SEARCH

SEARCH BY CITATION

Keywords:

  • need satisfaction;
  • affect;
  • autonomous motivation;
  • self-determination theory

Abstract

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

The relationship between need satisfaction and motivation is well established within self-determination theory (SDT). However, less is known about the affective mechanism that underlies this relationship. In this study, we extend SDT by focusing on the exact role of affect in the need satisfaction–motivation relationship. To this end, we conducted a daily diary study (= 72) and an experience sampling study (= 37) in which we tested the mediating role of positive and negative affect in the relationship between satisfaction of the autonomy, competence, and relatedness need on the one hand and autonomous motivation on the other hand. Moreover, alternative models were tested. The results of both studies demonstrated that affect did mediate the need satisfaction–intrinsic motivation relationship. Implications for theory and practice are discussed.

Practitioner points

  • Organizations can influence the intrinsic motivation of employees by changing working conditions to fulfil employees' needs.
  • Organizations can influence intrinsic motivation by changing the appraisals of working conditions.

Background

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

One of the most comprehensive motivation theories is without doubt the self-determination theory (SDT; Deci & Ryan, 1985; Gagné & Deci, 2005; Sheldon, Turban, Brown, Barrick, & Judge, 2003). This metatheory of motivation combines several aspects of the more traditional theories and is supported by a wide range of empirical studies (e.g., Haivas, Hofmans, & Pepermans, 2013; Lam & Gurland, 2008; Van den Broeck, Vansteenkiste, De Witte, & Lens, 2008; Van den Broeck, Vansteenkiste, De Witte, Soenens, & Lens, 2010). A central element in SDT is the relationship between need satisfaction and motivation. Although this relationship is well established, less is known about the affective mechanisms that underlie it. In the present study, we will tackle this limitation by focusing on the exact role of affect in the need satisfaction–motivation relationship.

Self-determination theory distinguishes two broad categories of work motivation: Controlled and autonomous motivation on the one hand and amotivation or demotivation on the other hand. When motivated in a controlled way, employees perform certain behaviours because they are told to do so or because they want to avoid feelings of shame and guilt (Deci & Ryan, 2002, 2008; Ryan & Deci, 2000). Controlled motivation includes two subtypes, namely external and introjected motivation. When employees are externally motivated, they engage in activities because of external reasons, such as to receive a reward or to avoid punishment. In contrast, employees who are motivated in an introjected way do things because of internal, yet instrumental reasons, such as avoiding feelings of worthlessness, shame, or guilt (Van den Broeck, Vansteenkiste, & De Witte, 2008). As opposed to controlled motivation, autonomous motivation implies that employees perform the behaviour with a sense of volition, that is, people engage in the activity because it is in line with their own interests and values (Gagné & Deci, 2005; Ryan & Deci, 2000). Autonomous motivation can further be divided into two subtypes, namely identified and intrinsic motivation. Identified motivation pertains to being motivated because of the fit with personal goals. Note that this type of motivation is still instrumental as people engage in activities because they are important for their personal goals. Intrinsic motivation in turn refers to performing the activity because it is interesting and enjoyable in itself.

Because autonomous motivation relates to a wide range of positive outcomes such as better psychological well-being, more effective performance, and greater persistence (Deci & Ryan, 2008), it is highly desirable in a work context. Moreover, Judge, Bono, Erez, and Locke (2005) have demonstrated that autonomous motivation is linked to affect-related concepts such as job satisfaction and life satisfaction, whereas controlled motivation does not relate to these outcomes. As the focus in this study is on the role of affect in the need satisfaction–motivation relationship, we chose to concentrate on autonomous motivation only.

According to SDT, the degree of autonomous (and controlled) motivation varies between and within individuals and depends on the fulfilment or thwarting of three basic psychological needs, that is, the need for autonomy, competence, and relatedness. The need for autonomy refers to people's desire to behave according to their own interests, to make their own choices, and to initiate their own behaviour. For example, a setting in which employees have the freedom to decide how and when their work can be done is beneficial for the need for autonomy. The need for competence refers to the desire to feel capable and effective. Hence, a task that is challenging may serve as a nutrient for this need. Finally, the need for relatedness refers to the necessity to be connected to others and to feel accepted by them. Pleasant contacts with the colleagues and superiors will therefore serve this need (Deci & Ryan, 2000). In general, SDT research has shown that fulfilment of the three basic psychological needs by the environment results in more autonomous motivation. This has been shown in a wide variety of contexts, such as sports (e.g., Ryan, Williams, Patrick, & Deci, 2009), work (e.g., Haivas et al., 2013; Lam & Gurland, 2008; Van den Broeck et al., 2011), and education (e.g., Ryan & Deci, 2009). Conversely, thwarting of the three basic psychological needs would lead to less autonomous motivation (Ryan & Deci, 2000).

