Habit and Intention: Multiple Processes 1
Habit and Intention: The Multiple Processes of Behavior Generation in Everyday Life
Wendy Wood Jeffrey M. Quinn
Texas A&M University
Preliminary draft of a chapter to appear in: J. Forgas, W. VonHippel, & K. Williams (Eds.), Sixth annual Sydney Symposium on Social Psychology.
Please do not cite or quote.
Habit and Intention: Multiple Processes 1
Habit and Intention: The Multiple Processes of Behavior Generation in Everyday Life
People seem to believe that much of their behavior is under their own control and that it reflects their intentions and plans. Everyday understanding of the mind-body problem is a pretty straightforward link between cognition, affect, and action (see Wegner, 2002). Thus, acting without thinking would seem to be the stuff of social gaffs and apologies.
Yet, absence of thought about actions may be a common occurrence in everyday life. Behavior can be guided through several processes, and these vary in the amount of attention they require (Heckhausen & Beckmann, 1990; Ouellette & Wood, 1998; Wegner & Bargh, 1998). With novel behavior or behavior in unfamiliar contexts, the uncertainties associated with performance require that people continuously attend to and evaluate new information as it is presented in order to respond appropriately. In contrast, for frequently performed behaviors in stable contexts, behavioral intentions become implicit as individual behaviors come to be incorporated into sequences of multiple actions (Ouellette & Wood, 1998; Wood, Quinn, & Kashy, 2002). Then recurring environmental cues can activate practiced behavior routines in a process that occurs largely outside of people=s awareness (i.e., Aat one extreme...accompanied by a complete mental blank,@ Heckhausen & Beckmann, 1990, p. 38). Thus, given the appropriate contextual cues, behavior can be initiated and performed with minimal, sporadic thought.
Habit and Intention: Multiple Processes 1
In the present chapter, we propose that the implicit processes guiding habit performance can proceed separately from explicitly intentional action. To the extent that these are independent processes, habitual and intentional performance modes will each be affected by a unique set of factors (i.e., in addition to such common factors as a person=s skill set). In addition, the two modes can generate separate, conflicting guides to action, as evidenced by action slips and other unintended behaviors (Hay & Jacoby, 1996; Reason, 1979). These errors illustrate yet another feature of the habitual and intentional systemsBalthough they proceed simultaneously, they also can influence each other in the generation of behavior (see Smith & DeCoster, 2000; Wilson, Lindsey, & Schooler, 2000). Not only can habits interfere with the best laid plans, but also, given sufficient effort and opportunity, the intentional system can override well-practiced behavior.
How Habits Guide Behavior
We define habits as behavioral tendencies to repeat well-practiced acts given stable contextual cues. Repetition of a behavior in a given setting promotes automaticity in the cognitive processing that initiates and controls the response, as the processing comes to be performed quickly, in parallel with other activities, and with the allocation of minimal focal attention (e.g., Posner & Snyder, 1975). Repetition of behavior can yield automaticity when people are intending to achieve some goal or when the repetition is unintended and people are unaware of what they have learned (Squire et al., 1993; see Lippa & Goldstone, 2001, for incidental development of automatic associations). Thus, habits that emerge from intentional repetition could be explicitly represented in memory as goal-action links (Aarts & Dijksterhuis, 2000; Verplankan & Aarts, 1999). Yet, our claim that people are often not aware of their habits and the contexts that trigger them suggests that implicit representations of habits will often not correspond to explicit understanding of behavior.
Habit and Intention: Multiple Processes 1
Stable contexts facilitate the propensity to perform repeated behaviors with minimal cognitive monitoring. Research on transfer of learning and on stimulus generalization have addressed the question of what makes features of stimuli and contexts interchangeable for learning and performance (e.g., Bouton, Nelson, & Rosas, 1999; Proctor & Dutta, 1993). For our purposes, contexts are stable to the extent that they present the same contextual cues integral to performing the response and to the extent that they are similarly conducive to fulfilling an actor=s goals. As Barker and Schoggen (1978) noted in their analysis of the genotype of behavior settings, contexts may vary in superficial attributes but be stable in the features supporting performance. Unstable contexts are ones in which shifts in the supporting environment implicate alternate goals or challenge the smooth initiation, execution, and termination of practiced responses.
