Abstract

A model explaining the relationship between task autonomy and performance is proposed that incorporates 3 different causal mechanisms. The performance benefits of task autonomy may be realized by increased motivation (motivational mechanisms), by capitalization of information asymmetries (informational mechanisms), or by better alignment with task and organizational structures (structural mechanisms). Further, it is proposed that these performance benefits are moderated by a variety of variables ranging from individual traits to organizational design. This model may provide a means for accounting for the sometimes inconsistent findings in the empirical literature exploring the relationship between autonomy and performance. The model also offers guidance in the search for additional boundary conditions as well as prescriptive guidelines for the allocation of autonomy in practice. Never tell people how to do things. Tell them what to do and they will surprise you with their ingenuity. -General George S. Patton The notion of task autonomy-giving the individual who performs a task considerable discretion and control in deciding how to carry it out-is as old as organizations themselves. The widespread and organized use of task autonomy in organizations, however, is a relatively modern phenomenon. Rarely used systematically before 1980, by the mid 1990s, task autonomy and related forms of employee participation were used in more than 90% of Fortune 1000 companies Although, in general, there is empirical support for the relationship between task autonomy and performance We believe it would be beneficial to develop a comprehensive model of the causal linkages between individual task autonomy and performance and to explore multiple mechanisms or processes that explain how task autonomy influences performance as well as the boundary conditions of these influences. Thus, the phenomenon of interest that our theory attempts to explain is the relationship between individual task autonomy and performance. As such, we are interested in factors that explain how and why autonomy affects individual performance. To these ends, we propose that there are three distinct mechanisms by which individual task autonomy may affect performance, namely motivational, informational, and structural. Further, we propose moderators that operate on these relationships to explain how and why the performance benefits of autonomy are contingent on factors ranging from individual differences to task structure. In addition to extending theory, our model also carries significant practical implications, as it informs organizational decisions about when, how, and to whom to grant autonomy. In describing the development of the theory, we first discuss how task autonomy relates to task performance via motivation (motivational mechanisms). In terms of this motivational effect, we also propose that the relationship is moderated by a variety of factors, which are based on both individual traits and situational states. Second, we explore how informational benefits derive from task autonomy (informational mechanisms), which depend on the amount of unique task-relevant information held by the individual and the complexity of the task itself. Third, we discuss the extent to which the effects of autonomy on performance depend on structural features of the task or organization (structural mechaClaus W. Langfred, Olin School of Business, Washington University; Neta A. Moye, Owen Graduate School of Management, Vanderbilt University. Both authors contributed equally to the article. We are very grateful for the comments and advice of the GOMERS research group at Washington University (Bill Bottom, Stuart Bunderson, Kurt Dirks, Chris Long, Judi McLean-Parks, Ryan Quinn and Raymond Sparrowe). We also thank Greg Oldham, Peter Kim, Michael Pratt and especially Bruce Barry for their excellent advice and suggestions. Correspondence 2004, Vol. 89, No. 6, 934 -945 0021-9010/04/$12.00 DOI: 10.1037 934 nisms), including task interdependence, task variability, and organizational formalization. Finally, the implications of our extended model for job design and organizational decisions regarding autonomy granting are discussed. Task Autonomy Defined In this article, task autonomy is defined as the degree to which an individual is given substantial freedom, independence, and discretion in carrying out a task, such as scheduling work and determining procedures to follow Current Theory on the Relationship Between Task Autonomy and Performance Drawing from the job design literature, the generally accepted causal mechanism linking task autonomy to task performance is motivation. The most explicit model of a motivational effect of task autonomy can be found within Much has been learned about the relationship between autonomy and performance from the research surrounding the job characteristics model. For researchers attempting to explore why task autonomy in reality only has a modest effect on performance We have developed a more specific, mid-range theory regarding task autonomy and performance. We incorporated the ideas from the job characteristics model but moved beyond the motivational mechanisms derived from that model by including two other mechanisms by which task autonomy influences performance, that is, informational and structural. Before discussing these two more novel mechanisms, we review the more commonly regarded mechanism by which task autonomy is expected to influence performance: motivation. Our Extended Model: Motivational Mechanisms As noted above, the job characteristics model provides a motivational explanation for how task autonomy relates to performance. More specifically, according to Proposition 1: The relationship between task autonomy and performance is mediated by motivation such that greater task autonomy leads to higher performance by increasing motivation. Although Proposition 1 suggests that all individuals will be more motivated when granted more autonomy, individual differences