theoretical predictions, suggesting that SDT motivations combine in an interesting and unique way at work. Second, perceptions of social support and job characteristics, psychological need satisfaction, and supervisor ratings of performance differed across profiles in ways that would not necessarily be predicted by or readily observed with a variable-centered approach to SDT motivations. For instance, being high on the three autonomous motivations (i.e., intrinsic, integrated, identified) and introjected motivation along with being low or high on external motivation (i.e., the self-determined and motivated profiles) was associated with more favorable job characteristics and better psychological need satisfaction. Ratings of job performance were highest when all five motivations were high. Further, the lowest levels of the correlates were generally observed for two profiles: the low autonomy profile (low integrated and intrinsic, moderate identified, introjected, and external) and the low introjection profile (low introjection, moderate on the other motivations). Finding that low introjected or low intrinsic/integrated motivation (along with moderate levels of other motivations) was linked to similarly unfavorable levels of correlates is not something that would be predicted by SDT research using a variable-centered approach. Describing motivation at the person-level is a useful complement to variable-centered investigations (Hayenga & Corpus, 2010). Although it may be argued that interactions among variables might reveal motivation patterns, doing so in the current study would require analyses containing five main effects, ten two-way interactions, numerous three-way and four-way interactions, and a five-way interaction. Since each interaction term introduces non-normality and potential collinearity into the analyses (Moosbrugger, Schermelleh-Engel, & Klein, 1997), the person-centered investigation of the present data provides a more interpretable and concise way of understanding of how various motivations tend to co-occur and the impact of these combinations on other variables. 7.1. Implications for self-determination theory The finding that employees high on all motivations (i.e., the motivated cluster [Cluster 5]) tend to perform well is consistent with investigations showing high-performing athletes tend to exhibit high levels of both types of motivation (e.g., Chantal, Guay, Dobreva-Martinova, & Vallerand, 1996). Compared with past profile analyses of SDT motivation, the only other known study to identify a five-cluster solution was Wang and Biddle (2001). Though we replicated two profiles from their study (i.e., self-determined and motivated), three of our profiles were different from what they found (i.e., low introjection, moderately motivated, and low autonomy in the present study vs. poorly motivated, amotivated, and moderate external in Wang and Biddle (2001)). Comparison of the distinct profiles suggests that the unique profiles in our study exhibit overall higher levels of motivation than the unique profiles in Wang and Biddle (2001). This difference suggests that the motivation levels of employees may generally be higher across the board than the motivation of athletes, suggesting that the reasons why employees work may be higher and more diverse than the reasons athletes engage in sport. Additionally, the finding that self-determination is associated with favorable workplace correlates echoes a large body of work showing positive associations between autonomous motivation and beneficial outcomes (e.g., Ryan et al., 1995). Although autonomous and controlled motivations are typically pitted against one another, our results show that they can co-occur and that

Table 5 Variable-centered results using regression. Predictors Dependent variable Social support Extrinsic motivation Introjected motivation Identified motivation Integrated motivation Intrinsic motivation (Variable-centered) Regression R2 (Person-centered) ANOVA R2 .11 .26 .23 .17 .17 .62 .31 Job autonomy .07 .11 .11 .12 .13 .36 .11 Skill variety .04 .06 .30 .27 .03 .45 .16 Task identity .05 .04 .19 .20 .01 .34 .11 Task significance .06 .02 .45 .05 .02 .50 .14 Feedback .06 .14 .15 .13 .05 .31 .07 Need satisfaction .01 .20 .27 .17 .21 .66 .39 Performance .06 .05 .09 .20 .02 .29 .07

Note. Regression coefficients are standardized betas (). p b .05. p b .01. p b .001.