Despite a large amount of studies that show that satisfaction of the three basic psychological needs relates positively to autonomous motivation, little is known about the (affective) mechanisms that underlie this relationship. The reason for this is that affect is hypothesized to play a distal role within SDT, that is, it is traditionally not considered to be part of the motivation-generative mechanism (Deci & Ryan, 1987). As a consequence, studies within the SDT framework have mainly focused on affect as an outcome of either need satisfaction or motivation. For example, Reis, Sheldon, Gable, Roscoe, and Ryan (2000), Sheldon and Bettencourt (2002), and Véronneau, Koestner, and Abela (2005) have studied affect as an outcome of satisfaction of the three basic psychological needs, whereas Carbonneau, Vallerand, and Lafrenière (2012) and Walls and Little (2005) have looked at affect as an outcome of autonomous motivation. However, the fact that SDT does not formally treat affect as a variable of interest in the motivation-generative mechanism is at odds with a large number of studies in the emotion literature that have demonstrated that affect is important in the coming about of motivation (e.g., Custers & Aarts, 2005; Frijda, 2008; Frijda, Kuipers, & ter Schure, 1989; Isen & Reeve, 2006; Martin, Ward, Achee, & Wyer, 1993; Meyer & Turner, 2006; Raghunathan & Tuan Pham, 1999; Scherer, 2004). Indeed, Isen and Reeve (2006) have demonstrated that the experience of positive affect increased people's intrinsic motivation towards activities, both at the behavioural and experiential level. Building on these findings, Isen and Reeve (2006) concluded that: ‘affect may play a more central role in understanding intrinsic motivational process than is currently recognized by self-determination theory’ (p. 321).

In the present study, we will contribute to the research on SDT by focusing on the exact role of both positive and negative affect in the need satisfaction–motivation relationship. In particular, our study goes beyond previous studies by testing the effects of positive and negative affect as mediators of the relationship between basic psychological need satisfaction and motivation. There are strong theoretical reasons to expect such mediation effects. First, one of the basic assumptions of SDT is that satisfaction (thwarting) of the basic psychological needs leads to positive (negative) well-being (Ryan & Deci, 2000). That is, factors that facilitate autonomy, competence, and relatedness should enhance well-being, while factors that detract from the satisfaction of these needs are thought to undermine well-being (Reis et al., 2000). The reason is that, when satisfied, the three needs offer different kinds of psychological rewards (Sheldon, Ryan, & Reis, 1996). In particular, autonomy implicates that one's goals and activities are in line with intrinsic interests and values. Competence involves the feeling that one can act effectively (Sheldon et al., 1996), and relatedness involves the feeling that one is close and connected to significant others (Reis et al., 2000). According to SDT, all three of the basic psychological needs have distinguishable effects and are essential for well-being, that is, ‘individuals cannot thrive without satisfying all of them, any more than people can thrive with water but not food’ (Ryan & Deci, 2000, p. 75). In line with this, Sheldon and Bettencourt (2002) have demonstrated that the satisfaction of the need for autonomy and relatedness was positively related to positive affect and that the satisfaction of the need for autonomy was negatively related to negative affect. In another study, Quested and Duda (2009) found that the satisfaction of the need for competence was positively related to positive affect and negatively to negative affect. Reis et al. (2000) showed that the satisfaction of the three needs was positively related to positive affect, while satisfaction of the need for autonomy and competence were negatively related to negative affect. Finally, Gillet, Fouquereau, Forest, Brunault, and Colombat (2012) demonstrated that the satisfaction of the three needs is related to more positive affect, while thwarting of the three needs is linked with less positive affect. Second, regarding the link between affect and motivation, several studies in the emotion domain have shown that affect has an impact on motivation (Custers & Aarts, 2005; Frijda, 2008; Frijda et al., 1989; Isen & Reeve, 2006; Martin et al., 1993; Meyer & Turner, 2006; Raghunathan & Tuan Pham, 1999; Scherer, 2004). For the specific case of autonomous motivation, it has been shown that positive affect increases people's enjoyment of and interest in mildly interesting activities because it increases the valence and evaluation of these activities (Erez & Isen, 2002; Kahn & Isen, 1993; Kraiger, Billings, & Isen, 1989). Apart from influencing the expectations about a task, positive affect also enhances the experience of interest, enjoyment, and sense of satisfaction derived from the activity during the execution of these activities (Erez & Isen, 2002; Isen, Daubman, & Nowicki, 1987; Staw & Barsade, 1993). Because the central component of intrinsic motivation pertains to inherent positive affective feelings associated with the execution of the task itself, an increase in positive affectivity should relate to an increase in autonomous motivation. Negative affect, on the other hand, has been found to relate negatively to intrinsic motivation. Meyer and Turner (2006) argue that, when one experiences negative affect, one will be amotivated and try to escape his/her work. Moreover, negative affect ‘tends to be incompatible with enjoyment as implied by interest and intrinsic motivation’ (Pekrun, Goetz, Titz, & Perry, 2002, p. 97). It is therefore expected that negative affect decreases autonomous motivation. In summary, we hypothesize that:

Hypotheses 1a and 1b: The relationship between satisfaction of the need for autonomy and autonomous motivation is positively mediated by positive (H1a) and negatively by negative affect (H1b).