can potentially alter this relationship. Knowing that needs fulfilled by task autonomy vary across individuals begs the question of whether they also vary within individuals across situations. To a certain extent, this issue is one of trait versus state. Traits are individual differences that are stable across a wide variety of situations. States are differences that are evoked or activated by particular situations and are, therefore, more transient in nature. In previous research on autonomy, variations in needs or desires for autonomy have been treated as a combination of state and trait. More specifically, Wageman (1995) defined a construct referred to as autonomy preferences. She uses the term preferences specifically as it represents personal characteristics that are more stable than states but can be influenced by experience SPECIAL SECTION: TASK AUTONOMY AND PERFORMANCE We contend that there are multiple factors on which individuals may differ that will affect the relationship between autonomy and performance. These differences are caused by both general individual personality traits and situational factors specific to each job or task. Trait-Based Determinants The influence of more enduring traits is expected to potentially moderate the motivating effect of task autonomy on performance. Whereas In our model, both need for autonomy and need for achievement are expected to influence the extent to which individuals are motivated by task autonomy. The need for autonomy Proposition 2: Need for autonomy moderates the relationship between task autonomy and motivation; when need for autonomy is high, the relationship between autonomy and motivation is positive, and when need for autonomy is low the relationship between autonomy and motivation is null or negative. Proposition 3: Need for achievement moderates the relationship between task autonomy and motivation; when need for achievement is high, the relationship between autonomy and motivation is positive, and when need for achievement is low the relationship between autonomy and motivation is null or negative. State-Based Determinants To generate a more specific, mid-range theory of task autonomy, we believe it is important to consider more contextually specific individual differences as well. As such, we have incorporated several state-based individual differences. We draw largely on Bandura's (1997) framework of social-cognitive theory to propose relationships about how context might influence the motivating effect of task autonomy. Social-cognitive theory (Bandura, 1997) suggests that people exercise control in order to realize benefits. According to The potential perceived benefits and costs of autonomy are numerous. One example of a benefit would be the case in which an individual considered productivity a desirable outcome and perceived that autonomy would allow him or her to be more productive. Other benefits include physical and psychological well-being The notion that human beings pursue and prefer things that they believe benefit them and will avoid things that they believe harm them underlies many theories of human behavior, from socialcognitive theory to microeconomics. It is important to note, however, that perceptions of benefits and costs are subjective. Two employees faced with the same stimulus may interpret it differently; one may interpret it to be positive and the other may interpret it to be negative. Thus, the exact same task autonomy can 936 LANGFRED AND MOYE be perceived and framed differently by two employees; one may perceive autonomy as an opportunity to perform and excel at work in order to impress superiors, whereas another might perceive it as a shameless attempt by management to extract more work and effort for the same pay. As a result, the motivational effects of such different perceptions are likely to be very different. In addition to the subjective nature of perceptions, there is also the possibility of inaccurate perceptions of the costs and benefits of the task, which could result from a lack of information or understanding about the task itself on the part of the individual. This notion of the effects of the utility of task autonomy has been observed before. Proposition 4: Perceived utility of task autonomy moderates the relationship between task autonomy and motivation; when the perceived utility is high, the relationship between autonomy and motivation is positive, and when the perceived utility is low, the relationship between autonomy and motivation is null or negative. Using Bandura's (1997) framework also leads to the suggestion that people will be motivated by autonomy only if they believe they have the ability to take advantage of it. If an individual does not believe that he or she can successfully perform the job with increased autonomy, then he or she will most likely not be able to secure the perceived benefits. Therefore, beyond utility, an individual's self-efficacy with regard to exercising task autonomy will contribute to his or her desire for autonomy and any resultant motivational effect. For example, a skilled software engineer who has written code for particular types of applications for many years is much more likely to want discretion in the job than an inexperienced new employee who has just completed training and is writing code alone outside the classroom for the first time. The experienced employee is much more likely to believe that he or she can successfully take advantage of task autonomy to improve his or her performance and is therefore more likely to express a preference for autonomy than is the second employee. The new employee, on the other hand, will have lower self-efficacy for a new and relatively unfamiliar task and is likely to not want autonomy over the task but rather to prefer more structure and guidance. Proposition 5: Self-efficacy moderates the relationship between task autonomy and motivation; when an individual's self-efficacy is high, the relationship between autonomy and motivation is positive, and when an