362

C.M. Moran et al. / Journal of Vocational Behavior 81 (2012) 354363

the existence of controlled motivations may not undermine self-determination. Thus, as long as an employee is autonomously motivated, it does not appear that controlled motivation is harmful for psychological need satisfaction. For example, if a person identifies with his/her work or truly enjoys the job, it may not matter for need satisfaction that he/she is also motivated because of pay. However, for job performance it seems that having both autonomous and controlled motivation is linked to better supervisor evaluations. This finding is particularly interesting given that researchers have been somewhat split as to whether controlled motivation will have favorable or detrimental effects in the workplace (e.g., Gagn & Deci, 2005). 7.2. Limitations, future research, and conclusion A limitation of the present study was that one of our motivation subscales was newly constructed and the other subscales were derived from past works. However, given the setting and sample of the present study (i.e., employees at work) and the fact that a measure of all five motivations did not exist, creating a new, theory-based measure was appropriate. Secondly, our data were cross-sectional thereby limiting inferences of causality. Given that this was the first attempt to link SDT profiles with organizational variables, we believe the cross-sectional results are of value. Nonetheless, future research should employ experimental and longitudinal designs aimed at better isolating the causal direction. Lastly, we cannot be certain that our results would generalize to all work settings and cultural contexts, as the present participants were sampled from twelve companies in seven industries in one country. Though we attempted to increase the generalizability of our results by sampling employees from multiple organizations in a variety of industries, we cannot be certain that the present results would replicate to other work settings in other countries. Further, culture may have played a role in shaping the motivation profiles observed in this study. For instance, it may be that the cultural context of China could have produced distinct profiles (e.g., the low introjection profile) that would not be observed in other countries, or that the observed effects of the profiles on other variables would not replicate in a Western cultural context (e.g., performance being highest when all motivations were high). Future work in which cultural variables are measured and participants are sampled from different countries could be useful for untangling the role of culture in the development of SDT profiles. Future research might also examine the antecedents that allow a person to be simultaneously high in autonomous and controlled motivations, as well as the temporal stability of the different motivation profiles (Ntoumanis, 2002). Only Hayenga and Corpus (2010) have examined SDT motivation using a person-centered approach in a longitudinal design. Though the academic sample used in Hayenga and Corpus's (2010) study may not be representative of the profiles at work, their results showed that 43% of participants changed cluster membership from fall to spring semesters. Thus, future research might also examine the factors that contribute to changes in motivational profile over time. The person-centered analyses presented above indicate that in an organizational setting, being more motivated and more self-determined is associated with favorable workplace perceptions and correlates, such as perceiving an enriched job and more social support, performing better, and experiencing greater need satisfaction. These results suggest that the different types of motivation posited by SDT are independent, and thus can occur in different amounts within each person (Ntoumanis, 2002). They also suggest that autonomous motivation is most important for the workplace, and that high levels of controlled motivation in the presence of high autonomous motivation do not negatively impact perceptions of job characteristics, social support, need satisfaction, or effectiveness on the job. Therefore, organizations should not fear having extrinsically motivated employees, so long as they are also autonomously motivated. References
Aldenderfer, M. S., & Blashfield, R. K. (1984). Cluster analysis. Beverly Hills, CA: Sage. Amabile, T. M. (1985). Motivation and creativity: Effects of motivational orientation on creative writers. Journal of Personality and Social Psychology, 48(2), 393399. Baard, P. P. (2002). Intrinsic need satisfaction in organizations: A motivational basis of success in for-profit and not-for-profit settings. In E. L. Deci, & R. M. Ryan (Eds.), Handbook of self-determination research (pp. 255275). Rochester: University of Rochester Press. Baard, P. P., Deci, E. L., & Ryan, R. M. (2004). Intrinsic need satisfaction: A motivational basis for performance and well-being in two work settings. Journal of Applied Social Psychology, 34, 20452068. Bergman, L. R., & Magnusson, D. (1997). A person-oriented approach in research on developmental psychopathology. Development and Psychopathology, 9(2), 291319. Boich, J. C. S., Sarrazin, P. G., Grouzet, F. M. E., Pelletier, L. G., & Chanal, J. P. (2008). Students' motivational profiles and achievement outcomes in physical education: A self-determination perspective. Journal of Educational Psychology, 100(3), 688701. Bono, J. E., & Judge, T. A. (2003). Self-concordance at work: Toward understanding the motivational effects of transformational leaders. Academy of Management Journal, 46(5), 554571. Brislin, R. W. (1986). The wording and translation of research instruments. In W. J. Lonner, & J. W. Berry (Eds.), Field methods in cross-cultural research (pp. 137164). Thousand Oaks, CA: Sage. Campion, M. A., Medsker, G. J., & Higgs, A. C. (1993). Relations between work group characteristics and effectiveness: Implications for designing effective work groups. Personnel Psychology, 46(4), 823850. Chantal, Y., Guay, F., Dobreva-Martinova, T., & Vallerand, R. J. (1996). Motivation and elite performance: An exploratory investigation with Bulgarian athletes. International Journal of Sport Psychology, 27(2), 173182. Deci, E. L., Koestner, R., & Ryan, R. M. (1999). A meta-analytic review of experiments examining the effects of extrinsic rewards on intrinsic motivation. Psychological Bulletin, 125(6), 627668. Deci, E. L., & Ryan, R. M. (2000). The what and why of goal pursuits: Human needs and the self-determination of behavior. Psychological Inquiry, 11(4), 227268.