Hypotheses 2a and 2b: The relationship between satisfaction of the need for competence and autonomous motivation is positively mediated by positive (H2a) and negatively by negative affect (H2b).

Hypotheses 3a and 3b: The relationship between satisfaction of the need for relatedness and autonomous motivation is mediated positively by positive (H3a) and negatively by negative affect (H3b).

As need satisfaction, motivation, and affect are highly dynamical concepts (Beal & Ghandour, 2011; Gooty, Gavin, & Ashkanasy, 2009; Reis et al., 2000), we focus on how they (dynamically) relate within an individual. This treatment of need satisfaction and motivation is novel in the research on SDT, which has traditionally relied on cross-sectional studies. However, the problem with cross-sectional studies is that they do not allow to capture dynamic, within-person relationships. To overcome this limitation, we studied the mediating role of affect in the need satisfaction–motivation relationship by means of a daily diary study and an experience sampling study.

STUDY 1

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

In this first study, we tested the mediating effect of positive and negative affect on the need satisfaction–motivation relationship by means of a daily diary study. In this study, we focused on intrinsic motivation only, as this is the most self-determined type of motivation. Moreover, intrinsic motivation is linked to a wide range of positive outcomes such as vitality and well-being (Gagné, Ryan, & Bargmann, 2003), and cognitive engagement (Walker, Greene, & Mansell, 2006). Also in the workplace, intrinsic motivation relates to desired outcomes such as more effective performance (Gagné & Deci, 2005), higher work effort (Bidee et al., 2013), and more knowledge sharing (Lin, 2007).

Method

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

Participants

Seventy-two Belgian employees and one trainee, 38 men and 34 women, of 22 different companies took part in this study. The mean age was 37.06 years (SD = 10.79). The majority of the participants were full-time employed (97.2%).

Procedure

At the end of each day, for 10 consecutive working days, participants received an email with a link to an online survey. In this survey, participants first had to think about one task they performed during the past working day. To facilitate memory, participants had to describe this task using keywords. Next, participants had to answer questions pertaining to the satisfaction of their basic psychological needs, intrinsic motivation, and positive and negative affect following from the execution of this specific task. In total, 522 responses were collected, which corresponds to 7.16 answers per person or a response rate of 72.29%.

Measures

All questions were selected from valid and reliable questionnaires. The English questions were translated into Dutch, and they were slightly changed to measure experiences referring to a single task. All items were answered on a 6-point rating scale, ranging from totally disagree to totally agree.

The basic need satisfaction in general scale (Johnston & Finney, 2010) was used to measure satisfaction of the three basic psychological needs (see also Gagné, 2003; Vansteenkiste, Lens, Soenens, & Luyckx, 2006; Wei, Philip, Shaffer, Young, & Zakalik, 2005). To lower the response burden on the participants, the two items with the highest factor loadings were selected to measure satisfaction of the need for competence (‘People at work told you that you worked good at your task’; ‘Working on your task gave you a sense of accomplishment’), satisfaction of the need for relatedness (‘While you were completing your working task, your colleagues were friendly towards you’; ‘During this task, you got along with the people you had to work with’), and satisfaction of the need for autonomy (‘You were free to express your ideas and opinions during this task’; ‘During this task, you felt you could pretty much be yourself’).

Positive and negative affect were measured with the Dutch questionnaire on the experience and evaluation of work (QEEW; Van Veldhoven & Meijman, 1994). This questionnaire was chosen because it measures the four quadrants of the affective circumplex model (Russell, 1980). An example item of this questionnaire is: ‘Did you feel calm during your specific work task of this working day?’ The QEEW consists of six indicators for positive affect (optimistic, contented, calm, relaxed, cheerful, and enthusiastic), and six indicators for negative affect (tense, gloomy, depressed, worried, miserable, and uneasy). The positive and negative affect scales showed good reliability with a Cronbach's alpha of .89 and .86, respectively.

Intrinsic motivation was assessed with two items from the interest/enjoyment subscale of the intrinsic motivation inventory (McAuley, Duncan, & Tammen, 1989; see also Flora & Flora, 1999; Ilardi, Leone, Kasser, & Ryan, 1993; Koka & Hein, 2003). To lower the response burden, the two items with the highest factor loadings that were substantively different from each other were selected (‘You found your task very interesting’; ‘I would describe the task as very enjoyable’).

Results

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

Descriptive statistics

The means, standard deviations, and correlations of all variables are reported in Table 1. As can be seen, the three basic psychological need satisfactions were positively correlated to each other. Also, the three need satisfactions and intrinsic motivation were positively correlated with positive affect and negatively with negative affect. Finally, intrinsic motivation was positively related to satisfaction of the three needs.

Table 1. Means, standard deviations, and correlations (i.e., correlations between the person-centred variables) of satisfaction of the need for autonomy, competence, and relatedness, and positive affect, negative affect, and intrinsic motivation
 Mean (SD)12345
Note
  1. **< .01.