individual's self-efficacy is low, the relationship between autonomy and motivation is null or negative. The propositions above are all contingencies that suggest the importance of "fit" between perceptions, values, or personality traits of individuals and the level of autonomy in their task. This notion of fit is conceptually similar to other models in organizational behavior and leadership, such as situational leadership In summary, we propose that motivational effects of task autonomy are contingent on a variety of factors that reside at the individual level, representing both state-and trait-based individual differences that differ both across and within individuals. As such, the relationship between individual task autonomy and performance is dependent on both situationally specific variables (perceived utility of task autonomy and self-efficacy of exercising task autonomy) and more enduring personal characteristics (need for autonomy and need for achievement). Ideally, the amount of autonomy granted to an individual by the organization would be consistent with all of these moderators to maximize motivational benefits of autonomy and to minimize any possible motivational losses. However, motivation is just one of three paths between task autonomy and performance. We propose that another means by which autonomy affects performance is through informational mechanisms. Our Extended Model: Informational Mechanisms Beyond motivational mechanisms, we suggest that autonomy affects individual performance because of issues related to information and decision making. Although not explicitly stated in task autonomy literature, an informational mechanism has been explicitly developed in the related literature on participation The logic of this informational mechanism is compelling regardless of whether the issue is the sharing of information between employee and supervisor (as in participation) or the employee is being given discretion or decision-making authority (as in autonomy). If the software engineer has better information than the supervisor about effectively programming code for that particular project, granting substantial autonomy to her would seem to be a good way to ensure that she can best take advantage of that knowledge. Taking advantage of that knowledge means that better decisions are made by the employee regarding how and when to complete the task. These informational benefits of autonomy are entirely separate from any motivational benefits Considering only the informational mechanism, the relationship between autonomy and performance is dependent on the informa-937 SPECIAL SECTION: TASK AUTONOMY AND PERFORMANCE tion asymmetry between the subordinate and the supervisor; the more task information the individual has, the greater the potential performance benefit of granting autonomy to that individual. If the individual has no more knowledge about the task than the supervisor, there will be no informational benefits from granting autonomy, as there is no information asymmetry to be exploited. Proposition 6: The relationship between autonomy and individual performance is dependent on information asymmetries between the subordinate and the supervisor such that the greater the asymmetries, the more positive the relationship between autonomy and individual performance. Whereas the performance benefits of granting autonomy to those individuals with the best information about the task seem self-evident, the introduction of the informational benefits of autonomy raises questions as to what other cognitive effects may also occur. The fact that negative effects of autonomy on performance have been found when individual, contextual, and motivational differences are controlled for The essential difference between a worker with task autonomy and a worker without task autonomy is that the former has the opportunity to make decisions about task implementation, whereas the latter simply carries out the task as instructed. We suggest that the worker with autonomy is more cognitively distracted from the performance of the task, as he or she is involved in two distinct activities simultaneously. One of these is the activity of performing the task itself. The other is the cognitive activity involved in evaluating earlier choices and making decisions about future choices. With attention focused simultaneously on task performance and decision making, individuals with autonomy experience additional cognitive activity that distracts them from performance of the task. We base this idea of cognitive distraction on the literature on dual-task processing. Dual-task processing refers to the cognitive activity that occurs when someone attempts to perform multiple tasks, either simultaneously or consecutively in rapid alternation These findings regarding dual-task processing have been related to the managerial term of multitasking as well We suggest that the attention paid to task decision making that comes with autonomy is a form of dual task processing (or multitasking), as autonomy changes a job from the single task of performance to the dual tasks of performance and evaluation and decision making. We believe that these two tasks, occurring either simultaneously or consecutively in rapid alternation, cause dualtask processing effects. It is our argument that the cognitive processes that occur when one is engaged in this autonomy-related dual task processing distract cognitive resources away from performing the task itself and that these distraction effects are heightened when tasks are more complex. This cognitive distraction then leads to lower task performance. On the basis of these concepts, we propose that when individuals are given autonomy, cognitive distraction will occur, and attention will be drawn away from task performance. Further, the more complex the task, the greater we expect the cognitive distraction to be. Proposition 7: Task autonomy leads to cognitive distraction, as cognitive resources are spent on decision making, evaluation, and switching