How do stable environmental events cue behavior? In classic learning theories, features of the environment directly cue well-practiced behavior through stimulus-response linkages (e.g., Hull, 1943; Spence, 1956). However, more recent models of cognitive processing outline how external events mobilize action by automatically triggering behavioral intentions and action sequences, which then can be implemented with minimal thought (e.g., Bargh & Ferguson, 2000; Heckhausen & Beckman, 1990). With repetition of behavior, intentions become relatively abstract, specifying broad goals, and they become incorporated into sequences of multiple actions (Vallacher & Wegner, 1987). We suggest further that, even for actions that initially may have been explicitly goal-directed, intentions become implicit with repetition. Thus, habitual action sequences are relatively automatically cued by the environment, often outside of awareness, and they do not require motivation, intention, or cognitive capacity.
Memory Systems and Habits
Habit and Intention: Multiple Processes 1
Research in cognitive psychology and the neurobiology of memory systems is consistent with our claim that separate habitual and intentional systems guide behavior. At a cognitive level, habits represent slow-learning systems in which contextual cues, sequences of motor performance, and sometimes behavioral goals are structured in associative memory by similarity of experiences and their repeated contiguity. These learned sequences are then retrieved by a fast, automatic pattern-completion mechanism (see Smith & DeCoster, 2000). The associative processing mode can be contrasted with intentional systems that include rule-based processing and relatively conscious, explicit reasoning that is applied in either a deliberative, systematic manner or a more spontaneous, heuristic way.[1]
Neuropsychological research on memory systems also suggests that habits are associated with a particular memory system. This work has examined patients with brain lesions that yield selective memory impairment or it has used functional neuroimaging techniques to examine activation of brain regions during performance of behavioral tasks (see reviews in Schacter, 1992, 1995). The findings suggest that noncognitive habit and skill memory are linked to a complex of specific brain systems involving the basal ganglia, cerebellum, and motor neocortex (Gabrieli, 1998; Squire, Knowlton, & Musen, 1993). These differ from the systems associated with priming and other forms of nonconscious memory and from the systems involved in declarative, conscious memory for facts and events. Thus, the memory systems involved in habitual responses appear to be separable at a neurological level from those involved in more explicit modes of behavior generation.
In addition, support for our claim that habits are stored as larger action sequences rather than discrete acts was provided by Jog, Kubota, Connolly, Hillegaart, and Graybiel=s (1999) study of the sensorimotor striatum of rats during learning of a maze. Because neuronal responses after successful acquisition emphasized the beginning and the end of the learned procedure, these authors concluded that an action template was developed for the behavioral unit as a whole (i.e., the full maze), and this was triggered by specific contexts at the start and the end of the maze. Thus, habits tend to be integrated into action sequences that are cued as a unit by stable features of the environment.
Intentions and Habits Guide Behavior
Habit and Intention: Multiple Processes 1
Despite the strong evidence of the separate cognitive and neurological processes associated with habitual and nonhabitual behavior, the standard behavior prediction models in social psychology have focused on explicit, intentional guides to action (see Eagly & Chaiken, 1993). According to the best-known such model, the theory of planned behavior (and its precursor, the theory of reasoned action), behavior is a function of behavioral intentions; these intentions in turn reflect attitudes toward a behavior, perceptions of normative pressures, and perceptions of efficacy to perform a behavior (Ajzen, 1987, 1999; Fishbein & Ajzen, 1984). Research that has measured the relevant cognitions and then used these to predict behavior has provided strong support for such models (see meta-analytic reviews by Armitage & Conner, 2000; Randall & Wolff, 1994; Sheppard, Hartwick, & Warshaw, 1988). However, this evidence of predictive power does not necessarily speak to the cognitive and affective mechanisms through which behavior is generated.[2]
Predictive models, like laws, are Ageneral summaries of the way things happen...and are deduced from observations of specific things happening...processes, on the other hand, are the intermediary events that explain why things happened as they did@ (Gilbert, 1998, p. 103). Developing the theory of reasoned action to address process, Fazio (1990; Fazio & Olson, 2003) proposed that the theory holds primarily when people are highly motivated and have sufficient opportunity to deliberate about their intentions and behavior. When motivation and opportunity are limited and when attitudes are strong and automatically activated in memory, then people respond in a more spontaneous fashion and their attitudes and perceptions of social norms directly guide action. Although Fazio=s depiction of spontaneous action can be faulted for providing little insight into the relevant mechanisms (e.g., how do the automatically-activated attitudes toward a target direct specific behaviors, see Eagly & Chaiken, 1993), the model is noteworthy for addressing the dynamics of intention-behavior processes.