C.M. Moran et al. / Journal of Vocational Behavior 81 (2012) 354363

363

Deci, E. L., Ryan, R. M., Gagn, M., Leone, D. R., Usunov, J., & Kornazheva, B. P. (2001). Need satisfaction, motivation, and well-being in the work organizations of a former Eastern Bloc country: A cross-cultural study of self-determination. Personality and Social Psychology Bulletin, 27, 930942. Frederick-Recascino, C. M. (2002). Self-determination theory and participation motivation research in the sport and exercise domain. In E. L. Deci, & R. M. Ryan (Eds.), Handbook of self-determination research (pp. 277294). Rochester: University of Rochester Press. Gagn, M., & Deci, E. L. (2005). Self-determination theory and work motivation. Journal of Organizational Behavior, 26, 331362. Gillet, N., Vallerand, R. J., & Rosnet, E. (2009). Motivational clusters and performance in a real-life setting. Motivation and Emotion, 33, 4962. Gower, J. C. (1967). A comparison of some methods of cluster analysis. Biometrics, 23, 623637. Grant, A. M. (2008). Does intrinsic motivation fuel the prosocial fire? Motivational synergy in predicting persistence, performance, and productivity. Journal of Applied Psychology, 93(1), 4858. Greguras, G. J., & Diefendorff, J. M. (2010). Why does proactive personality predict employee life satisfaction and work behaviors? A field investigation of the mediating role of the self-concordance model. Personnel Psychology, 63, 539560. Grolnick, W. S., & Ryan, R. M. (1987). Autonomy in children's learning: An experimental and individual difference investigation. Journal of Personality and Social Psychology, 52, 890898. Guay, F., Vallerand, R. J., & Blanchard, C. M. (2000). On the assessment of situational intrinsic and extrinsic motivation: The Situational Motivation Scale (SIMS). Motivation and Emotion, 24, 175213. Hackman, J. R., & Oldham, G. R. (1975). Development of the job diagnostic survey. Journal of Applied Psychology, 60, 159170. Hayenga, A. O., & Corpus, J. H. (2010). Profiles of intrinsic and extrinsic motivations: A person-centered approach to motivation and achievement in middle school. Motivation and Emotion, 34, 371383. Ilardi, B. C., Leone, D., Kasser, T., & Ryan, R. M. (1993). Employee and supervisor ratings of motivation: Main effects and discrepancies associated with job satisfaction and adjustment in a factory setting. Journal of Applied Social Psychology, 23, 17891905. Kasser, T., Davey, J., & Ryan, R. M. (1992). Motivation and employeesupervisor discrepancies in a psychiatric volitional rehabilitation setting. Rehabilitation Psychology, 37, 175188. Koestner, R., & Losier, G. F. (2002). Distinguishing three ways of being internally motivated: A closer look at introjections, identification, and intrinsic motivation. In E. L. Deci, & R. M. Ryan (Eds.), Handbook of self-determination research (pp. 101121). Rochester, NY: University of Rochester Press. Levesque, C., Sell, G. R., & Zimmerman, J. A. (2006). A theory-based integrative model for learning and motivation in higher education. In S. Chadwick-Blossey (Ed.), To improve the academy, Vol. 24. (pp. 86103)Bolton, MA: Anker Publishing. Liu, W. C., Wang, C. K. J., Tan, O. S., Koh, C., & Ee, J. (2009). A self-determination approach to understanding students' motivation in project work. Learning and Individual Differences, 19, 139145. Marsh, H. W., Ldtke, O., Trautwein, U., & Morin, A. (2009). Classical latent profile analysis of academic self-concept dimensions: Synergy of person- and variable-centered approaches to theoretical models of self-concept. Structural Equation Modeling: A Multidisciplinary