1. Need for autonomy4.55 (0.92)    
2. Need for competence3.87 (1.02).56**   
3. Need for relatedness5.46 (0.98).26**.14**  
4. Positive affect1.86 (0.49).41**.43**.24** 
5. Negative affect1.37 (0.57)−.36**−.32**−.22**−.55**
6. Intrinsic motivation3.94 (1.19).63**.58**.16**.54**−.45**

Multilevel path modelling

To test the mediating effect of positive and negative affect on the relationship between need satisfaction and intrinsic motivation, we estimated a two-level path model with measurements nested within participants (Preacher, Zhang, & Zyphur, 2011) using Mplus version 7 (Mplus, Los Angeles, CA, USA). Multilevel path modelling (and the more general method of multilevel structural equation modelling) is especially well suited to test a mediation analysis on multilevel data as it partitions the variance in a within and a between part, something that cannot be done in multilevel regression models (Preacher et al., 2011).

In particular, we modelled (1) satisfaction of the three basic psychological needs as predictors of positive and negative affect; (2) positive and negative affect as predictors of intrinsic motivation; and (3) satisfaction of the three needs as predictors of intrinsic motivation. Following the suggestions of Preacher et al. (2011), the mediators (i.e., positive and negative affect) were allowed to correlate. Moreover, the three need satisfactions were also allowed to correlate. Finally, to account for serial dependency in the data (due to the high-frequency repeated measurements), we included relationships from the lagged dependent variable (i.e., intrinsic motivation) and the mediators (i.e., positive and negative affect) to their current observations in the predictive model (see Beal & Weiss, 2003; West & Hepworth, 1991). Note that the inclusion of time-lagged variables in the predictive model implies that we evaluate whether changes in the mediators and the dependent variable can be predicted using the predictors in our model. As the impact of the time-lagged variables was of no substantive interest, we do not report them. Moreover, because we are interested in relationships at the within-person level, we only report the within-component of the multilevel path model.

The parameter estimates of the within-component of the model are shown in Table 2. First, regarding the relationships between the independent variables and the mediators, all need satisfactions related in a positive way to positive and in a negative way to negative affect. Second, positive affect was positively related to intrinsic motivation, while the relationship between negative affect and intrinsic motivation was negative. In addition, positive affect and negative affect were negatively correlated (= −.45; < .001). Third, after controlling for the mediation effect, the positive relationships between satisfaction of the need for autonomy and competence on the one hand and intrinsic motivation on the other hand remained, while the direct relationship between satisfaction of the relatedness need and intrinsic motivation was no longer significant.

Table 2. Unstandardized path estimates and their corresponding p-value
 Effectp-value
Predictor[RIGHTWARDS ARROW]Mediator
Autonomy[RIGHTWARDS ARROW]positive.389<.001
Autonomy[RIGHTWARDS ARROW]negative−.454<.001
Competence[RIGHTWARDS ARROW]positive.426<.001
Competence[RIGHTWARDS ARROW]negative−.266<.001
Relatedness[RIGHTWARDS ARROW]positive.222<.001
Relatedness[RIGHTWARDS ARROW]negative−.270<.001
Mediator[RIGHTWARDS ARROW]Outcome
Positive[RIGHTWARDS ARROW]intrinsic.197<.001
Negative[RIGHTWARDS ARROW]intrinsic−.084.002
Predictor[RIGHTWARDS ARROW]Outcome
Autonomy[RIGHTWARDS ARROW]intrinsic.512<.001
Competence[RIGHTWARDS ARROW]intrinsic.287<.001
Relatedness[RIGHTWARDS ARROW]intrinsic−.065.120
Predictor[RIGHTWARDS ARROW]Mediator[RIGHTWARDS ARROW]Outcome
Autonomy[RIGHTWARDS ARROW]positive[RIGHTWARDS ARROW]intrinsic.077.002
Competence[RIGHTWARDS ARROW]positive[RIGHTWARDS ARROW]intrinsic.084.001
Relatedness[RIGHTWARDS ARROW]positive[RIGHTWARDS ARROW]intrinsic.044.003
Autonomy[RIGHTWARDS ARROW]negative[RIGHTWARDS ARROW]intrinsic.038.008
Competence[RIGHTWARDS ARROW]negative[RIGHTWARDS ARROW]intrinsic.022.034
Relatedness[RIGHTWARDS ARROW]negative[RIGHTWARDS ARROW]intrinsic.023.057

Table 2 also reports the indirect relationships (i.e., the effects of the predictors on the outcomes via the mediators). In line with Hypothesis 1a, positive affect positively mediated the relationship between the satisfaction of the need for autonomy and intrinsic motivation. The relationship between the need for competence satisfaction and intrinsic motivation was also positively mediated by positive affect. Therefore, Hypothesis 2a was confirmed. Hypothesis 3a was also supported as positive affect positively mediated the relationship between the satisfaction of the need for relatedness and intrinsic motivation. Negative affect did also mediate the relationships between satisfaction of the need for autonomy, competence, and relatedness on the one hand and intrinsic motivation on the other hand (with the indirect effect being marginally significant for the need for relatedness). Together, these findings confirmed Hypotheses 1b, 2b, and 3b, respectively.