Habit and Intention: Multiple Processes 1
To extend predictive models to consider the dynamic impact of habits, Ouellette and Wood (1998) conducted a meta-analytic synthesis of the conditions under which behavior reflects conscious intentions versus habits. Specifically, habitual repetition should occur when a behavior has been performed with sufficient frequency in a stable contextBas should be the case with behaviors such as drinking coffee and wearing sealtbelts. In contrast, intentions should be a guide primarily when a behavior is novel or performed in an unstable contextBas with behaviors such as getting a flu shot or donating blood (Triandis, 1977). The top panel of the path model in Figure 1 displays the findings for studies in the synthesis that examined behaviors that could be performed frequently. As anticipated, people who had established habits for these behaviors simply repeated their past behavior and were minimally influenced by their stated intentions. The bottom panel in Figure 1 displays the findings for studies that examined behaviors that occur only a few times a year, and typically in unstable contexts. With these behaviors, people were likely to carry out their intentions, and the frequency with which they had performed the action in the past had little effect. This overall pattern of findings suggests that when people have sufficient opportunity to form habits, behavior is not a function of conscious intentions.
Additional demonstration that habit and intention provide separate guides to behavior has been provided by research that has examined behavior prediction for people with and without habits in a given behavioral domain (e.g., Albarracin, Kumkale, & Johnson, 2002; Ferguson & Bibby, 2002; Ouellette & Wood, Study 2). For example, in a study of car use, Verplanken, Aarts, van Knippenberg, and Moonen (1998) found that people who used their cars frequently and thus had established a driving habit did not base their travel decisions on their intentions to driveBinstead, they repeated their past behavior. In contrast, people who used their cars less often were guided by their intentions regarding mode of transport. Thus, habitual and explicit intentional processes appear to provide separate guides to behavior, and these guides interact in their effects such that intentions have minimal influence when habits have been formed.
Habit and Intention: Multiple Processes 1
Despite the independent effects of habit and intention on past behavior, people=s judgments of these factors appear to be linked. For example, in both of the path models in Figure 1, intention is correlated with past behavior. This relation could reflect a tendency for people to practice to habituation intended behaviors as well as a tendency for people to infer their intentions from their past behavior (e.g., reasoning, AI did it in the past, I will probably do it in the future@). Inference of intention from past behavior should occur especially when people are uncertain of their intentions (see Bem, 1972), a circumstance that is likely when behavior emerges from implicit processes. Thus, the relation between intention and frequency of past behavior plausibly reflects a variety of factors unrelated to the processes through which habits guide behavior.
Is past behavior a valid measure of habit? The evidence that behavior is generated through multiple processes does not challenge models of explicit intention so much as complement them by suggesting that implicit processes also guide behavior. Yet, a recurring question from the perspective of prediction models is whether frequent behavior in stable contexts reflects habit or some other variable that might already be represented in the model (e.g., Ajzen, 2001).
Participant reports of past performance frequency have been the standard measure of naturally-occurring habits (e.g., Ronis, Yates, & Kirscht, 1989; Triandis, 1977). Similarly, manipulations of frequent performance in stable contexts have been the standard procedure to establish strong habits in the laboratory (e.g., Hay & Jacoby, 1996; Shiffrin & Schneider, 1977). These two paradigms converge in suggesting that the processes guiding behavior are automatized with frequent practice, especially when the practice occurs in stable contexts (Schneider & Shiffrin, 1977; Wood, Quinn, & Kashy, 2002).[3]
Habit and Intention: Multiple Processes 1
Furthermore, suggesting that performance frequency effects cannot easily be dismissed as proxies for other, nonmeasured variables, habits have been found to maintain their predictive impact in designs that control for a variety of additional factors (e.g., attitude accessibility, self-concept, perceived behavioral control; Ouellette & Wood, 1998). Additional reassurance is provided by the pattern of effects associated with habits; although main effect findings that past behavior predicts future behavior would be vulnerable to explanation through nonhabitual factors, the interactive effects of habit and intention that emerge in specific conditions (see Figure 1, also Verplanken et al., 1998) are not easy to explain in terms other than automaticity.