costs. Proposition 8: The cognitive distraction associated with autonomy increases in proportion with the complexity of the task. We further argue that the cognitive distraction associated with task autonomy will detract from task performance, as proportionally less cognitive attention is focused on the actual performance of the task. Proposition 9: Increases in cognitive distraction resulting from autonomy lead to decreases in performance. It is worth noting, however, that the cognitive distraction effects outlined above may decrease in severity as learning increases. It is especially during the early stages of tasks that individuals with autonomy will be experimenting with different ways to carry them out and will suffer the most cognitive distraction. Similarly, errors and poor performance are likely to occur more often early in tasks and will be greatly reduced as familiarity with the task grows. Overall, we propose that the informational benefits of autonomy discussed previously will co-occur with the downside of cognitive distraction. We are not suggesting that the informational benefit of autonomy does not exist. Rather, we are drawing attention to the cognitive costs occurring simultaneously with the informational benefit. The overall information-related effect of autonomy on performance is thus the sum of the informational effect, which can 938 LANGFRED AND MOYE range from zero to a large positive value, and the cognitive distraction effect, which can range from zero to a large negative value. Understanding these informational mechanisms is but one element of the larger question of how autonomy relates to performance. These mechanisms operate independently, but cumulatively, in relation to motivational mechanisms. That is, being allowed to make task-related decisions will not only have the informational effects noted above but simultaneously may have the effect of influencing motivation, depending on the motivational mechanisms at work. In addition, we believe there is a third set of mechanisms that will alter the relationship between task autonomy and performance, namely, structural mechanisms. Our Extended Model: Structural Mechanisms When exploring the relationship between autonomy and performance, it is important to do so in the context of characteristics of the task and the organization. Benefits of task autonomy may depend on a variety of structural mechanisms, related to both the task itself and the organization in which it is embedded. One task may be characterized by a particular structure that is incompatible with a high level of task autonomy, for example, whereas another task might have a structure that is especially suited to task autonomy. In one instance, autonomy might harm task performance, while in another it may have significant benefit, or even be necessary for successful task completion. Thus, the effects of autonomy on performance are contingent on a number of features associated with the task and the organizational structure, and we therefore consider task and organizational structures to be important mechanisms in our theory. We elaborate on two categories of structural mechanisms-task and organizational structure-in the following sections. First, we propose two structural characteristics of the task that will influence the effect of autonomy on performance via task processes: task interdependence and task variability. Second, we consider a structural characteristic of the organization, formalization, and how it influences the effects of autonomy on performance. Task Structures Although tasks can be described in a variety of ways, we believe that task interdependence and task variability are of particular importance to the relationship between autonomy and performance. Both have strong implications for how work processes should be structured in organizations, and both are key features of tasks in organizational units (in the sense that all tasks can be described in terms of interdependence and variability). Task interdependence represents the degree to which completion of a task is dependent on coordination with other tasks in order to be accomplished. It is a feature central to organizations It has been noted that individual autonomy can be at odds with tasks that require high interdependence within units, as such tasks require high levels of interaction and close coordination of members' actions in timing and sequence In contrast, when tasks are characterized by lower interdependence, individuals can work relatively independently, and greater individual autonomy would allow them to take advantage of unique task-specific knowledge (the informational mechanism discussed above) without interfering with any required coordination with others. Orton and Weick (1990) pointed out how important it is for individuals to adapt and independently adjust to change when interdependence is low, which individual autonomy allows team members to do. Similarly, Stewart and Barrick The logic outlined above suggests that the benefits and costs of individual autonomy can be very dependent on task interdependence. Although granting individual autonomy when task interdependence is low allows individuals to take advantage of informational and motivational benefits of autonomy without any loss in performance associated with process or coordination losses, granting it under conditions of high task interdependence can result in lowered performance, as necessary coordination with other organizational members is jeopardized. Proposition 10: Task interdependence moderates the relationship between task autonomy and performance such that the performance benefits of task autonomy are greater when task interdependence is lower. We expect a disordinal moderating effect of task interdependence on the relationship between individual autonomy and performance, suggesting, among other things, that individuals with high individual autonomy and high task interdependence will perform worse than those with high individual autonomy and low task interdependence. A