The Rocky Road from Intentions to Task Completion

The Rocky Road from Intentions to Task Completion

Attitude researchers long ago identified the match between the predicted and actual “object of judgment” to be a key determinant of attitude–behavior consistency (LaPiere, 1934; Lord, Lepper, & Mackie, 1984). The construal problem faced by a planner is even more daunting. What will the target task actually look like day by day? What goals and values will be most salient? How will the planner’s emotions and motivations change over the course of the project? What other tasks might intervene, and how will they affect the perceived importance of the target task?

The analysis of temporal construal offered by Liberman and Trope (1998; Trope & Liberman, 2000) suggests that distant tasks (in time) are analyzed at a higher level of abstraction than are tasks close at hand. Thus, a student considering an honors’ research project from a year’s distance may construe the task as a problem of finding a good problem, getting to work on collecting data, and budgeting plenty of time for statistical analysis. In contrast, a student considering the same project from the distance of only 1 or 2 weeks may be more likely to construe the task as competing for a professor’s attention, searching for appropriate reference material, and booking laboratory rooms at a convenient time. In general, a more detailed and low-level construal of the task highlights difficulties, and so predictions and judgments made at a more distant time should tend to be more optimistic. Both the tendency to construe distant events at a high level of abstraction and the associated relation between temporal distance and optimism has been confirmed in a number of experiments (Liberman & Trope, 1998). This suggests an intriguing avenue of investigation for debiasing through encouraging more detailed, lower-level construals of projects.

Another difficulty in matching predicted intentions and actual behavior is the problem of emotional construal. A planner faces a double inference problem: (1) How to account for the effect of emotional states during the project period; and (2) How to adjust for the contaminating effects of current moods and emotions on the planning process. A person may anticipate that as the promised completion date looms, she or he will feel strongly committed to the goal of finishing early, give the project high priority, and so forth. Over time, however, people are likely to experience unforeseen changes in their hedonic states as a result of such basic hedonic processes as adaptation. Thus, the feeling of urgency associated with anticipated failure becomes less intense as the days go by, and a delayed completion – at first treated as a painful loss – gradually becomes the status quo. For example, an author may initially believe that he will be terribly embarrassed if his manuscript is not completed by a promised date; however, as he slips further and further behind his predicted pace, his frame of reference may change so that he feels only relief when his manuscript is completed within a year of the suggested day. That people may experience such unanticipated hedonic shifts is consistent with demonstrations of the problems people have in anticipating how their own feelings, preferences, and desires will change over time (Kahneman & Snell, 1992; Gilbert, Pinel, Wilson, Blumber, & Wheatley, Chapter 16, 2001, this volume; Wilson, Wheatley, Meyers, Gilbert, & Axsom, 2000; Loewenstein & Frederick, 1997; for a review, see also Loewenstein & Schkade, 1999).

Consider Loewenstein’s (1996) concept of the “empathy gap” that exists between different hedonic states. People with a full stomach underestimate the difficulty of dieting when hungry, presumably because their intention to diet is very strong when they are sated. Despite having considerable experience with failed diets, people do not adjust for the low validity of these predictions due to the empathy gap. Similarly, people who are not sexually aroused underestimate the likelihood of failing to use condoms, presumably because their intentions are strong and they do not realize that fluctuating levels of sexual arousal can lower the validity of such intentions. We suspect that people possess general theories about how their task-relevant hedonic experiences change over time (e.g., “By the time the kids are in bed, I’m often too exhausted to finish the things I want to do”; “I usually spend too much time on the fun parts and avoid the unpleasant parts”; “Once I start browsing the newsgroup, I’m lost for the evening”), but this set-based information about the effect of hedonic states on the predictive validity of plans is ignored in favor of a case-based evaluation of current feelings and intentions.

Elster and Loewenstein (1992) liken people’s future selves to their children at college; although people believe that such selves will continue to behave in accordance with present priorities and values, they tend to be less predictable and more open to temptation than we imagine. This analogy fits some additional findings from our comparison of planning biases in Japanese and Canadian students (Buehler et al., 2000). Participants in both cultures were asked to rate the strength of the emotional reactions they would feel if they failed to complete the task by the time they predicted; later, we asked the tardy members of those groups (the majority) how they actually felt. Members of both cultural groups significantly overestimated how badly they would feel about their prediction failure, consistent with a general intensity bias in affective forecasting (Buehler & McFarland, in press).

