“What Motivates You Doesn't Motivate Me”: Individual Differences in the Needs Satisfaction–Motivation Relationship of Romanian Volunteers



The present study examines individual differences in the relationship between two core concepts of the self-determination theory (SDT), namely satisfaction with the autonomy, competence, and relatedness need and motivation (autonomous vs. controlled). Based on the values component of SDT we hypothesised at least two different subpopulations with different need satisfaction–motivation patterns. Data from 349 Romanian volunteers revealed that two groups (or subpopulations) of volunteers can be distinguished, supporting our hypothesis. For the first and largest group, the pattern is in line with the SDT assumption that satisfaction of the autonomy and competence need has an effect on the autonomous forms of motivation. This group is in line with people endorsing intrinsic values. The second group of volunteers, however, revealed that satisfaction with the relatedness need links up with the controlled forms of motivation, and satisfaction of autonomy and competence needs does not predict autonomous motivation. This group is expected to favor extrinsic values. Both relationship patterns were further linked to work engagement and intention to quit, in order to shed light on the practical importance of the observed differences.


Since the beginnings of motivation theories (Maslow, 1954; McClelland, 1961) the concept of needs, defined as the content of motivation that provides the basis and direction for action, has grounded the studies on motivation. Starting around the mid-1960s, a new approach, which emphasised the process of motivation (e.g. Vroom, 1964), shifted the focus from needs towards goals selection and pursuit as dominant motivational concepts (Steers, Mowday, & Shapiro, 2004). Finally, in the last 30 years, Self-Determination Theory (SDT; Deci & Ryan, 1985) has reframed the concept of psychological needs by emphasising its role in both the content and the process of behavior.

SDT is a greatly studied metatheory of human motivation (Deci & Ryan, 1985, 2000) built on the assumption that all individuals have innate tendencies to develop towards a more elaborated self. SDT has evolved in a series of five interrelated mini-theories, each receiving a considerable amount of empirical support (e.g. Baard, Deci, & Ryan, 2004; Deci, Koestner, & Ryan, 1999; Ryan & Connell, 1989; Sheldon, Ryan, & Reis, 1996; van Beek, Hu, Schaufeli, Taris, & Schreurs, 2012; Van den Broeck, Schreurs, De Witte, Vansteenkiste, Germeys, & Schaufeli, 2011; Williams & Deci, 1996). In this article, we study individual differences in the relationship between elements from two of these mini-theories, namely the basic needs theory and the organismic integration theory, in a sample of Romanian volunteers.

Basic Psychological Needs and Organismic Integration

SDT argues that everyone has three basic psychological needs, that is, the need for autonomy, competence, and relatedness. Satisfaction of these needs provides the basis for optimal psychological growth and well-being of the individual (Deci & Ryan, 2000). The need for autonomy relates to a sense of personal choice, free will, and ownership of one's actions. The need for competence refers to the need to be effective in the interaction with one's environment and in the achievement of desired outcomes. The need for relatedness reflects the desire to have close relationships and to obtain a sense of communion. Contrary to other motivation theories that are concerned with need strength (e.g. McClelland, 1961, 1985), SDT focuses on (a) the extent to which the individual experiences satisfaction of these needs within various social environments and (b) the consequences of this variation in need satisfaction.

According to SDT, need satisfaction is closely linked to intrinsic motivation and the internalisation of external values (Deci & Ryan, 1985). Internalisation is an active process in which individuals transform social requests into self-regulation or values that are personally accepted. The internalisation process can be successful to different degrees, thus resulting in different forms of regulation, from less to fully self-determined behaviors. External regulation refers to the classical case of extrinsic motivation when people are acting to obtain desired consequences or to avoid punishment. Introjected regulation represents a partial, but only very limited form of internalisation in which people act to avoid guilt and shame or to obtain ego enhancements and feelings of worth. Identified regulation appears when people accept and recognise the value of a certain behavior; hence their behavior is more autonomous but still instrumental, as it serves to attain a personal goal. Integrated regulation represents a more complete form of internalisation and refers to integrating the activity with other aspects of one's values and identity. Finally, Intrinsic motivation emerges when people engage in an activity because they consider it interesting in itself and when they do so without any external demands or pressures. The four regulation types, plus intrinsic motivation, differ in the extent to which they represent either controlled or autonomous motivation. Thus, SDT distinguishes between controlled motivation (extrinsic and introjected regulations) and autonomous motivation (intrinsic motivation, identified and integrated regulations) (Deci & Ryan, 2000; Vansteenkiste, Lens, Dewitte, De Witte, & Deci, 2004; Bidee, Vantilborgh, Pepermans, Huybrechts, Willems, Jegers, & Hofmans, in press; Haivas, Hofmans, & Pepermans, in press b).