Apart from the hypothesized mediation model, we also tested two alternative models. Because, according to SDT, autonomous and intrinsic motivation follow from need satisfaction (e.g., Lam & Gurland, 2008; Ryan & Deci, 2009; Ryan et al., 2009), we limited the alternative models to those models where need satisfaction preceded motivation (i.e., affect–need satisfaction–motivation, and need satisfaction–motivation–affect). In the first alternative model, satisfaction of the three basic psychological needs was tested as a mediator between positive and negative affect on the one hand and intrinsic motivation on the other hand. Whereas statistically significant mediation effects were obtained for the need for autonomy and competence, the need for relatedness turned out to be no significant mediator. Moreover, the fit of this alternative model to the data was worse than that of our hypothesized mediation model (i.e., Bayesian Information Criterion [BIC] values of 19819.50 and 18924.81, respectively). In the second model, we tested whether intrinsic motivation mediated the relationship between satisfaction of the three basic psychological needs and positive and negative affect. The results of this model revealed that a mediation effect could be confirmed for the needs for autonomy and competence, but not for the need for relatedness. Moreover, the fit of this model to the data was worse than that of the hypothesized model (BIC values of 18926.17 and 18924.81, respectively).

Discussion

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

In general, the results of the first study supported our hypothesis that affect mediated the relationship between need satisfaction and intrinsic motivation. This is in line with SDT's basic assumption that the satisfaction of needs results in positive outcomes such as increased positive and decreased negative affect (Ryan & Deci, 2000). Further, our results also supported the claim of Isen and Reeve (2006) that affect is important for the understanding of the process of intrinsic motivation by demonstrating the mediating role of positive and negative affect in the need satisfaction–intrinsic motivation relationship.

While a diary study has some important strengths (Bolger, Davis, & Rafaeli, 2003), the downside is that the variables are not measured when they are experienced. This period between the experience and measurement may give rise to recall biases. A second limitation of our daily diary study was that all variables were collected at the same point in time (i.e., at the end of the day). Apart from the possible common method bias that may be caused by such a procedure (Podsakoff & Organ, 1986), it also prevents us from inferring directionality from the results. To (1) replicate the findings of our first study and (2) to resolve for the problems of recall bias and directionality, we tested the mediating effect of positive and negative affect in an experience sampling study as well.

STUDY 2

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

In study two, we tested the mediating effect of positive and negative affect on the need satisfaction–autonomous motivation relationship by means of an experience sampling design. Experience sampling studies are specifically designed to capture real-life experiences at random moments in time; therefore, the ratings are less susceptible to recall bias (Bolger et al., 2003). Moreover, to be able to hint towards directionality, we included a time lag between the measurements of need satisfaction and affect on the one hand and autonomous motivation on the other hand (Thrash, Elliot, Maruskin, & Cassidy, 2010).

Method

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

Participants

Thirty-seven Belgian employees, 17 men and 20 women, from the same company, took part in this study. Their mean age was 29.44 years (SD = 6.96). The majority of the participants worked full time (97.3%).

Procedure

The data were collected using HTC TOUCH2 Smartphones (Taoyuan City, Taiwan). Each participant received a Smartphone for 10 consecutive working days. The phone was programmed to beep four times a day at semi-random time intervals, that is, two times before lunch break and two times after lunch break. The first beep before and after lunch break appeared randomly, while the second beep followed 30 min after the first one. After the first beep, participants had to report on their level of need satisfaction and on their affective state, while at the second beep, their level of autonomous motivation was assessed. Using this time-lagged design, we made sure that the antecedents (i.e., need satisfaction and affect) preceded the effect (i.e., autonomous motivation), which allows for directional conclusions (Thrash et al., 2010). In total, 372 responses were collected, which implies that on average people answered 10.05 times (a response rate of 50.27%).

Measures

All items were selected from valid and reliable questionnaires. To cross-validate our results, we used different questionnaires than in Study 1. The phrasing of the original items was changed to measure experiences of participants at that particular moment. To lower participants' response burden, need satisfactions and motivation items were selected from existing questionnaires based on their factor loadings. Only the items with the highest factor loadings were retained. All items were answered on a 6-point rating scale going from totally disagree to totally agree.

The three basic need satisfactions were assessed with nine items from the Dutch version of the work-related basic need satisfaction scale (Van den Broeck et al., 2010) measuring satisfaction of the need for autonomy (e.g., ‘I feel like I have to follow other people's commands for this task’), competence (e.g., ‘I feel competent in this task’), and relatedness (e.g., ‘During this task, I don't really feel connected with other people at my job’).