second characteristic of task structure that we believe influences the performance effects of autonomy is task variability. Task variability refers to the number of exceptional cases in the work requiring different methods or procedures for doing the work We believe that task variability will influence the extent to which autonomy has positive process-related effects on performance. Individuals performing highly variable tasks need to be able to change or modify their procedures to successfully complete those tasks. When individuals have task autonomy, they can do so efficiently, as they have discretion over the implementation of tasks. When individuals do not have task autonomy, however, they either cannot complete tasks characterized by high variability or they must wait for new instructions from their supervisor or manager before proceeding. The combination of high variability and low autonomy is highly inefficient at best and completely dysfunctional at worst. Thus, when task variability is high, some degree of individual autonomy is necessary to successfully carry out the task. On the other hand, if task variability is low, individual task autonomy is not necessary and the absence of it will not be catastrophic. In that instance, task autonomy may not have great benefit but is unlikely to cause harm, as it can in the case of high task interdependence. Therefore, we suggest that task autonomy moderates the relationship between autonomy and performance such that the benefits of task autonomy are greatest when task variability is highest. Proposition 11: Task variability moderates the relationship between task autonomy and performance such that the performance benefits of task autonomy are greater when task variability is higher. Organizational Structure In addition to the structural characteristics associated with the task itself, we suggest that there are contingencies resulting from organizational structures, such as the degree of formalization the organization imposes on tasks. Whereas task interdependence and variability are driven by the task, organizational structures can be driven by a variety of other factors, such as management beliefs and preferences or the organization's culture, history, or degree of bureaucratization. We consider the degree of formalization in the organization to be a characteristic of organizational structure that is of critical importance. Formalization refers to the degree of job codification and rule observation High task formalization presents a problem when granting task autonomy, in that it limits the extent to which the individual can actually exercise discretion in carrying out the task. The more formalized a particular task, the less discretion is possible in its implementation. This has been demonstrated in a variety of settings, particularly with respect to professionals and knowledge workers Beyond the issue of the exercise of discretion, there are other possible negative effects of autonomy under the condition of high formalization. Although it may be possible to exercise autonomy in carrying out a task when formalization is high, doing so would violate organizational rules and procedures and possibly work-unit norms and values. The findings that groups and organizations sanction those who violate norms are consistent, ranging from the literature on small groups These considerations suggest the following proposition: Proposition 12: Task formalization moderates the relationship between task autonomy and performance such that the performance benefits of task autonomy are greater when formalization is lower. It is worth noting that The influence of task and organizational structures outlined above suggests that the realization of positive performance effects from autonomy will partially depend on the structural context in which autonomy is granted, encompassing both task and organizational structure. Again, we are not suggesting that autonomy will not have motivational or informational benefits but rather that these benefits must be balanced against any process-related costs that might emerge if autonomy is granted in a structural context 940 LANGFRED AND MOYE that does not fit, such as high task interdependence, low task variability, and/or high formalization. Discussion In this article, we have developed theory regarding the relationship between task autonomy and performance. Existing models and studies of task autonomy have not only fallen short of providing a complex model incorporating several different mechanisms but also have not addressed the mechanisms that might contribute to null or negative effects of autonomy. We have attempted to remedy those shortcomings by presenting a model that combines multiple mechanisms and explores multiple contingencies by which task autonomy affects performance. In particular, our model illustrates a variety of ways in which autonomy can have null or negative effects on performance, processes that have not previously been theorized or studied. Overall, we argue that the effect of task autonomy on performance is the sum of three independent mechanisms. One mechanism is the motivational effect of autonomy, which is contingent on state-and trait-based individual differences. Another mechanism is the informational effect, as autonomy may capitalize on the asymmetry of information between those who do the work and those who supervise it, although this positive effect may be limited by increased cognitive distraction. Finally, structural features of the task may enhance or limit the process benefits of task autonomy, as illustrated in our discussion of the moderating effects of task interdependence, task variability, and organizational formalization. We believe that our elaboration and integration of current thinking creates a much more complete picture of the relationship between task autonomy and performance and illuminates whyand under what conditions-task autonomy will and will not lead to performance gains. With this model, we hope to bring the theory of autonomy