Breaking and creating habits on the working floor: A field-experiment on the power of implementation intentions

Abstract

Previous research has shown that implementation intentions are effective tools to promote new behavior. The present study aimed to provide the first evidence that conscious planning is an effective tool in replacing well-learned habits with new habits. This was tested in a field-experiment on repetitive behavior in the domain of recycling, using 109 employees of a tele-company as participants. Recycling behavior of the participants was observed by the actual amount of paper and the number of plastic cups in their personal wastebaskets. Following a pre-measure, participants were assigned to either implementation intention conditions, conditions in which an eye-catching facility was placed to promote recycling behavior, or control conditions. Recycling behavior was substantially improved in the facility as well as the implementation intention conditions in week 1 and week 2 and still 2 months after the manipulation. These data supported our hypothesis that planning breaks down unwanted habits and creates new ones.

Introduction

Habits are efficient modes of goal-pursuit. When our goal-directed behaviors are well-learned through repetition in stable environments, goal-relevant situational cues may automatically elicit these behaviors. This way, we may attain our goals without conscious thought. However, when we try to change our habits, e.g., because other contradictory goals or action programs have become more important in the habitual situation, it seems very difficult to alter our automatic habitual responses (Aarts and Dijksterhuis, 2000a, Heckhausen and Beckmann, 1990, Reason, 1990).

What kind of strategies would be successful for breaking unwanted habitual behavior and creating ‘wanted’ habits? Habits are considered to be situationally guided goal-directed behaviors, and hence, behavioral responses are automatically elicited when the situation arises (Aarts and Dijksterhuis, 2000b, Bargh and Gollwitzer, 1994). Theoretically, then, strategies to successfully break habits should fit in with these habitual processes. In the present paper, we will investigate two potentially powerful tools to change habitual behavior. The first refers to conscious planning. Strikingly resembling habitual processes, recent studies have shown that after forming an implementation intention, a goal-directed behavior becomes strongly linked to situational cues and becomes automatically activated because of these situational cues. The processes underlying effects of planning suggest that the formation of implementation intentions may be a strong tool in order to break habitual behavior and create new habits.

Second, the importance of situational cues in the onset of habitual behavior suggests that eye-catching changes in the situation may affect habitual behavior by facilitating the performance of alternative behaviors and turn them into new habits. Therefore, we also investigate the installation of such facilities as environmentally prepared cueing tools in breaking old unwanted habits and creating new ones, and to compare its effects with mentally prepared cueing—that is, by conscious planning.

Gollwitzer, 1993, Gollwitzer, 1999 distinguishes goal intentions, which refer to intentions to achieve a certain goal, from implementation intentions, which refer to intentions that specify, where, when, and how these goals are acted upon. Goal intentions have the form of “I intend to do X (“I intend to exercise”) whereas implementation intentions follow the proposition “In situation Y, I will do Z” (e.g., “When I come home from work, I put on my sneakers and go for a run”).

Several studies have shown that furnishing goals with implementation intentions can have a strong effect on goal-pursuit. Typically, participants are first instructed to attain a certain goal and then asked to write and visualize when, where and how they will act on that goal. The impact of implementation plans on behavioral performance has been shown across various behaviors, ranging from completing an assignment (Gollwitzer & Brandstätter, 1997) to shopping at a bioshop (Bamberg, 2002; see for overviews Gollwitzer, 1999, Gollwitzer and Sheeran, in press) and can be stunningly strong. For example, Orbell, Hodgkins, and Sheeran (1997) showed that participants who had formed implementation intentions for performing breast self-examination performed at the level of 100%, compared to 53% for control participants.

Subsequent studies have provided more clarity with regard to the mechanism producing these strong effects of planning. These studies revealed striking similarities in the processes underlying planning and habits. For example, Aarts and Dijksterhuis (2000b) established that non-habitual travel mode users who formed implementation intentions as to using the bicycle showed automatic bicycle responses to travel goal situations, just as habitual bike users. Forming implementation intentions creates a strong mental link between a situation and a behavioral response. Therefore, when the person encounters the goal-relevant situation specified in the implementation intention, the goal-directed behavior is automatically initiated (Aarts and Dijksterhuis, 2000b, Brandstätter et al., 2001, Webb and Sheeran, 2003).

Furthermore, forming implementation intentions increases the mental accessibility of situational cues, and thus facilitates the activation and subsequent execution of the associated behavior (Aarts, Dijksterhuis, & Midden, 1999). Hence, the onset and the proceeding of behavior have been delegated to the environment. Both habits and implementation intentions thus seem to consist of automatic behavioral responses elicited by situational cues. The former as a result of repeated actions in a stable environment, the latter as a result of a single act of planning (see also Bargh and Gollwitzer, 1994, Orbell et al., 1997).

The similarities between the mental representation of habits and implementation intentions suggest that implementation intentions are potentially powerful tools to create new habits. If so, forming implementation intentions renders the desired and planned behavior more likely to be performed frequently over time. Tests for these effects are generally lacking.