Both intrinsic motivation and the four types of regulation are linked to fulfillment of the basic needs (Deci & Ryan, 2002). In particular, the degree to which people engage in interesting activities or are able to internalise external demands, values, and regulation, is mainly a function of the extent to which the activities allow the autonomy, competence, and relatedness needs to be satisfied. Yet, at present research on this crucial link between needs satisfaction and the different types of motivation is subject to two major limitations.

First, most research on SDT studies need satisfaction as an aggregated construct (e.g. Baard et al., 2004; Deci, Ryan, Gagné, Leone, Usunov, & Kornazheva, 2001; Gagné, 2003). For this reason, the link between satisfaction of the autonomy, competence, and relatedness need on the one hand, and the different types of motivation on the other is not often tested. The few studies that have differentiated between the three needs showed that separating the needs reveals interesting independent contributions (Boezeman & Ellemers, 2009; Greguras & Diefendorff, 2009; Haivas, Hofmans, & Pepermans, in press a; Lynch, Plant, & Ryan, 2005). By differentiating between the three needs, these studies allow for a better understanding of the role of need satisfaction in the internalisation process. In particular, it could be tested whether the three needs are all equally important for full internalisation, which should not be the case according to SDT. Whereas feelings of competence and autonomy contribute to the facilitation of full internalisation, satisfaction of the relatedness need is not sufficient to ensure full internalisation of activities (Deci & Ryan, 2000; Gagné & Deci, 2005). Thus, the first contribution of the present paper is to study the relationships between satisfaction of the autonomy, competence, and relatedness needs, and the different types of motivation. In line with SDT, we hypothesise that satisfaction of the competence and autonomy needs, in particular, will relate to the more autonomous forms of motivation.

A second limitation is that most of the previous studies that relate needs satisfaction to motivation do not pay attention to the fact that these relationships may differ between individuals. Note that we do not suggest that individual differences have been neglected in former research on SDT. Rather, these former studies have focused either on individual differences in regulatory styles (Vallerand & Ratelle, 2002; Losier, Perreault, Koestner, & Vallerand, 2001) or on the way people are oriented towards the social context, referred to as causality orientations (Deci & Ryan, 1985; Lam & Gurland, 2008; Williams, Grow, Freedman, Ryan, & Deci, 1996; Gagné, 2003). Basically, and in line with our argumentation, both approaches suggest that not all individuals react in the same way. However, they both relate individual differences to the way single variables are perceived or assessed and not to the interrelationships between those variables. Indeed, whereas previous research acknowledges that people may differ in the satisfaction of their needs and therefore also in their type of motivation, the relationship between need satisfaction and motivation is implicitly assumed to be similar for everyone. In other words, each individual is supposed to be well characterised by an “average pattern”, thereby ignoring individual deviances from that average. Recent research, although still scarce, has started to question the dominant paradigm that generalises the universality of such psychological relationship patterns (e.g. De Gieter, Hofmans, & Pepermans, 2011). In addition, a number of theoretical arguments suggest that individual differences in these relationships are an empirical reality.

Individual Differences

First, it has been found that different individuals display different motivational profiles (a profile being a specific combination of autonomous and controlled motivation), even within the same situation or for the same activity (Gillet, Vallerand, & Rosnet, 2009; Ratelle, Guay, Vallerand, Larose, & Senécal, 2007; Vansteenkiste, Sierens, Soenens, Luyckx, & Lens, 2009). In other words, even when the environment provides the same means for satisfying the three needs, people show different types of motivation. Such individually based variations may be due to the fact that people differ in the way satisfaction of their needs relates to motivation, and that these interindividual differences lie at the basis of differences in motivational profiles.

