In the past decade, there has been an affective revolution in the study of organizational behavior (Barsade, Brief, & Spataro, 2003). The affective reactions of employees in response to workplace events is a topic of substantial importance to managers and researchers alike. Positive emotional reactions have been linked to numerous desirable outcomes such as increased productivity, job satisfaction and empowerment, and decreased stress and turnover. Conversely, negative emotional reactions have been shown to predict a wide array of undesirable outcomes such as tension, turnover, decreased productivity and even workplace violence (Bagozzi, 2003). However, we argue that our current understanding of the relationships between affect and behavioral outcomes is incomplete. Specifically, we posit that the influence of emotional intelligence on the attributional component of the stimulus-emotion-behavior process (see Weiner, 1985) has not been adequately investigated.
The purpose of this paper is to present a theoretical framework intended to better understand the cognitive mechanisms underlying the relationship between workplace outcomes, attributions, emotional reactions, and subsequent behaviors. Our theoretical model is based on past research on attribution theory and recent work on emotional intelligence. Specifically, we suggest that the relationship between attributions, affect, and behavioral outcomes may be moderated by an individual’s level of emotional intelligence.
We begin by providing an overview of research on attribution theory, the relationship between attributions and emotions, and emotional intelligence. We then present our theoretical model of affective reactions to workplace events. We conclude with a discussion of the implications of this process for several facets of organizational research, including motivation (empowerment and learned helplessness), stress, deviance, and aggression.
In order to manage the scope of the paper, our discussion will focus on employees’ perceptions of negative outcomes. We reason that these outcomes are of heightened interest for two reasons. First, as noted by Adams (1965) and Festinger (1957), negative outcomes are more likely than positive ones to trigger an in-depth cognitive evaluation process because they are less likely to be anticipated. For this reason, the attributional framework developed in this paper may be less relevant when positive outcomes occur. Second, the psychological and behavioral ramifications, including stress, workplace deviance, and aggression, resulting from negative outcomes can be far more severe than those cued by positive outcomes (Martinko & Gardner, 1987).
Attribution theory and emotions
Attribution theory is the study of the process by which individuals make assessments of causality in response to the outcomes they observe. The theory is based on Heider’s (1958) notion that people have an inherent tendency to be «naïve psychologists» that attempt to determine the causes of events that are important to them. A wide array of causal attributions can be made, but attributions of ability, effort, luck and task difficulty are among the most common (Weiner, Frieze, Kukla, Reed, Rest, & Rosenbaum, 1971). As with all perceptions, however, attributions do not necessarily reflect one’s objective reality. Different people have shown systematic tendencies, known as attribution styles, toward making certain attributions for outcomes across both time and situation (Abramson, Seligman, & Teasdale, 1978). These tendencies are most likely to influence attributions in situations characterized by high levels of ambiguity, where the causes of outcomes are unclear. Thus, our discussion of the attribution process will focus on relatively ambiguous situations.
Research on attribution theory has identified several key attributional dimensions (see Kelly, 1967; Abramson et al., 1978; Weiner et al., 1971), four of which appear to have implications for emotional responses to negative outcomes. The first, locus of causality, refers to the perceived internality or externality of the cause of an outcome (Weiner et al., 1971). Internal causes are those perceived to be a result of internal characteristics of a person (e.g., ability or intelligence), whereas external causes are those outside the control of individuals, such as luck or chance. The second dimension, stability, pertains to the variability of a perceived cause across time and situation (Kent & Martinko, 1995). Causes such as intelligence are assumed to be relatively stable while causes such as effort are more variable. The third dimension, controllability, describes the extent to which individuals believe they, or someone else, are able to control the cause of an outcome (Weiner, 1979). The fourth dimension, intentionality, refers to the degree to which individuals perceive that the cause of an outcome was intentional.
Attributions and emotions
Weiner (1985) noted that attributions are related to affective outcomes. More specifically, emotions are considered reactions to the causal evaluations individuals make in response to outcomes (Weiner, Graham, & Chandler, 1982). Regarding the locus of causality dimension, individuals that make an internal attribution for a negative outcome, such as "I failed because I am unintelligent", are likely to experience shame. Alternatively, if a negative outcome is attributed externally to another person, the resulting emotion is most likely to be anger or frustration toward that person (Weiner, 1985).
