Buehler_planning Fallacy of Planning

8/3/2019 Buehler_planning Fallacy of Planning

1/16

Journal of Personality and Social Psychology1994, Vol. 67, No .3.366-381

Copyright 1994 by the American Psychological Association, Inc.0022-3514/94/S3.00

Exploring the "Planning Fallacy": Why People UnderestimateTheir Task Completion Times

Roger Buehler, Dale Griffin, and Michael Ross

This study tested 3 main hypotheses concerning people's predictions of task completion times: (a)People underestimate their own but not others' completion times, (b) people focus on plan-basedscenarios rather than on relevant past experienceswhile generating their predictions, and(c) people'sattributions d iminish the relevance of past experiences. Results supported each hypothesis. Ss' pre-dictions of their completion times were too optimistic for a variety of academic and nonacademictasks. Think-aloud procedu res revealed thatSs focused primarily on future scenarios when predict-ing their completion times. In Study4, the optimisticbias was eliminated forSs instructedto connectrelevant past experiences with their predictions. In Studies3 and 4, Ss attributed the ir past predictionfailures to relatively external, transient, and specific factors. In Study 5, observer Ss overestimatedothers' completion times and m ade greater use of relevant past experiences.

In 1871, the colony of British Colum bia agreed to join thenew country of Canadaon the condition thata transcontinentalrailway reach the west coast by 1881. In fact, because of theintervention of an economic depression and political changes,the last spike was not driven until 1885, 4 years after the pre-dicted date ofcompletion.Nearly 100 years later, in 1969, themayor of Montreal proudly announced th at the 1976 Olympicswould feature a state-of-the-art coliseum covered by the firstretractable roof ever built on a stadium. According to mayorJean Drapeau, the entire Olympic venture would cost$ 120 mil-lion and "can no more have a deficit than a man can have ababy" (Colombo, 1987, p. 269). Because of economic prob-lems, strikes, and other construction delays, the stadium roofwas not in place until 1989, 13 years after the predicted dateof completionand cost $120 million byitself! Many peopleconsider the Sydney Opera House to be the champion of allplanning disasters. According to original estimates in 1957, theopera house would be completed early in 1963 for$7 million.Ascaled-down version ofthe opera house finally opened in 1973at a cost of $102 million (Hall, 1980).

The history of grand construction projects is rife w ith opti-

Roger Buehler, Department of Psychology, Simon Fraser University,Burnaby, British Columbia, Canada; Dale Griffin and Michael Ross,Departm ent of Psychology, University of Waterloo, Waterloo, Ontario,Canada.

This article is based on a doctoral dissertation submitted by RogerBuehler to the University of W aterloo. The researchwas supported by adoctoral fellowship to Roger Buehler and a research grant to each au -thor from the Social Sciences and Humanities Research Council ofCanada.

We thank the following individuals who commented on earlier ver-sions: Robyn Dawes, Cathy McFarland, Sandra M urray, Chris Roney,Ewart Thomas, Jacquie Vorauer, Mark Zanna, and three anonymousreviewers.

Correspondence concerningthis article should be addressed to RogerBuehler, Department of Psychology, Simon Fraser University, Burnaby,British Columbia, Canada, V5A 1S6. Electronic mail may be sent [email protected].

mistic, even unrealistic, predictions (Hall, 1980). Vet currentplanners seem to be unaffected by this bleak history: The build-ers of the Channel tunnel connecting Britain and France pre-dicted that the first rains would run between London and Parisin June 1993, after an expenditure of 4.9 billion pounds. Thereal cost is expected to be at least10 billion pounds, and at thispoint the projected opening date is May 1994. The tendency tohold a confident belief that one's own project will proceed asplanned, even while knowing that the vast majority of similarprojects have run late, has been termed theplanning fallacy(Kahneman& Tversky, 1979).

Great construction projects are often undertaken by govern-ments. Proponents of these schemes may deliberately provideoverly optimistic assessments of cost and time to win politicalapproval for the projects. In addition, some of these projectsinvolve the adoption of new technologies that turn out to bemuch more complex and expensive than their advocates envi-sioned (Hall, 1980).