Second, Kasser's (2002) self-determination theory of values offers some insights that support our questioning of the universality of certain relationships between SDT-related variables. It is suggested that individuals seek out activities that they value, with the valuing process depending on the current state of the individual, making it a dynamic process. Kasser (2002) suggests two sets of values, i.e. intrinsic and extrinsic ones (see also Van den Broeck, Vansteenkiste, Lens, & De Witte, 2010). As Kasser mentions that “some values are conducive to growthful, intrinsically motivated actions, others tend to prompt extrinsically motivated behaviors, focused on rewards and people's praise” (Kasser, 2002, pp. 127–128), intrinsic values are more likely to satisfy the self's needs than extrinsic values. In line with this reasoning, Sheldon and Kasser (1995) showed that people valuing intrinsic goals were more autonomy oriented, while people relying on extrinsic goals were more control oriented, even experiencing different relatedness need satisfactions (Kasser & Ryan, 2001; Sheldon & Kasser, 1995). The latter were even “likely to ignore their needs and to engage in activities which work against their health and well-being” (Schmuck, Kasser, & Ryan, 2000, p. 226). This suggests that, depending on the values an individual endorses at a particular moment, the relationship between the three needs and the different types of motivation may be different.

Third, the functional theory of Clary, Snyder, Ridge, Copeland, Stukas, Haugen, and Moiene (1998) advances six possible functions that can be served through volunteer activities: (1) expressing personal values such as altruism or humanitarian concerns (i.e. values); (2) expanding knowledge, skills, and abilities (i.e. understanding); (3) increasing one's social network or building a favorable image in front of relevant others (i.e. social); (4) extending personal career related networks (i.e. career); (5) protecting one's ego and reducing potential negative feelings (i.e. protective); (6) increasing personal development and esteem enhancement (i.e. enhancement). As can readily be seen, all these functions can be linked to either intrinsic (1, 2, 6) or extrinsic values (3, 4, 5). As such, this approach further supports the idea that there exists a (dichotomous) differentiation between why and how people get motivated to volunteer.

In short, previous research suggests that there are individual differences in the way need satisfactions relate to the different types of motivation. In line with Kasser's (2002) self-determination theory of values, we expect to find two groups of volunteers: (1) one group for which satisfaction of the needs (and especially the competence and autonomy needs) relates to the more autonomous forms of motivation (i.e. people who endorse intrinsic values); and (2) one group for which satisfaction of the needs is unrelated to the more autonomous forms of motivation (i.e. people who endorse extrinsic values).

To study the potential impact of the hypothesised individual differences, subgroup-related relationships will further be linked to two important organisational outcomes, namely volunteers' work engagement and intention to quit the (volunteering) organisation (Gagné, 2003; Grube & Piliavin, 2000; Millette & Gagné, 2008). Psychological engagement in one's work or activities is often used as an important indicator that relates to personal involvement, triggered by autonomous motivation in paid (e.g. Baard et al., 2004; Deci et al., 2001) as well as in voluntary work (e.g. Koestner, Franz, & Weinberger, 1990; Gagné, 2003). Intention to quit is considered a good predictor of actual turnover, even in the volunteering sector (Omoto & Snyder, 1995) where retaining volunteers is a well-known practical consideration of present-day volunteering organisations (Bussell & Forbes, 2002; Vantilborgh, Bidee, Pepermans, Willems, Huybrechts, & Jegers, 2011).



Participants were 349 Romanian volunteers aged between 18 and 58 years (mean age 22.9 years; SD = 4.8) and 61.3 per cent were female. With respect to professional status, 28.7 per cent had paid jobs and 71.3 per cent were students. With respect to the level of education, 0.3 per cent of the participants had completed primary school, 47 per cent had completed secondary education, 44 per cent had acquired a professional school or bachelor degree and 8.7 per cent had obtained a masters or PhD degree. This sample was representative of Romanian volunteers in several aspects: the majority of volunteers in Romania are female (two-thirds female, one-third male); they are mostly young, that is, within the age range 19–25 years, and are highly educated, with more than half currently enrolled in an educational program (Rigman, 2009; European Union, 2010).