Research has also indicated that the perceived stability of causes can affect emotional responses, given their impact on individuals’ expectations for future outcomes (Weiner, 1985). In case of negative outcomes, stable attributions can exacerbate the negative effect of internal attributions on emotions. This is because the cause of the negative outcome is not only thought to be the self, but also unchangeable (or at least very difficult to change), meaning that the same negative outcome is likely to occur again in the future. Conversely, unstable attributions for negative outcomes are likely to attenuate the negative effect of internal or external attributions on emotions because the perceived cause (e.g., level of effort) may not be present in future attempts at completing the task (Martinko, 2002).
Gundlach, Douglas and Martinko (2002) argued that attributions of control in response to negative outcomes are likely to trigger emotions such as anger and resentment when the cause is also attributed to internal and stable characteristics of another person. The logic behind this argument is that attributions of stability and control are related to assignments of responsibility. If a failure is attributed to stable factors that reflect internal, controllable traits of another person (e.g., dishonesty), that person is likely to be held responsible for the outcome. As a result, frustration and anger toward that person is likely. Conversely, failures attributed to internal and controllable factors of the perceiver such as inadequate effort should not have as strong a negative affect on emotions because action can be taken to remove the cause (e.g. work harder).
Finally, the influence of internal, stable and controllable attributions on emotions in response to negative outcomes is argued to be exacerbated by attributions of intentionality (Gundlach et al., 2002). The authors cite evidence (i.e., Berkowitz, 1993; Geen, 1990; Kidd & Utne, 1978; Utne & Kidd, 1980) showing that attributions of intent increase one’s emotional sensitivity. This increased affective responsiveness is predicted to raise the likelihood and intensity of negative emotional reactions to stable and controllable attributions resulting from negative outcomes. This argument is consistent with Dasborough and Ashkanasy’s (2002) model of emotions and attributions of intentionality. Their model predicts that subordinates that make negative attributions of intentionality (i.e., self-serving, manipulative intentions) for their leader’s behaviors are likely to experience negative affect toward the leader, decreasing leader-member relationship quality.
Given the apparent relationship between emotions and attributions, we suggest that emotional intelligence should be considered in the study of attributions also. Research has indicated that emotional intelligence contributes to work performance (i.e., Brotheridge, 2006; Cherniss & Adler, 2001; Goleman, 1998; Lopes, Grewal, Kadis, Gall, & Salovey, 2006; Mayer, Caruso, & Salovey, 2000). As such, emotional intelligence has become a popular topic among organization scientists and managers, as evidenced by the many book sales (e.g., Goleman, 1998) and article publications (e.g., Druskat & Wolff, 2001) in recent years.
Emotional intelligence is defined as «the ability to monitor one’s own and others’ feelings and emotions, to regulate them, and to use emotion-based information to guide thinking and action» (Salovey, Bedell, Detweiler, & Mayer, 1999, p. 141). The concept of emotional intelligence is rooted in the idea that there is a link between reason and emotion (Ciarrochi, Forgas, & Mayer, 2001), and that both emotional and intellectual reasoning are integral to human success and survival (Brackett & Salovey, 2006).
For the purposes of this paper, we adopt the four-dimensional model of emotional intelligence put forth by Mayer and Salovey (1997) that encompasses perception, facilitation, understanding, and management of emotions (referred to as dimensions 1-4, respectively). Perception concerns the identification of emotions in others and in oneself; that is, whether one can read emotions accurately. Facilitation concerns the use of emotions to help achieve outcomes, solve problems, capitalize on opportunities, and motivate. Understanding emotions is the ability to understand complex emotions, and how they change over time. The final dimension is the ability to manage emotions in oneself and in others. These four dimensions represent iterative steps, in that each one contributes to the development of others (for example, one needs to perceive emotions prior to being able to assimilate, understand, and then manage them).
Salovey et al. (1999) and Jordan, Ashkanasy, Härtel, and Hooper (2002) have argued that, taken together, the four dimensions of emotional intelligence can moderate employees’ emotional reactions to negative outcomes and their ability to cope with the associated stress. Salovey et al. (1999) argued that high levels of emotional intelligence can promote effective coping by decreasing the extent to which individuals ruminate on negative events, by promoting emotional disclosure, and by increasing individuals’ proclivity to seek social support after negative events occur. Similarly, Jordan et al. (2002) suggested that people high on emotional intelligence tend to deal with their negative emotional reactions in ways that promote a productive result. The essence of their argument is that people high on emotional intelligence are able to cope with stressful events and therefore have less extreme emotional reactions to such events.
Although we are also interested in the moderating role of emotional intelligence, our approach is different in that we argue that by considering each of these dimensions or branches separately, we can learn more about emotional intelligence and the role it plays in the attribution-emotion-reaction process.