A key aspect of the intensity bias in affective prediction, like the planning fallacy itself, is that people know about but do not learn from the past. Why should students with poor grades overpredict how bad they will feel, and why should students with good grades overpredict how good they will feel – given that people have experienced the same type of outcomes many times in the past? More poignantly, why should people expect winter holidays to produce warmer feelings than they actually do, given a rich past diet of disappointment (Buehler & MacFarland, in press)? Why should both our Japanese and Canadian students overestimate how sad and guilty they will feel if they fail to complete a task by a predicted date (Buehler et al., 2000)? If one reason is that people neglect the lessons of the past when making case-specific predictions in the “inside mode,” then encouraging them to think more distributionally may reduce the extremity bias in hedonic prediction. Consistent with this reasoning, Buehler and McFarland (in press) found that the intensity bias was reduced in participants who were encouraged to think about a set of relevant past experiences.

The “inside–outside perspective” account of the planning fallacy implies that people focus on the task at hand and neglect background factors. People may also exaggerate the emotional impact of a future event (such as failing to complete a task when promised) because their predictions are focused on that event alone and fail to account for the mitigating effects of the many other events, internal and external, that affect a person at any given time. Wilson et al. (2000) term this a focalism bias, and Schkade and Kahneman (1998) describe a similar phenomenon they call a focusing illusion. Focalism may play a dual role in the planning fallacy: First, the focus on a single case may prevent recognition of the time taken up by rival activities; second, the neglect of the background distractions created by competing activities may lead the predictor to exaggerate both the pleasure that results from completing the task on time and the pain that results from missing the predicted date. When background factors are made salient by asking participants to consider the entire field of events that might occur at a given point in their lives, affective forecasts become less extreme and thus more realistic (Wilson et al., 2000). Such manipulations might similarly improve task-completion forecasts. It is not clear whether the manipulation truly induces an outside perspective, but it does broaden the planner’s field of view to include a wider range of background information and thus may weaken the reliance on specific plans for the task at hand.

Focalism may also affect predictions of task motivation, as well as predictions of emotional or hedonic response. The act of task prediction necessarily leads the planner to focus on one particular task among many, and this may lead the planner to exaggerate the ongoing importance of the target task. When people make plans to lose weight, improve their grades in college, or complete their income tax returns early this year, they may well overestimate how much their current motivation to complete the task will be present throughout the project period, and may also overestimate how salient their plans and intentions will remain throughout the project period.

This analysis suggests two ways to bring behavior in line with predictions: ensure that the original level of motivation is maintained throughout the project period and enhance the salience of goal-directed plans throughout the same period. Pham and Taylor (1999) instructed participants to imagine when, where, and how they would study for an upcoming exam in order to achieve a high grade, and carry out this exercise of imagination every morning during the week leading up to the exam. Naturally, most students are motivated to do well, plan to do well, and intend to act in ways that allow them to do well. Carrying out the imagination exercise served to link students’ motivations, plans, and intentions to their actual everyday behavior, and therefore led to improved exam performance. Similarly, in an investigation of the planning fallacy students were trained to envision themselves performing each of the steps needed to complete a school-related project by the time they had predicted (Taylor et al., 1998). Again participants rehearsed these simulations each day of the week during which they carried out the project. The exercise increased the proportion of students who finished by the predicted time from 14% (in a control condition) to 41%.

Gollwitzer (1999) has used similar manipulations at the time of prediction to make goal-related intentions and plans more mentally accessible in the relevant environments. He demonstrated that participants with a challenging goal (e.g., to finish a writing project during the winter vacation) were more likely to meet that goal if they were prompted to generate implementation intentions that specified precisely when, where, and how they would accomplish the goal. When the specified opportunity to act on their plans is encountered, people are more likely to initiate and carry out the actions required to meet their goals. Gollwitzer also reviewed a number of related studies in the area of health promotion showing that making vivid, detailed plans can increase the likelihood of self-protective behaviors such as breast self-examination (e.g., Orbell, Hodgkins, & Sheeran, 1997). Note that in such studies, the relative effect of an implemental focus on prediction versus behavior is not important; as long as the outcome is more likely to occur, the manipulation is successful. In the planning fallacy studies, in contrast, virtually all participants complete the task – the critical phenomenon is the match between predictions and completion times. As noted previously, our studies on school projects and holiday shopping found that requiring participants to generate detailed implementation intentions affected only predictions without bringing behavior into line. A challenge for future research is to identify factors that alter the impact of plans and mental simulations on behavior relative to their impact on prediction.