To obtain a representative volunteer sample, we recruited from 10 non-governmental organisations operating in the social domain (61.3% of respondents) and in the educational domain (38.7% of respondents). The questionnaires were distributed at the end of regular meetings in the respective organisations where volunteers are briefed about new and ongoing volunteering projects. This was accompanied by a short verbal clarification of the study, where special emphasis was put on confidentiality, anonymity of responses, and voluntary participation. Our procedure resulted in a response rate close to 100 per cent (a small number of questionnaires could not be used due to incompleteness).


The questionnaire included four measurement scales (Motivation at Work Scale-R, Basic Needs Satisfaction at Work Scale, Volunteers' Work Engagement and Intention to Quit). In addition, demographic questions concerning age, gender, level of education, and professional status were included in the questionnaire. The four measurement scales were translated from English into Romanian by a professional translator. An English back-translation was done by another translator and this back-translation was compared with the original version of the scales. Remaining and additional ambiguities were solved by a team of four psychologists who mastered both English and Romanian.

Motivation at Work Scale-R (MAWS-R)

Motivation for voluntary work was measured using the Motivation at Work Scale-R (Gagné et al., 2012). In this scale, participants were asked to rate the reasons for engaging in volunteer work according to the different forms of autonomous (and controlled) motivation, on a scale ranging from 1 (not at all because of this reason) to 7 (exactly because of this reason). The scale measures the five forms of motivation: external regulation (four items, e.g. “Because others put pressure on me”), introjected regulation (four items, e.g. “Because it makes me feel proud of myself”), identified regulation (four items, e.g. “Because what I do in this job has a lot of personal meaning for me”), integrated regulation (four items, e.g. “Because I am made for this type of work”), and intrinsic motivation (four items, e.g. “Because I enjoy this work very much”). Consistent with SDT and previous studies (Deci & Ryan, 2000; Vansteenkiste et al., 2004), we used the distinction between controlled motivation, composed of the items that measure external and introjected regulation (α = 0.68), and autonomous motivation, composed of items that measure identified and integrated regulation and intrinsic motivation (α = 0.89) for subsequent analyses.

Basic Needs Satisfaction at Work Scale

We adapted the need satisfaction scale developed by Deci and Ryan (2000) to the volunteer context by substituting the words “work” and “job”, respectively, with “work as volunteer” or “volunteering activities”. In total, 20 items were used to assess autonomy satisfaction (six items, e.g. “I am free to express my ideas and opinions when I work as a volunteer”, α = 0.72), competence satisfaction (six items, e.g. “I don't feel very competent at my volunteering activities”, reversed coded, α = 0.62), and relatedness satisfaction (eight items, e.g. “I really like the people I am volunteering with”, α = 0.74). All items had to be answered on a 7-point scale ranging from 1 (totally disagree) to 7 (totally agree). One item from the original autonomy satisfaction subscale was deleted because of its decreasing impact on Cronbach's α (from α = 0.72 to α = 0.60; “When I am volunteering, I have to do what I am told”).

Volunteers' Work Engagement

Work engagement was measured by the adapted version of the Volunteers' Work Engagement Scale proposed by Gagné (2003) which measures behavioral and emotional engagement in volunteer work on a 7-point scale, from 1 (totally disagree) to 7 (totally agree). Example items are: “I really like to devote myself to my volunteer work” or “When I am volunteering, I often feel bored” (reversed). Cronbach's α of the 12-item scale was 0.78.

Intention to Quit

Volunteers rated the following two questions on a scale from 1 (totally disagree) to 7 (totally agree): “I frequently think about leaving this organisation” and “It is likely that I will leave this organisation in the next year”, which is based on earlier measures of turnover intention (e.g. Currall, Towler, Judge, & Kohn, 2005). Combining both questions resulted in a Cronbach's α of 0.74.


In a first analysis, we fitted a path model in which the satisfaction with the three basic psychological needs (autonomy, competence, and relatedness) predicts the different types of regulation.