Theoretical model and propositions
Our conceptual model of the cognitive process that links workplace events with behavioral outcomes via attributions and emotional intelligence is shown in Figure 1. Given that Weiner (1985) has already conceptualized the outcome - affect - attribution - emotion portion of this model, our propositions focus on the role of emotional intelligence and the specific behavioral, motivational, and psychological outcomes that we have added to the initial attributional model.
As suggested by Weiner’s (1985) attributional theory of motivation and emotion, our model shows that outcomes are expected to create an initial affective reaction which triggers the attribution process. We argue that the emotional intelligence dimensions of emotion perception, facilitation, and understanding moderate this affect-attribution relationship. Consistent with the past research cited above, the resulting attributions are predicted to influence emotional outcomes. We argue, however, that this relationship between attributions and emotions is also moderated by emotional intelligence. Specifically, we predict that the emotion management dimension of emotional intelligence will influence the emotions felt by individuals. Finally, these emotions are predicted to cue certain motivational, behavioral, and psychological responses. We will now discuss each of these relationships in greater detail.
Outcome and initial affective response
There has been some debate concerning the primacy of emotion over cognition (see Lazurus, 1984; Zajonc, 1984). For the purpose of the present discussion we adopt the view put forth by Weiner (1985) which argues that when outcomes relevant to an individual occur, that individual experiences an immediate emotional response which is subject to subsequent cognitive adjustment. To illustrate, if an employee experiences a negative outcome such as being denied a promotion, it is predicted that an immediate negative emotion such as anger, sadness, or shock will be felt. This response then triggers the rest of the cognitive process outlined in the following sections, possibly resulting in an outcome based on an emotion that is different from the initial affective response.
Initial affective response and attributions
Research has shown that the attribution process is triggered when outcomes are considered important and unexpected by perceivers (Weiner, 1977). As noted above, negative outcomes commonly meet both criteria. Weiner (1985) indicated that the attribution process serves as a mechanism to try to explain the cause of these unexpected and important outcomes and the initial affective reactions they produce. Thus, we do not conceptualize attributions as being a direct result of outcomes but as a result of the initial affective responses to those outcomes. Specifically, initial outcome-dependant affect is thought to mediate the relationship between workplace outcomes and the resulting attributions.
The moderating role of emotional intelligence (Dimensions 1-3)
We argue that emotional intelligence may be an important component of the attribution process. As such, our model suggests that the degree to which individuals’ attributions are constructive, given their current situation, may be a product of their level of emotional intelligence. We suggest that high levels of emotional intelligence are associated with more constructive attribution styles (i.e., attribution styles that promote desirable responses to outcomes). The basis for this assertion is the idea that emotional intelligence allows individuals to effectively deal with negative outcomes. We argue that emotional intelligence does not promote coping by directly influencing their ultimate emotional reactions, but rather by moderating the attributions individuals make based on their initial affective responses, as shown in Figure 1.
To understand this moderating relationship, it is necessary to consider the impact of the emotional intelligence dimensions of perception, facilitation and understanding. The perception dimension of emotional intelligence is associated with the ability to perceive emotions in oneself (Mayer & Salovey, 1997). Those with high levels of emotional intelligence should be able to use this ability to become aware of the emotions they feel and to familiarize themselves with the behavioral implications of those emotions. The second dimension, according to Mayer and Salovey, is facilitation, or the ability to use emotions in order to achieve desirable outcomes. With respect to the third dimension, people with the ability to understand emotions are likely to understand the source of their initial emotional response to a negative outcome, and this heightened understanding should impact their attributions.
Implicit in the preceding discussion is the notion that emotional intelligence enables individuals to comprehend the cause of their initial affective reaction and the implications of their emotional responses. This implies an understanding of the causes of outcomes that produce initial affect (per the understanding dimension) which may promote accurate attributions. In cases where the exact cause of an outcome is ambiguous and open to debate, we argue that the facilitation and perception dimensions might promote a tendency toward attributions linked to the most constructive emotional responses.
Based on this argument, Table 1 indicates that high levels of emotional intelligence are predicted to cue unstable, unintentional attributions that are either internal and controllable or external and uncontrollable when negative events occur. These attributions are believed to produce the most constructive outcomes. For example, those with high levels of emotional intelligence should recognize that attributing negative outcomes to a leader is likely to cause tension and conflict with the leader. Such an attribution is therefore unproductive (unless the leader clearly is the cause of the outcome). Conversely, those with low levels of emotional intelligence are predicted to be less cognizant of the implications of their attributions and may exhibit a bias toward less constructive attributions. We argue that when the cause of an outcome is ambiguous, these individuals will have a tendency toward stable, intentional, and internal and uncontrollable or external and controllable attributions for negative outcomes.