The job emotion scale (Fisher, 2000), which measures work-related affect, was translated into Dutch. It consists of eight indicators of positive affect (liking someone and something, happiness, enthusiasm, pleasure, pride, enjoyment, contentment, and optimism) and eight indicators of negative affect (depression, frustration, anger, disgust, unhappiness, disappointment, embarrassment, and worry). A sample item of this questionnaire is: ‘During this task, I experience frustration’. The job emotion scale had good reliability for positive affect (Cronbach's α = .91) and negative affect (Cronbach's α = .92).

Autonomous motivation was assessed with two items from the situational motivation scale (Guay, Vallerand, & Blanchard, 2000), which measures state motivation. In particular, the validated Dutch translation of Bos-Nehles (2010) was used in this study. The selected items were ‘I do this task because I think that performing this task is interesting’ and ‘I do this task for my own good’.

Results

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

Descriptive statistics

Table 3 reports the means, standard deviation, and correlations. As can be seen from this table, satisfaction of the need for autonomy and competence were positively related to each other. Autonomous motivation was positively related to positive affect and negatively to negative affect. Positive affect was positively related to the three types of need satisfaction, and negative affect was negatively related to satisfaction of the need for autonomy and competence.

Table 3. Means, standard deviations, and correlations (i.e., correlations between the person-centred variables) of the satisfaction of the need for autonomy, competence, and relatedness, and positive affect, negative affect, and autonomous motivation
 Mean (SD)12345
Note
  1. *< .05; **< .01.

1. Need for autonomy3.26 (0.92)    
2. Need for competence5.01 (0.66).12*   
3. Need for relatedness3.16 (0.80).13*.01  
4. Positive affect3.69 (0.88).17**.19**.16** 
5. Negative affect1.94 (0.82)−.14**−.15**.04−.50**
6. Autonomous motivation4.06 (1.04).06.10.07.27**−.13*

Multilevel path modelling

Multilevel path modelling was used to test whether positive and negative affect mediated the relationship between satisfaction of the need for autonomy, relatedness, and competence on the one hand and autonomous motivation on the other hand. As our data have a nested data structure with measurements nested within days, and days nested within participants, we modelled our data using a three-level path model. In particular, we predicted (1) positive and negative affect by the three basic psychological need satisfactions; (2) intrinsic motivation by positive and negative affect; and (3) intrinsic motivation by the three need satisfactions. Moreover, positive and negative affect and the three need satisfaction variables were allowed to correlate (Preacher et al., 2011). Finally, we controlled for serial dependency in the data by including relationships from the lagged dependent variable and mediators to their current observations (see Beal & Weiss, 2003; West & Hepworth, 1991). Again, this implies that we are predicting changes in the mediators and the dependent variable. As these lagged variables were of no substantive interest, we did not include their results in this study. Because we were interested in relationships at the within-person level, we only report the within-person effects.

The parameter estimates of the model are shown in Table 4. Regarding the links between basic need satisfaction and affect, our results showed that the satisfaction of the need for autonomy related in a (marginally) negative way to negative affect, whereas it was unrelated to positive affect. The satisfaction of the needs for competence and relatedness related positively to positive affect, whereas negative affect was marginally related to satisfaction of the need for competence only. Regarding the link between affect and motivation, the results showed that positive affect related positively to autonomous motivation, whereas no relationship was found between negative affect and autonomous motivation. Positive affect and negative affect were correlated negatively (= −.51; < .001).

Table 4. Unstandardized path estimates and their corresponding p-value
 Effectp-value
Predictor[RIGHTWARDS ARROW]Mediator
Autonomy[RIGHTWARDS ARROW]positive.441.178
Autonomy[RIGHTWARDS ARROW]negative−.437.076
Competence[RIGHTWARDS ARROW]positive1.002<.001
Competence[RIGHTWARDS ARROW]negative−.803.064
Relatedness[RIGHTWARDS ARROW]positive.533.025
Relatedness[RIGHTWARDS ARROW]negative.237.212
Mediator[RIGHTWARDS ARROW]Outcome
Positive[RIGHTWARDS ARROW]autonomous motivation.087.005
Negative[RIGHTWARDS ARROW]autonomous motivation.020.551
Predictor[RIGHTWARDS ARROW]Outcome
Autonomy[RIGHTWARDS ARROW]autonomous motivation.075.437
Competence[RIGHTWARDS ARROW]autonomous motivation.124.054
Relatedness[RIGHTWARDS ARROW]autonomous motivation−.026.830
Predictor[RIGHTWARDS ARROW]Mediator [RIGHTWARDS ARROW]Outcome
Autonomy[RIGHTWARDS ARROW]positive[RIGHTWARDS ARROW]autonomous motivation.038.264
Competence[RIGHTWARDS ARROW]positive [RIGHTWARDS ARROW]autonomous motivation.087.022
Relatedness[RIGHTWARDS ARROW]positive [RIGHTWARDS ARROW]autonomous motivation.046.087
Autonomy[RIGHTWARDS ARROW]negative [RIGHTWARDS ARROW]autonomous motivation−.009.603
Competence[RIGHTWARDS ARROW]negative [RIGHTWARDS ARROW]autonomous motivation−.016.554
Relatedness[RIGHTWARDS ARROW]negative [RIGHTWARDS ARROW]autonomous motivation.005.524

The indirect relationships in Table 4 revealed that positive affect mediated the relationship between satisfaction of the need for competence on the one hand and autonomous motivation on the other hand, thereby confirming Hypothesis 2a. Also for satisfaction of the need for relatedness, a marginally significant mediation effect through positive affect was found. As negative affect was not predictive of autonomous motivation, it did also not mediate the relationship between satisfaction of the three basic psychological needs and autonomous motivation, thereby rejecting Hypotheses H1b, H2b, and H3b.