Second, to test whether individual differences exist in the relationship between satisfaction of the three basic needs and motivation for performing volunteering activities, we relied on a statistical method called mixture path analysis (MPA). Similar to the traditional path analysis model, a series of dependent variables is modeled as a function of a set of predictors. The major difference, however, is that it is assumed that the data originate from various (unknown or latent) clusters or groups, with each of these clusters showing different relationships in the path model (DeSarbo & Cron, 1988). MPA identifies these clusters by simultaneously estimating (1) cluster memberships and (2) a separate path model per cluster. In particular, for each individual in the data set, the MPA model estimates the probability Pij that person i belongs to cluster j. These probabilities can then further be used to cluster the persons by assigning each person to the cluster with maximum probability. For each of these clusters we can then establish the path model. As the number of clusters is unknown, the most commonly used approach is to fit models with an increasing number of clusters and compare their model fit (e.g. Brusco, Cradit, Steinley, & Fox, 2008). In this study, MPA models with one to five clusters were fitted. Moreover, for each model, 100 random restarts were performed in order to minimise the risk of converging to a local maximum (e.g. Leisch, 2004). To compare these models, we made use of the Bayesian Information Criterion (BIC), which is an information criterion that balances model fit and parsimony (Schwarz, 1978). The BIC is generally considered the best index to compare mixture models (Leisch, 2004). In general, a model with a lower BIC is preferred over a model with a higher value. Our analyses were carried out in Mplus (Muthén & Muthén, 2010), a statistical program that allows fitting a wide range of finite mixture models.

In a final step, independent samples t-tests were performed to test whether the clusters of volunteers differed on work engagement and intention to quit.


Overall Analysis

Table 1 shows the means, standard deviations, and correlation coefficients for each variable. All variables are measured on a 1 to 7 scale.

Table 1. Means, Standard Deviations, and Correlations of all Variables
  1. ** p <.01; *** p <.001.
 1. Autonomy Need5.40.99         
 2. Competence Need5.40.93.65**        
 3. Relatedness Need5.44.93.73**.62**       
 4. Extrinsic1.96.89−.16**−.17**−.12**      
 5. Introjected3.671.02.12*.15**.11*.39**     
 6. Identified5.241.10.32**.27**.23**.03.45**    
 7. Integrated4.551.27.36**.35**.31**.10.47**.71**   
 8. Intrinsic5.281.05.35**.32**.29**.02.40**.65**.65**  
 9. Engagement5.11.78.57**.61**.46**−.27**.11**.37**.36**.31** 
10. Intention to quit1.861.32−.40**−.37**−.34**.02−.16*−.16**−.17**−.18**−.43**

Subsequently, we fitted a path model in which the autonomy, competence, and relatedness needs predict the different types of regulation. From Table 2 it can be seen that satisfaction of the autonomy and competence needs relates to the autonomous forms of motivation, whereas satisfaction of the relatedness need does not. As satisfaction of the relatedness need was significantly related to all forms of motivation when computing the zero-order correlations (see Table 1), this result implies that, when accounting for the effects of satisfaction of the autonomy and competence needs, satisfaction of the relatedness need no longer explains variance in the autonomous forms of motivation. In other words, the effect of relatedness on the autonomous forms of motivation is already accounted for by the autonomy and competence needs.

Table 2. Path Coefficients for the Relationship between Autonomy, Competence, and Relatedness Need Satisfaction and the Extrinsic, Introjected, Identified, Integrated, and Intrinsic Regulation for the Whole Sample
 Controlled motivationAutonomous motivation
  1. * p <.05; ** p <.01; *** p <.001.

Individual Differences Analysis

To test for individual differences in the relationship between basic need satisfaction and motivation, mixture path models with one to five subpopulations were fitted. Moreover, for each model, 100 random restarts were performed. In Figure 1 the best BIC out of the 100 restarts is plotted for each of the five mixture models. Based on this information, we decided to retain the model with two mixture components.

Figure 1.

BIC as a function of the number of subpopulations in the mixture path model.