Proposition 1: Emotional intelligence (EI dimensions 1-3)
moderates the relationship between individuals’ initial affective responses
and their attributions
The moderating role of emotional intelligence (Dimension
As noted earlier, Jordan et al. (2002) argued that emotional intelligence allows people to cope with stress by moderating emotional reactions to negative job outcomes. We suggest that this moderation influences the relationship between outcome-based attributions and emotional responses described above. This argument is partially based on an early finding by Dasborough (2004), who examined how emotional intelligence moderates individual followers’ emotional reactions to attributions of leader intentions. Dasborough found that different levels of emotional intelligence were associated with different emotional responses to attributions of leader intentionality. Specifically, participants lower on emotional intelligence had more extreme emotional responses than their more highly emotionally intelligent counterparts.
To better understand this relationship, recall that the fourth dimension of emotional intelligence involves the ability to manage one’s emotions. As suggested by Mayer and Salovey (1997), emotional responses and their effect on behavior are influenced by individuals’ ability to manage their emotional response. People high on this dimension of emotional intelligence will manage the affect of their emotions on their behavior to promote the most effective outcomes (Salovey et al., 1999). For example, in the case of negative emotional responses to attributions, we argue that people high on emotional intelligence will manage this negative emotion to ensure it does not negatively impact work performance.
Proposition 2: Emotional intelligence (EI dimension 4) moderates
the relationship between member attributions and emotional responses
Psychological, motivational and behavioral consequences
As indicated in the preceding sections, there is theoretical and empirical evidence indicating that emotional intelligence and the attribution process can influence the emotional responses of individuals. In this section we will briefly review prior research on the relationship between emotions and four emotion-driven reactions that may result following a negative outcome, depending on the attributions individuals make: empowerment, learned helplessness, stress, and aggression.
Empowerment and learned helplessness
Empowerment and learned helplessness represent high and low levels of motivation, respectively (Campbell & Martinko, 1998). Empowerment is defined as an increase in an individual’s effort-performance expectancies (Conger & Kanungo, 1994), and is associated with high levels of effort, innovation, and managerial effectiveness (Spreitzer, 1995). Conversely, learned helplessness is a state of very low motivation in which minimal effort is exerted. Learned helplessness is typically a result of repeated failures at a task, and often persists even after the factors leading to the failure have been removed (Martinko & Gardner, 1982; Overmier & Seligman, 1967).
Campbell and Martinko (1998) found that unstable, externally uncontrollable, or internally controllable attributions for failures are associated with the positive emotions that cause empowerment. As discussed above, these attributions are similar to those we expect to be made by individuals with high levels of emotional intelligence. Campbell and Martinko also found that internal, stable, uncontrollable attributions for failures are associated with the negative emotions that promote learned helplessness. Again, we argue that individuals with high levels of emotional intelligence will make similar attributions, suggesting that low levels of emotional intelligence increase one’s propensity to experience learned helplessness. Some indirect empirical support for this argument was provided by Ciarrochi, Deane, and Anderson (2002), who concluded that emotional intelligence moderates the relationship between stressful outcomes and hopelessness – a construct similar to learned helplessness – such that high levels of emotional intelligence reduced the likelihood of hopelessness in response to negative outcomes.
It should be noted that our framework includes the attributional dimension of intentionality, whereas Campbell and Martinko did not include this in their study. We argue, however, that the attributions to unintentional causes that we expect emotionally intelligent people to favor in ambiguous situations will further facilitate empowerment because the perceiver will not feel that future attempts at a task will be purposefully thwarted by an external source.
Thus, it is our contention that high levels of emotional intelligence produce attributions that are amenable to empowering emotions while low levels produce attributions that promote negative emotions and learned helplessness.
Proposition 3: Due to the attributions they make and their
emotional responses, individuals with high levels of emotional intelligence
will display empowerment more frequently than those with low levels of emotional
Proposition 4: Due to the attributions they make and their
emotional responses, individuals with low levels of emotional intelligence will
display learned helplessness more frequently than those with high levels of
Given that relationships between attributions, emotions and stress have been established in past research (e.g., Zellars & Perrewe, 2001), it follows that the same attributions that are linked to emotions should also be indirectly related to stress. Perrewe and Zellars (1999) developed a conceptual model of this process indicating that internal, unstable and controllable attributions for negative outcomes will cue emotions such as guilt that are not pleasant but that promote constructive coping responses to the outcome (e.g. working harder in the future). Conversely, failures that are attributed to stable, external and uncontrollable causes are predicted to promote emotions such as anger that may exacerbate stress. Based on our discussion of the moderating role of emotional intelligence above, we argue that high levels of emotional intelligence might promote the internal, but unstable and controllable, attributions for negative outcomes that Perrewe and Zellars associated with lower levels of stress.