As satisfaction of the three basic psychological needs and positive and negative affect were measured at the same time, we also tested an alternative mediation model in which need satisfaction mediated between positive and negative affect on the one hand and autonomous motivation on the other hand. In this model, there were no significant mediation (or indirect) effects. Moreover, our hypothesized model with affect as a mediator of the need satisfaction–autonomous motivation relationship had a better fit to the data (BIC values of 12345.20 and 12330.78 for the alternative and hypothesized model, respectively).

Discussion

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

The results of the experience sampling study partially confirmed the results that were found in the daily diary study. That is, positive affect appeared to play a key role in the need satisfaction–autonomous motivation relationship, and the proposed mediation chain provided a better fit to the data than a mediation model in which satisfaction of the three needs mediated the relationship between affect and autonomous motivation. Moreover, by including a time lag between the measurements of need satisfaction and affect on the one hand and autonomous motivation on the other hand, we established temporal precedence between our measurements of affect and autonomous motivation. Therefore, stronger claims about the directionality of the results can be made.

At the same time, some of the results of the experience sampling study diverged from those of the daily diary study. First, in the experience sampling study, positive affect did not follow from satisfaction of the need for autonomy. Because this finding was counter to our expectations, we ran a three-level model to test the relationship between satisfaction of the need for autonomy and positive affect without accounting for the effects of competence and relatedness need satisfaction. This analysis revealed that there was a positive relationship between satisfaction of the need for autonomy and positive affect (β = .63; = .086). Hence, it seems that in our experience sampling study, the effect of autonomy need satisfaction on positive affect was already accounted for by the effects of competence and relatedness need, a finding that was also found in the meta-analysis of Patrick, Knee, Canevello, and Lonsbary (2007). Second, satisfaction of the need for relatedness was not predictive of changes in negative affect. Because of this reason, we also tested the effect of relatedness satisfaction without controlling for the effects of the other two needs. This analysis showed that satisfaction of the relatedness need was not linked to negative affect (β = .18; = .393; see also the non-significant correlation in Table 3). Whereas this result was not anticipated, other studies have also found that relatedness is primarily linked to variation in positive affect (e.g., Quested & Duda, 2009; Reis et al., 2000; Sheldon & Bettencourt, 2002). Indeed, it has been shown that positive affect increases when people socialize, whereas negative affect results from stressful or aversive events (Watson & Clarck, 1994). In situations in which one's need for relatedness is satisfied (i.e., a situation with people one feels connected with), the likelihood of both positive and negative (e.g., arguments) interactions increases (Reis et al., 2000). As a result, negative, stressful events such as interpersonal conflicts do not readily imply a lowered satisfaction of the need for relatedness. In other words, satisfaction of the need for relatedness is not straightforwardly linked to negative affect. Third, we failed to find a relationship between negative affect and autonomous motivation. One explanation for this difference might be that we measured intrinsic motivation in the first study, while the broader category of autonomous motivation was measured in the second one. However, when identified and intrinsic motivation were tested in two separate models, we only found a significant effect of positive affect on intrinsic motivation, while identified motivation was unrelated to both positive and negative affect. This finding has two important implications. First, it implies that (positive) affect is particularly important for intrinsic motivation, while it is less important for other types of motivation. This makes sense as the definition of intrinsic motivation, that is, ‘the engagement in an activity for its own sake, that is, for the satisfaction and enjoyment experienced during the course of the activity itself’ (Van den Broeck et al., 2008, p. 4) directly refers to positive emotions such as satisfaction and enjoyment, while this is not the case for the other types of motivation. Second, the results of the separate models also imply that the failure to replicate the relationship between negative affect and motivation cannot be explained by the fact that we measured autonomous instead of intrinsic motivation in the experience sampling study. Instead, we believe that this finding should be interpreted as an indication for the secondary role of negative affect in the coming about of intrinsic motivation. Whereas positive affect (and in particular enjoyment and interest) is by definition required for intrinsic motivation to emerge, negative affect is primarily relevant because it tends to be incompatible with enjoyment and interest (Pekrun et al., 2002). In other words, negative affect is mainly important because it has the possibility to block the (positive affective) feelings that are crucial in the coming about of intrinsic motivation. In line with this interpretation, the results of the diary study showed that the effect of negative affect on intrinsic motivation was substantially smaller than the effect of positive affect, and because of the introduction of the time lag in the experience sampling study, this small effect may have disappeared.