It appeared that the data can best be described by a model including two subpopulations of 293 (i.e. Subpopulation 1 – SP1) and 56 (i.e. Subpopulation 2 – SP2) volunteers, respectively. When comparing the path coefficients for both subpopulations (see Table 3), it can be seen that satisfaction of the autonomy and competence needs has an effect on the autonomous forms of motivation for SP 1, whereas for individuals belonging to SP 2, satisfaction of autonomy and competence needs hardly predict any form of autonomous motivation. In contrast, relatedness need satisfaction relates positively to the controlled forms of motivation in SP 2.

Table 3. Path Coefficients for the Relationship between Autonomy, Competence, and Relatedness Need Satisfaction and the Extrinsic, Introjected, Identified, Integrated, and Intrinsic Regulation for Subpopulations 1 (SP1) and 2 (SP2)
  Controlled motivationAutonomous motivation
  1. * p <.05; ** p <.01; *** p <.001.

Finally, in order to obtain a more detailed picture of the volunteers constituting the two subpopulations, we tested for between-subpopulation differences using independent samples t-tests (see Table 4). First, there were no significant differences according to gender, professional status (student or employed), age, or the type of NGO (educational vs. social). Second, compared to volunteers from SP2, volunteers belonging to SP1 have a significantly higher satisfaction of their three needs. Conversely, the latter group scores significantly higher on controlled motivation than individuals from SP1.

Table 4. Mean Differences between the Two Subpopulations (SP1 and SP2)
 SP 1 (n = 293)SP 2 (n = 56)t-valuep-value
Needs Satisfaction5.574.896.367<.001
Autonomous Mot.>.05
Controlled Mot.2.653.67−10.037<.001
Work Engagement5.204.654.930<.001
Intention to quit1.792.19−2.058<.05

Second, we also evaluated whether membership of one of the subpopulations relates differently to volunteers' work engagement and intention to quit the organisation. As can be seen in Table 4, the results showed that individuals belonging to SP1 score significantly higher on work engagement and lower on intention to quit than individuals of SP2 (see Table 4).


The present study presents a thorough assessment of the complex relationship between two fundamental elements of SDT, namely that between basic psychological needs (autonomy, competence, and relatedness) and the different regulatory types. First of all, and in line with our expectations, the results show that the relatedness need is indeed less relevant for full internalisation of the regulations. In particular, our results show that, when accounting for autonomy and competence need satisfaction, satisfaction of the relatedness need has no additional explanatory power for predicting variation in the regulatory types. Second, and most important, the results reveal that there are major individual differences in the relationship between the basic psychological needs and the different regulatory types. These individual differences can best be summarised using two subpopulations, each with a different need satisfaction–motivation relationship. In line with our expectations, we find that for one subpopulation (i.e. SP1), satisfaction of the autonomy and competence needs relates to facilitation of autonomous motivation, whereas for the other subpopulation (i.e. SP2), satisfaction of the autonomy and competence needs is unrelated to autonomous motivation. The observation that the first subgroup is composed of the largest number of volunteers (n = 293) has two important implications. First, it implies that the SDT assumption which states that satisfaction of the autonomy and competence needs is more central to autonomous motivation than satisfaction of the relatedness need holds for the majority of individuals. Second, it also indicates why, if no individual differences are taken into account, research often supports this SDT premise, i.e. because the results are “biased” towards the results of the largest group.

An explanation of the observed individual differences in the need satisfaction–motivation link can be found in the theory of values orientation. Kasser and Ryan (1996) propose two types of values: intrinsic and extrinsic ones. Intrinsic values are desires that are congruent with actualising and growth tendencies, while extrinsic values do not provide satisfaction in themselves: they focus on the power or sense of worth that can be derived from attaining them. It is likely that intrinsic values, rather than extrinsic ones, will more easily satisfy our basic needs. Indeed, it has been found that people oriented towards intrinsic values are likely to differ from those oriented towards extrinsic values in the way they experience need satisfaction (Kasser & Ryan, 2001; Schmuck et al., 2000). Thus, we suggest that volunteers from SP1 place greater importance on intrinsic values, such as concern for growth, autonomy, or self-regard. In accordance with Kasser (2002), this determines that they will engage in behaviors that help them satisfy their basic needs, such as performing volunteering activities, for which they will display autonomous forms of motivation. In contrast, and also in line with Kasser's (2002) self-determination theory of values, we suggest that volunteers belonging to SP2 place greater importance on extrinsic values and therefore get involved in volunteering activities to obtain appreciation and affection from others and get their relatedness need satisfied. Hence, they perform behaviors based on what other people value, which is reflected in the higher level of controlled motivation.