Proposition 5: Due to the attributions they make and their
emotional responses, individuals with high levels of emotional intelligence
will experience less stress than those with low levels of emotional intelligence.
Deviance and aggression
The most potentially troubling outcomes that may result from attributional processes and emotional reactions can be grouped under the umbrella of organizational deviance. Douglas and Martinko (2001) explained that two forms of organizational deviance exist – passive and aggressive. Passive deviance is said to result from internal and stable attributions for negative outcomes and refers to actions such as petty theft and tardiness that are detrimental to an organization but are not of an overly serious nature. This form of deviance can also take self-targeted forms such as alcohol and drug abuse that affect organizations indirectly through diminished performance.
Aggressive forms of organizational deviance can be far more severe in their consequences for organizations and their employees. Aggression can take different forms of varying severity, ranging from vocal outbursts to homicide. Martinko and Zellars (1998) argued that stable, externally controllable and intentional attributions for failures are likely to trigger such reactions in certain people. They explain that the anger and frustration that result from such attributions can push some individuals toward aggressive retaliation.
Because we argue that low levels of emotional intelligence will promote attributions similar to those associated with both forms of deviance, it appears that individuals low on emotional intelligence will be more be more likely to commit acts of deviance than those with high levels of emotional intelligence. Conversely, those high on emotional intelligence might not make the attributions associated with deviance, and should therefore be able to more constructively manage negative emotions. This argument is in line with Quebbeman and Rozell’s (2002) prediction that high levels of emotional intelligence, along with constructive attribution styles, will decrease the perceptions of injustice often associated with acts of workplace aggression.
Proposition 6: Due to the attributions they make and their
emotional responses, individuals with high levels of emotional intelligence
will exhibit fewer deviant/aggressive behaviors than those with low levels of
Limitations and future research
The model we have presented in this paper should be viewed as a tool to explicate the nature of the attribution-emotion relationship and, in particular, the role that emotional intelligence plays in this process. The model that we have presented is not all-encompassing and we acknowledge that other variables may influence the relationships we have discussed. While these other variables may be important, they are outside the scope of our proposed model, which focuses specifically on the attribution-emotion-reaction process for parsimony.
Another limitation concerns the attributional dimension of intentionality. This dimension has not received the same amount of attention in attributional research as the other dimensions. As a result, its proposed role in our model is more speculative than the roles of stability, controllability and locus of causality.
Additionally, it must be reiterated that the causal relationships proposed in this paper are predicted to be relevant only when the cause of an event is ambiguous and open to interpretation. In many instances of failure, perhaps the majority, cause-and-effect relationships are clear and factual attributions can be made. In these cases our cognitive model is less likely to be relevant to either emotional or behavioral outcomes.
It could be suggested that the greatest limitation of our theoretical model is the central role of emotional intelligence. This construct is still being developed, and to date has been widely criticized in the academic arena (e.g. Barrett, 2000). Despite these criticisms, Matthews, Zeidner, and Roberts (2003) conclude that emerging constructs such as emotional intelligence can play a positive role in science by stimulating research. It is our hope that by incorporating emotional intelligence into a nomological network, future research will be stimulated in the field.
To date, our specific model remains to be tested. In this respect, opportunities exist for laboratory and field research to test the model we have proposed as a means to understand in more detail the role of emotional intelligence in the attribution-emotion-reaction process. Future research might also investigate the possibility that emotional intelligence might, over time, reduce the frequency with which individuals experience the negative outcomes thought to initiate the cognitive process described in our model. Salovey et al. (1999) argued that emotional intelligence helps individuals learn from negative events and use this wisdom to shape future behaviors so that such events become less likely. Based on the arguments presented in this article, it seems possible that the attribution process might also influence this aspect of emotion-based coping.
In conclusion, we concur with the growing number of scholars who argue that emotional intelligence is an important predictor of coping ability and workplace behavior. We argue, however, that future research on this construct should consider the possible interaction of causal attributions and emotional intelligence. Similarly, we feel that future attributional research based on Weiner’s (1985) framework should consider the influence of emotional intelligence on the attribution-emotion-behavior process.
This article was written while Marie Dasborough was a doctoral student at the University of Queensland (Australia).