This phenomenon is not limited to commercial mega-proj-ects,however, and its occurrence does not depend on deliberatedeceit or untested technologies. From a psychological perspec-tive, the planning fallacy can perhapsbe studied most profitablyat the level of daily activities. Consider one familiar example:Academics who carry home a stuffed briefcase full ofwork onFridays, fully intending to complete every task, are often awarethat theyhave never gone beyondthe first one or two jobs on any

previous weekend. The intriguing aspect of this phenomenon isthe ability of people to holdtwo seemingly contradictory beliefs:Although aware that most of their previous predictions wereoverly optimistic, they believe that their current forecasts arerealistic. It seems that people can know the past and yet stillbe doomed to repeat it. The phenomenon, we propose, is notpeculiar to academics. Ina classroom surveywe conducted, stu-dents reported having finished about two thirds of their previ-ous projects(M = 68%) later than they expected.

The planning fallacy is an important topic of study for bothapplied and theoretical reasons. Inaccurate completion esti-mates can have economic, social, and personal costs. Surpris-

366


2/16

PREDICTION OF TASK COMPLETION TIME 367

ingly, however,we were able to locate little research on this phe-nomenon in the psychological literature. In one relevant study,professional engineers were foundto underestimate consistentlywhen equipment overhauls in electricity-generating stationswould be finished (Kidd, 1970). An important feature of thestudy distinguishes it from the present investigation, however:

The repair projects involved teams of technicians, and the engi-neers' predictions reflected group judgments, which are oftenmore extreme than those made by individuals (Janis,1982;My-ers & Lam m, 1976). Researchers have also examined how peo-ple make plans to accomplish tasks assigned to them in the lab-oratory, such as carrying out a sequence of errands in a hypo-thetical town (Hayes-Roth, 1981; Hayes-Roth & Hayes-Roth,1979). Subjects typically overestimate how much they can ac-complish in a given time period and continue to do so in theface of repeated negative feedback. The present study extendsthese previous investigations by documenting the prevalenceand magnitude of overly optimistic personal predictions forreal-world task s.

A second purpose of our studywas to explore the psychologi-cal mechanisms that underlie these forecasts.We focus, in par-ticular, on the mechanisms by which people segregate their gen-eral theories about their predictions (i.e., that they are usuallyunrealistic) from their specific expectations for an upcomingtask. Unlike the op timistic or self-enhancingbiases documentedby many researchers (e.g., Taylor & Brown, 1988; Weinstein,1980), the planning fallacy features the combination of rela-tively pessimistic general theories with optimistic specificjudgments.

Our research also differs in severalways from most studies onbiases in predictive judgment (for sum maries, see Dawes, 1988;Kahneman, Slovic,& Tversky, 1982; Rehm& Gadenne, 1990;Yates, 1990). The current experiments involve time estimates

rather than the occurrence or nonoccurrence of predictedevents and real tasks varying in familiarity instead of hypothet-ical and unfamiliar activities. We examine the everyday pro-cesses by which people maintain their optimistic outlooks inthe face of pessimistic past experience. We do not search forindividual differences in motivational constructs, such as opti-mism or self-enhancement, that might moderate the level ofop-timistic prediction; we seek out com mon processes that lead tooptimistic predictions in m ost people.

Processes Underlying Task Prediction

In their theoretical analysis of the planning fallacy, Kahne-man and Tversky (1979) suggested that people can use singular

and distributional information when predicting task comple-tion. Singular information relates to aspects of the specifictarget task that m ight lead to longeror shorter completion times.Distributional information concerns how long it took to com-plete other, similar tasks. In the present studies, where individ-uals make predictions about everyday activities, the distribu-tional information could either be their own past experiences(personal base rates) or the experiences of others (populationbase rates). Kahneman and Tversky (1979) suggested tha t peo-ple who focus on case-based or singular information adopt aninternal perspective: They concentrate on working outhow theywill complete the target task. In contrast, people who primarily

consider distributional information embrace an external per-spective: They compare the present task with past projects.Thus,the two general approaches to prediction differ primarilyin whether individuals treat the target task as a unique case oras an instance of an ensemble of similar problems.

In most cases, people should derive their predictions from

both case-based and distributional information. However, theexistence of the planning fallacy implies that people typicallyadopt an internal perspective when predicting their own com-pletion times; they seemingly fail to consider such relevant dis-tributional information as their previous experiences with sim-ilar tasks.