The relations that are displayed by the two volunteering subpopulations suggest that there are significant interindividual differences in how the satisfaction of the autonomy, competence, and relatedness needs are aligned when one is extrinsically and introjected motivated versus identified, integrated, or intrinsically motivated. From an SDT perspective, all people are naturally inclined to develop towards greater autonomy (integration with the self) and greater relatedness (integration with the social community), and the three needs together enhance the internalisation of activities performed. However, it is also assumed within SDT that incompatibility between the needs might arise, that is, when the social context structure turns the needs against each other. In our particular study, the results could be rooted both in the specifics of the volunteers' activity domain and in the specifics of the Romanian culture. First, it might be that volunteers belonging to SP2 engage in volunteering mainly because it creates a social context to be with one's friends, and because it offers the opportunity to be viewed favorably by other important people (see also Snyder & Omoto, 2008). Second, the volunteering population in Romania is fairly distinct as it has been formed in the last two decades in a social climate less conductive to social participation, responsibility for self, and autonomy (Voicu, 2001). Therefore, it mainly consists of young and educated people from urban areas. Thus, young Romanian volunteers might be more interested in satisfying their autonomy and competence needs and less in connecting with others, maybe due to the fact that the relatedness need might already be satisfied during other day-to-day youth activities. In addition, it can be suggested that the collectivistic features of the Romanian culture might also allow volunteers to satisfy their relationship need by means of particularly strong connections with their families and other social groups.

From our results it appears that the group differences relate to important differences in certain organisational behaviors by volunteers, namely, work engagement and intention to quit. Members of SP1, that is, volunteers for whom autonomy and competence need satisfaction predicts autonomous motivation, experience a stronger satisfaction for the three basic needs and a stronger work engagement as well as a lower intention to quit. This is in line with other SDT studies, also in the voluntary sector (e.g. Baard et al., 2004; Van den Broeck, Vansteenkiste, De Witte, & Lens, 2008; Vansteenkiste, Neyrinck, Niemic, Soenens, De Witte, & Van den Broeck, 2007; Gagné, 2003), and provides support for the SDT assumption that satisfaction of the basic needs results in positive outcomes through the facilitation of self-determined behavior. It further emphasises the importance of taking into account individual differences when studying the factors that influence volunteers' organisational behavior. The fact that no differences in demographic variables are found between the two subpopulations may be due to the rather homogeneous nature of the sample, consisting of mainly young educated volunteers.

In addition to the theoretical implications, our study tentatively suggests that when designing retention and motivational programs for volunteers, organisations ought to consider subpopulation differences and design practices that address both groups of volunteers. For volunteers belonging to the first subpopulation, it is important to provide tasks with optimal challenges, free choices, and constant feedback in order to satisfy their autonomy and competence needs, while for the second subpopulation it is of key importance to foster social interactions between volunteers or with beneficiaries, in order to satisfy the volunteers' relatedness need.

Several limitations of the present study should be noted. First, it is evident that our findings on volunteers cannot be generalised to paid workers. This is due to differences in the task structure and available resources, as mentioned in the introduction to this paper, and also because paid workers and volunteers differ in work attitudes (De Cooman, De Gieter, Pepermans, & Jegers, 2011). Second, our sample of volunteers comes from an East European country with a specific social background, and hence, a replication of the study in another non-Western country with a comparable social background is recommended, as would be replications in Western countries.

Despite these limitations, this study is the first one to reveal individual differences in one of the core relationships of SDT. The results show that an individual differences analysis can reveal additional insights when investigating the relationship between individual need satisfaction and motivation. Yet, further research is needed in order gain a more complete picture of the magnitude and consequences of individual differences in predicting motivational behavior for volunteers as well as for other populations.