closer to the reality of practice. In the next sections, we discuss the general research contribution and implications of our theory. We then discuss the implications for practice of this expanded view of task autonomy and performance and attempt to clarify the issue of when and under what conditions organizations should grant autonomy to individuals. Finally, we include suggestions for empirical research and thoughts on measures and possible study designs. Research Contribution and Implications By exploring motivational, informational, and structural mechanisms in the same model, we emphasize the balance among various forces in their effect on performance. When considering individuals in specific situations and contexts, the general need or desire for autonomy and control can be overwhelmed by situational factors, for instance. One example of this is self-efficacy. Given a particular task, an individual with considerable task experience and skill is much more likely to want more control over various aspects of the task than is an individual who has no experience at the task. Overall, it can be said that we are presenting an interactionist view of the motivational effects of autonomy, driven by both internal traits and situational factors specific to the setting and context. We believe that using both views enriches the construct, allowing for further exploration of how both stable individual differences and the context in which an individual works might influence motivation. We also add to theory on task autonomy by considering the informational mechanisms linking task autonomy to performance. We adopted ideas from the participation literature by formally incorporating the concept that granting task autonomy can have positive effects on performance by capitalizing on information asymmetry between supervisor and worker. An important implication of this informational effect is that it offers an explanation for situations in which there may be null effects of task autonomy on performance, specifically when the individual granted autonomy does not have better information than the supervisor and thus has no information asymmetry on which to capitalize. Incorporation of this effect may be primarily a formalization of a generally accepted or implicit effect of autonomy but offers a more complete formal theory of how task autonomy relates to performance. In addition, incorporation of contingent motivational effects as well as the informational effects of task autonomy allowed us to describe how autonomy can have negative effects on motivation and performance. With respect to informational effects, we proposed that task autonomy redirects cognitive attention away from the focal task to decision making and evaluation. This cognitive distraction is always going to occur, a natural by-product of task autonomy. Cognitive distraction thus provides an important refinement to the informational mechanism. Although the current implication for informational effects is that without information asymmetry, there will not be a positive effect of task autonomy on performance, the inclusion of cognitive distraction suggests that in fact there will be a negative effect. In other words, without the gain from information asymmetries, asking individuals to make decisions merely drains cognitive resources and distracts them from task performance. Finally, the addition of task and organizational structure to our theory of autonomy is a significant step toward developing a more comprehensive and complex model of autonomy. By taking into account the effects of structural task characteristics such as interdependence and variability, as well as organizational characteristics such as formalization, we are adding a new dimension to the study of individual autonomy. Although the organizational design literature has discussed the effect of task variability on how to structure an organization, for example, it has not been applied to individual task autonomy. It is also worth considering that there may be other organization structure factors that moderate the relationship between task autonomy and performance. The degree of access to organizational resources, for instance, may be a valuable future addition to this theory. In terms of developing this theory further, incorporating task structure and organizational structure into our model of individual autonomy also opens the door to considering cross-level determinants and contingencies of the relationship between task autonomy and performance. Implications for Practice Our expanded model of task autonomy has several implications for organizations or managers who wish to incorporate task autonomy, or have already incorporated task autonomy, in organizational settings. In general, our expanded model helps delineate the conditions under which autonomy might be most appropriate as well as the conditions under which it might be most inappropriate. SPECIAL SECTION: TASK AUTONOMY AND PERFORMANCE The model highlights the importance of considering multiple factors at once when trying to determine whether granting task autonomy will result in improved performance. The model also implies possible avenues for actions that could be taken to improve performance effects in situations in which task autonomy already exists. Supplementing the conventional wisdom of hiring people who desire autonomy, our model suggests that some of the contingencies affecting the motivational mechanism are partially malleable. Organizations can work to improve either individual's selfefficacy or their perceptions of the utility of task autonomy in order to reap better performance gains. From the rich research in social-cognitive theory (Bandura, 1997), there are many interventions that could be used to improve an individual's self-efficacy, including training, feedback on performance, and modeling. With regard to utility, organizations could attempt to influence the perceived benefits and costs individuals associate with the introduction of task autonomy. For