GENERAL DISCUSSION

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References

Our results confirmed the central role of (positive) affect in SDT by showing that affect mediates the relationship between satisfaction of the three basic psychological needs and intrinsic motivation. By means of a daily diary and an experience sampling study, we demonstrated that satisfaction of the three basic psychological needs is related to variations in (positive) affect and that these variations are related to variations in intrinsic motivation. Moreover, by testing alternative models with different causal effects, we were able to exclude viable alternative explanations. In general, our findings are in line with SDT's assumption that need satisfaction leads to positive outcomes (Gillet et al., 2012; Quested & Duda, 2009; Reis et al., 2000; Ryan & Deci, 2000; Sheldon & Bettencourt, 2002) and that positive affect fosters intrinsic motivation (Custers & Aarts, 2005; Frijda, 2008; Frijda et al., 1989; Isen & Reeve, 2006; Martin et al., 1993; Meyer & Turner, 2006; Raghunathan & Tuan Pham, 1999; Scherer, 2004).

Theoretical and practical implications

A first important theoretical implication of our results is that they call for an extension of SDT. Whereas SDT hypothesizes that affect plays a distal role in the coming about of motivation (Deci & Ryan, 1987; Isen & Reeve, 2006), we have shown that this is not the case. Instead, our results supported the claim of Isen and Reeve (2006, p. 321) that ‘affect may play a more central role in understanding intrinsic motivational process than is currently recognized by self-determination theory’. Moreover, we went beyond Isen and Reeve (2006) by demonstrating that a better understanding of the (positive) affective mechanisms that underlie the need satisfaction–motivation link is necessary if we want to get a grip on the mechanisms through which need satisfaction impacts on intrinsic motivation.

On a general level, our findings stressed the key role of (positive) affect in the coming about of intrinsic motivation. Whereas this has also been recognized by others (e.g., Dowson & McInerney, 2003; Isen & Reeve, 2006; Seo, Barrett, & Bartunek, 2004), motivation researchers have traditionally focused on cognitive rather than on affective antecedents of motivation. Our results dispute this choice and pave the way for studies that focus on both cognitive and affective determinants of motivation. At this point, we believe that affect and cognition should not be conceived as opposing forces, but rather as inherently integrated and interacting (see also Hofmans, Gelens, & Theuns, 2013; Storbeck & Clore, 2007).

Our results have practical implications as well. The central role of (positive) affect in the need satisfaction–intrinsic motivation relationship stresses the highly subjective nature of the motivation process. Because affect results from an appraisal of the situation or environment (Scherer, 2004), not the objective situation but rather the situation as it is appraised determines the level of autonomous motivation. Our results thus imply that organizations can influence the level of intrinsic motivation of their employees by (1) changing the working conditions in a way that they fulfil their needs; and (2) changing the appraisals of these working conditions by, for example, communicating about them.

Limitations and future research

When interpreting the results of this study, some limitations should be taken into account. First, important events may have happened between the consecutive measurement moments, and these events may have affected the relationships between the variables in our mediation models. Second, whereas we used time lags to establish temporal precedence while controlling for previous levels of affect and motivation, our results can only hint towards causal relationships. To test for true causality, three conditions have to be fulfilled: (1) causes and effects should be correlated; (2) causes must precede effects; and (3) possible spurious causes should be controlled for (Thrash et al., 2010). While the first condition was fulfilled in Study 1, the first and second conditions were fulfilled in Study 2, and the third one can only be addressed through randomization. As such, experimental research on the mediating role of affect is needed in order to make strong causal inferences. Third, positive and negative affect, need satisfaction, and motivation were assessed with different measurement instruments in the daily diary and the experience sampling study. These methodological differences may be partly responsible for the differences that were found in the two studies. However, an unmistakable advantage of this choice is that we were able to cross-validate our research with different instruments, which makes our research less likely to result from method bias. Finally, there are several ways in which this line of research can be expanded. First, future research might focus on the mediating role of discrete affect, as different discrete emotions are known to have different action tendencies (Frijda et al., 1989; Scherer, 2009), which suggests that they might have different effects on intrinsic and autonomous motivation. Second, in this study, we only focused on intrinsic and autonomous motivation. Other (sub)types of motivation are also important in a working context and should thus be included in future research. Third, few studies have focused on need thwarting. Therefore, an interesting avenue for future research is to study the mediating role of affect in the need thwarting–motivation relationship. Finally, as people differ in the degree to which they vary in their positive and negative affect (Beal & Ghandour, 2011), it would be interesting to study the impact of interindividual differences in affect variability on the affect–motivation relationship.

References

  1. Top of page
  2. Abstract
  3. Background
  4. STUDY 1
  5. Method
  6. Results
  7. Discussion
  8. STUDY 2
  9. Method
  10. Results
  11. Discussion
  12. GENERAL DISCUSSION
  13. References