In a study of Sheeran and Orbell (1999) on vitamin C supplement intake, it was shown that planning affected repetitive behavior across a period of 2 weeks. However, because participants were aware of the fact that their behavior was monitored, it is possible that their results were caused by processes of public commitment rather than planning. Studies on public commitment have shown that behavior is more in line with goals or intentions when (1) these intentions are explicitly stated in front of other people (e.g., a group or specific person) and (2) one is aware that these individuals will monitor the behavior (e.g., Schlenker, Dlugolecki, & Doherty, 1994). For participants within the implementation intention condition, both elements of public commitment were present. First, participants in the planning conditions expressed their intentions to take the vitamin pills in the presence of the experimenter, by responding to the questions concerning where, when and how they planned to take the vitamin pills. Second, they knew that the experimenter would visit the participants in their dormitories and, in the presence of the participants, would count the number of pills that were actually taken. This enhanced public behavioral commitment may have induced all kinds of motivational differences (e.g., self-presentation motives) that may explain the differences between the implementation intention and the control condition.

Thus, a strong test of the idea that planning may create habits is as yet not available. To convincingly demonstrate this idea, a study should contain a measure of overt behavior that is frequently performed. Furthermore, and importantly, participants should not be aware of the fact that their behavior is being monitored.

Is it also possible to break habits¹ by forming implementation intentions? Nearly, all studies on planning thus far focused on the facilitation of (new) behaviors (e.g., breast self-examination, completing an assignment). In these cases, the situations that are part of the plans are not yet strongly tied to a behavioral response (e.g., people may behave in these situations in multiple ways rather than showing one fixed response). In other words, prior to the planning, the context does not automatically elicit a behavioral response. In the case of a habit, however, the situation automatically instigates the habitual behavioral response. Hence, it is more difficult to override strong habitual responses than to learn a new behavior in a situation that was not strongly linked to a behavioral response. Indeed, it has been proven to be very difficult to break habitual behavior by means of changing goal intentions. For example, interventions aimed at changing habitual health and environmental behaviors are often unsuccessful, even though people have positive intentions to alter their behaviors (Aarts et al., 1997, Aarts et al., 1998, Gifford, 1997, Verplanken and Faes, 1999).

Although several authors have stressed the possibility that forming implementation intentions may break habits, there is no direct empirical evidence supporting this claim. Verplanken and Faes (1999) showed that forming implementation intentions to eat healthy, resulted in more healthy food intake, but it failed to overrule bad eating habits. Their null-findings can be considered as another example of the difficulty of breaking habitual behavior and may suggest that implementation intentions, though powerful tools to facilitate new behaviors, cannot break habits.

However, the absence of habit change in the Verplanken and Faes (1999) study may be interpreted in a different way. Participants formulated implementation intentions directed at a broad, abstract goal (“to eat healthy”). Because of the generality of the abstract goal, the implementation intentions may not directly compete with concrete bad eating habits in specific situations (e.g., eating a candy bar everyday at 4 p.m. during the break at school). In other words, by planning their healthy behavior, people probably linked new behavioral responses to situations (e.g., eating fruit at home) that were different from the habitual situations.

These ideas thus suggest that a prerequisite for planning to be effective in altering habitual behavior, is that planning links a new behavioral response to the same situational cues that elicit the habitual behavior. After planning, the situation now activates both the habitual and the planned response, which may compete for action control (Macrae and Johnston, 1998, Norman and Shallice, 1986). As implementation intentions are more recently installed, the planned response may be more accessible and, at least temporarily, override the habitual response.

In addition to these theoretical notions on the circumstances that make planning a strong candidate to alter habitual behavior, there are also several necessary empirical elements that should be included to provide convincing evidence for the idea that planning breaks habits. One such crucial element is a measure of behavior over a long period of time. Bamberg (2002), for example, showed that habitual non-public transport users were more likely to use a new bus route after planning. Behavior was observed only once at the time the participants used their free bus-tickets that were provided by the experimenter. Since no subsequent behavior was measured, it is not clear whether habits were truly “broken.” It is very well possible that participants returned to their old habits after getting their single time free ride.

To sum up, to convincingly show that planning can break habits and people do not return to their old habits, it should be demonstrated that (1) the behavioral pattern to be broken is habitual in nature; (2) planning results in behavior change; (3) this behavioral change is enduring and stable; and (4) the link between previous behavior and future behavior is stably reduced.

Another potentially powerful way to alter habitual behavior is by changing the situation. Obviously, people are forced to act differently if their habitual behavior is blocked by changes in the situation. However, changes in the situation may also alter habitual behavior if the new situation facilitates alternative responses. For example, the very same person who habitually throws his old paperwork in a regular dustbin instead of recycling it by using a special paper-recycling box, may be propelled to do so if such a recycling box is placed at his own desk. Indeed, vivid reminders or “prompts” have been shown to turn trashy habits into pro-environmental action (Zimbardo & Leippe, 1991; see also Intons-Peterson & Newsome, 1992). Moreover, such attention-grasping facilities are likely to cause behavioral changes that should be stable and observable over a longer period of time, that is, they can turn a person’s paper-in-the-dustbin habit into a new paper-recycling habit (Aarts et al., 1997). In the present study, we thus investigated such a change in the situation as a tool to break old habits and create new ones.