Obstacles to Using Past Experiences

When individuals make their time estimates, they may focuson the problem at hand, constructing a story about how theywill complete the task. A number of theorists have offered re-latedviews of the prediction process, emphasizing people's ten-dency to construct scenarios or narratives as they generate in-ferences and forecasts (Dawes,1988;Griffin, Dunning,& Ross,1990; Johnson & Sherman, 1990; Jungermann & Thuring,1987; Kahneman & Lovallo, 1991; Kahneman & Tversky,1982a; Klayman & Schoemaker, 1993; Read, 1987; Zukier,1986). Zukier suggested that for many judgments and predic-tions, people adopt a "narrative mode" of thinking concernedwith sequential relationships among events, action-relatedstructuring, and the integration of available information into aconnected narrative. Once individuals are in the planning ornarrative mode, there are a number of obstacles that preventthem from incorporating their past experiences into their story.We consider three particular impedim ents: (a) the forward na-ture ofprediction,(b) the elusive definition of "sim ilar" expe-riences, and (c) attributional processes that diminish the rele-vance of the past to the present.

The act of prediction, by its very nature, elicits a focus on thefuture rather than on the past; a future orientation may preventindividuals from looking backward in time. However, a failureto use personal base rates need not always result from neglect ofthe past. People may sometimes attend to their past experiencesbut nevertheless fail to incorporate this information into theirpredictions. The connection between past experiences and aspecific prediction task is not straightforward. The person mustfirst select an appropriate standard for comparison, a past expe-rience or class of experiences similar in important ways to theone under consideration. Often it may be difficult to detect theapprop riate set of past experiences; the various instances seem

so different from each other that individuals cannot comparethem meaningfully (Kahneman& Tversky, 1979).Even if people are able to identify a distribution of similar

experiences, they may not apply this information to the curren tprediction. Considerable research suggests that people tend toneglect background data (e.g., base rates) when they possesscase-based information on which to form their judgments (forreviews see Bar-Hillel, 1983; Kahneman et al., 1982). Peopleappear to make use of base-rate information only if they canconnect it to the judgment at hand (Bar-Hillel, 1980; Borgida&Brekke, 1981), such as when they incorporate base rates intotheir intuitive theories of causation (Ajzen, 1977). Thus, people


3/16

368 R. BUEHLER, D. GRIFFIN, AND M. ROSS

may use their previous experiences as a basis for prediction pri-marily when they can draw a causal connection between thepast and the present (e.g., the earlier tasks took longer than Iexpected becauseI tend to procrastinate).

Furthermore, people might actively process informationabout the past in a manner that reduces its pertinence to thecurrent prediction. The meaning and relevance of any past be-havior depends largely on an individual's explanation of why itoccurred (Jones & Davis, 1965; Kelley, 1967; Weiner, 1985).Certain types of attributionswill have the effect of linkinga pastevent to the present and future; other attribu tions will serve toisolate the past. To the extent that people perceive a previousepisode to be caused by external, unstable, and specific factors,they need not connect its outcome to future occasions. For ex-ample, an optimistic academic may attribute her or his inabilityto complete past weekend tasks to visits by her or his in-laws.Thus, the academic may generalize the previous failures onlyto weekends when that external and specific factor is present.Knowing that the in-laws are away this weekend, the academicmay suppose that she or he can readily attain her or his

objectives.We suggest that people often make attribution s that diminish

the relevance of past experiences to their current task. Peopleare probably most inclined to deny the significance of their per-sonal history when they dislike its apparent implications (e.g.,that a project will take longer than they hope). If they are re-minded ofa past episode that could challenge their optimisticplans, they may invoke attributions that render the experienceuninformative for the present forecast. This analysis is consis-tent with the view that individuals are inclined to explain awaynegative personal outcomes (for reviews see Snyder & Higgins,1988;Taylor& Brown, 1988).

Actor-Observer Differences in PredictionThere are contexts, however, in which people do use distribu-

tional information in the service of their predictions. In partic -ular, individuals may rely on distributional information whenthey make predictions forothers rather than for themselves. An-ecdotically, it seems that the planning fallacy vanishes when in-dividuals forecast other people's task completions. We are no tsurprised when our colleagues' journal reviews are late or whentheir house renovations take twice the time th at they predicted.Even without the benefit of hindsight, we would have antici-pated these outcomes.

An actor-observer difference in prediction may be partly dueto differing attributions for past behaviors. Because they are not

explaining their own inability to meet predictions, observersmay be less motivated than actors to discount previous predic-tion failures by attributing them to external, transitory, or un-stable causes. Also, observers tend to ascribe acto rs' behaviorsto characteristics of those individuals; conversely, the actors,themselves, are disposed to attribute their behaviors to externalcircumstances (Jones & Nisbett, 1972). When exp laining fail-ures to m eet previous predictions, observers may see actors asprocrastinators or as dilatory, but actors see themselves as thevictims of circumstances. Thereis another possible basis forac-tor-observer differences in predictions. In com parison with ac-tors, observers may be relatively unaware of the actors' future

activities and commitments. Consequently, it may be difficultfor observers to develop plans of how and when another individ-ual will complete a task. If observers cannot construct futurescenarios with confidence, they may rely on available sourcesof distributional information, including the other individual'sprevious performance.