example, organizations could consider increasing pay, decreasing hours, or other means for attaching benefits to the introduction of task autonomy (as long as such incentives are framed as being directly related to the introduction of autonomy). Such interventions, and the resulting effects on self-efficacy and utility perceptions, are a means for increasing the motivating effect of task autonomy. Most important, our model points organizations in these new directions and gives them additional options for improving the performance effects of task autonomy. It is worth noting that although an organization might be able to manipulate employees' preference for (and motivation from) autonomy with some of the interventions discussed above, this does not suggest that the organization should not also select employees on the basis of personality traits. If the organization has a particular task structure that is particularly well suited to individual task autonomy (e.g., the combination of low task interdependence and high task variability), then the organization should definitely employ a selection system to identify employees who have a strong trait-based preference for autonomy. The practical implications of the costs of cognitive distraction involve a closer look at the allocation of autonomy in two ways. First, it is very important to balance the expected benefits of autonomy against this performance loss. That is, in situations where individuals do not have adequate information to make decisions, the introduction of task autonomy may actually have overall negative effects due to cognitive distraction. In addition, in tasks where cognitive attention demands are high (e.g., safety or inspection jobs), the distracting effects of autonomy may be difficult to overcome. Therefore, organizations need to carefully consider both the nature of the individuals being granted task autonomy (how informed, how well trained, how prepared to make decisions) as well as the nature of the task (complex task, high cognitive attention demands). In general, our model points to additional avenues for organizations to consider for improving the informational gains of introducing task autonomy. The structural contingency of task interdependence has clear, but not necessarily easy-to-implement, implications. If the task technology dictates high levels of task interdependence in the organization, in which individuals' work is tightly coupled, then organizations need to be very careful about implementing task autonomy, and must try to balance the structural cost against other potential motivational or informational benefits. By the same token, the contingency effects of task variability must be carefully attended to by organizations. If task variability is high, task autonomy is practically a requirement for successful task completion (in lieu of tight supervisory monitoring and continual feedback), and the cost of not granting autonomy to individuals under these conditions is expected to be considerable. The extent to which a task is formalized into rules and procedures also has clear implications for the granting of autonomy, to the extent that individual autonomy is clearly at odds with restrictive procedures and rules for carrying out tasks. The contingent nature of the relationship between autonomy and performance as it depends on these structural factors illustrates the complex balance implied by our model. The more comprehensive nature of our model is what makes it useful both for prescriptive purposes and to guide directions for future research. The notion of balance is crucial, as all three of the mechanisms can be occurring simultaneously, and the benefit of one mechanism may be offset by the cost of a second mechanism or, alternatively, completely dependent on a third mechanism. One contingency may have a positive effect, for example, whereas another simultaneously occurring contingency may have a negative effect. Depending on combinations and distributions of individual traits, motivational states, information asymmetries, and task and organizational structures, there are practically limitless combinations spawned by our model. Ideally, for a manager to be able to decide whether to grant a particular individual autonomy on a specific task, he or she would have perfect information about a range of different questions relating to how motivated the employee is by autonomy (both in general and for this specific task), how much task-specific knowledge the employee has about the task, how complex the task is, how interdependent it is with other tasks, how much variability there is (and is likely to be) for the task, and how formalized the rules and procedures governing the task are. Given this information, the manager then must be able to judge the relative strength of one mechanism versus another. For example, should autonomy be granted to an employee who values it and would be very motivated by it but does not have much task-specific knowledge on which to draw? Further, how would that balance be influenced by a moderate degree of task interdependence but a low degree of task variability? In general, we can certainly conclude that autonomy should definitely not be granted to an employee who is not motivated by autonomy and who has little knowledge about a complex and highly interdependent task that has low variability and many rules and procedures. We can also conclude that autonomy should definitely be granted to an employee who is motivated by autonomy and who has considerable knowledge about a relatively simple task that can be performed independently and has high variability but few rules and procedures. Although our theory can only provide general answers to these questions, it does present a clear framework within which to explore all three mechanisms simultaneously and a guide by which managers and organizations can understand the mechanisms that are important to take into account and the ways in which each one influences the relationship between autonomy and performance.