Journal, 16(2), 191225. Miserandino, M. (1996). Children who do well in school: Individual differences in perceived competence and autonomy in above-average children. Journal of Educational Psychology, 88, 203214. Moosbrugger, H., Schermelleh-Engel, K., & Klein, A. (1997). Methodological problems of estimating latent interaction effects. Methods of Psychological Research, 2(2), 95111. Morgeson, F. P., & Humphrey, S. E. (2006). The work design questionnaire (WDQ): Developing and validating a comprehensive measure for assessing job design and the nature of work. Journal of Applied Psychology, 91(6), 13211339. Muthn, L. K., & Muthn, B. O. (19982011). Mplus v. 6.11. Los Angeles: Muthn & Muthn. Ntoumanis, N. (2002). Motivational clusters in a sample of British physical education classes. Psychology of Sport and Exercise, 3, 177194. Porter, L. W., & Lawler, E. E. (1968). Managerial attitudes and performance. Homewood, IL: R. D. Irwin. Ratelle, C. F., Guay, F., Vallerand, R. J., Larose, S., & Sencal, C. (2007). Autonomous, controlled, and amotivated types of academic motivation: A person-oriented analysis. Journal of Educational Psychology, 99(4), 734746. Reeve, J. (2002). Self-determination theory applied to educational settings. In E. L. Deci, & R. M. Ryan (Eds.), Handbook of self-determination research (pp. 184203). Rochester, NY: University of Rochester Press. Ryan, R. M., & Connell, J. P. (1989). Perceived locus of causality and internalization: Examining reasons for acting in two domains. Journal of Personality and Social Psychology, 57, 749761. Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 6878. Ryan, R. M., Deci, E. L., & Grolnick, W. S. (1995). Autonomy, relatedness, and the self: Their relation to development and psychopathology. In D. Cicchetti, & D. J. Cohen (Eds.), Developmental psychopathology. Theory and methods, Vol. 1. (pp. 618655) Oxford: John Wiley & Sons. Ryan, R. M., & Grolnick, W. S. (1986). Origins and pawns in the classroom: A self-report and projective assessment of children's perceptions. Journal of Personality and Social Psychology, 50, 550558. Sansone, C., & Harackiewicz, J. (Eds.). (2000). Intrinsic and extrinsic motivation: The search for optimal motivation and performance. San Diego, CA: Academic Press. Sheldon, K. M., & Elliot, A. J. (1998). Not all personal goals are personal: Comparing autonomous and controlled reasons for goals as predictors of effort and attainment. Personality and Social Psychology Bulletin, 24(5), 546557. Tukey, J. W. (1953). The problem of multiple comparisons. In H. I. Braun (Ed.), The collected works of John W. Tukey: Vol. VII: Multiple comparisons: 19481983 (pp. 1300). New York: Chapman & Hall. Vallerand, R. J., & Bissonette, R. (1992). On the predictive effect of intrinsic, extrinsic, and amotivational styles on behavior: A prospective study. Journal of Personality, 60, 599620. Vandenberg, R. J., & Lance, C. E. (2000). A review and synthesis of the measurement invariance literature: Suggestions, practices, and recommendations for organizational research. Organizational Research Methods, 3(1), 469. Wang, C. K. J., & Biddle, S. J. H. (2001). Young people's motivational profiles in physical activity: A cluster analysis. Journal of Sport & Exercise Psychology, 23, 122. Ward, J. (1963). Hierarchical grouping to optimize an objective function. Journal of the American Statistical Association, 58, 236244. Williams, L. J., & Anderson, S. E. (1991). Job satisfaction and organizational commitment as predictors of organizational citizenship and in-role behaviors. Journal of Management, 17(3), 601617.