Deadlines

We examined one additional factor that may influence peo-ple's predicted and actual completion times. Anecdotically, itappears that although people fail to m eet their predictions, theydo typically meet important deadlines. As teachers, we noticethat most students turn their assignments in on time, but fewsubmit them early. In our classroom survey, students indicatedthat they finish approximately three quarters of their projects(M = 73%) on the same day as the deadline. We suspect thatdeadlines may sometimes exert a greater impact on behaviorthan on predictions. Although people are aware that they havecompleted previous tasks only shortly before a deadline, theyremain optimistic that they will finish the current assignmentwith plenty of time to spare. Actors may fail to apply their pastexperiences with deadlines to their current predictions for thesame reasons that they generally fail to base their forecasts ontheir past experiences.

In addition, we anticipated an actor-observer difference inthe impact of deadlines on predictions. Relative to observers,actors are more motivated to deny the relevance of episodes thathave unpleasant implications for the present or futureas well asmore able to generate scenarios ofhow they would like the fu-ture to unfold. Therefore, we expect actors to depend less ondeadlines when generating their predictions than wouldobservers.

Overview of the Presen t Studies

We tested implications of the above analysis in a series of fivestudies. In the first study, we examined whether university stu-dents' time estimates for an important academic task wereoverly optimistic. In Study 2, we assessed the accuracy of par-ticipants' time estimates for a variety of academic and nonaca-demic tasks that they intended to complete in the followingweek. In the third study, we recorded the on-line narratives ofparticipants as they predicted their completion times for vari-ous academic tasks, some of which had deadlines.We analyzedthese narratives for evidence that people focus on plan-based

scenarios for the task at hand rather than on distributional in-formation, such as their previous experiences. In Study 4, wemanipulated the focus of the predictors' thoughts and the im-mediacy of their deadlines to test experimentally the hypothesesdeveloped in o ur na rrative analyses. In Studies3 and 4,we alsoexamined participants' explanations for past performances, an-ticipating that their attributions would diminish the relevanceof past failures to meet optimistic goals. In the final study, weaskedobservers to predict when another individual would finisha targettask. Each observerwas yoked to a participant in Study4, and we manipulatedthe type of information the observerpos-sessed about the actor (e.g., distributional or case-based infor-


4/16


mation). This experiment allowed us to test our suppositionsconcerning actor-observer differences in p rediction.

Study 1

Our initial study was designed to provide evidence that peo-ple's task com pletion estimates tend to be optimistically biased.To provide a relatively stringent test of the hypothesis, we se-lected individuals engaged in a project of considerable impor-tance, assessed their predictions when the projects were nearcompletion, and obtained an objective measure of completiontimes.

Method

Subjects and procedure. Thirty-seven psychology students (27women and10 men) enrolled in the final semester of the Honors Thesiscourse at the University of W aterloo, Waterloo, Ontario, Canada, werecontacted for a brief telephone survey conce rning their ongoing thesis.The interviewer asked respondents to predict as accurately as possiblewhen they would submit their finished thesis. In addition, the in-

terviewer asked respondents to forecast when they would com plete thethesis "if everything wentas well as it possibly co uld" and "if everythingwent as poorly as it possibly could." The order of the optimistic andpessimistic predictions was counterbalanced across participants. Thecoordinator of the thesis course recorded the date on which each thesiswas submitted.

Results and Discussion

Predicted and actual completion times were recorded as thenumber ofdays from the date of the survey. T heresults are sum-marized in Table 1. When asked for their best estimate, respon-dents predicted, on average, that they would finish n 33.9 days,but they actually took 55.5 days, /(32) = 3.43,p < .002.' An

additional 4 respondents were not included in this analysis be-cause they had not completed their thesis when our recordswere discontinued two semesters after the survey. Fewer thanone third of the respondents (29.7%) finished in the time theyreportedas their most accurate prediction.

Despite the optimistic bias, respondents' best estimates wereby no means devoid of information: T he predicted com pletiontimes were highly correlated with actual com pletion times (r =.77, p < .001). Compared with others in the sample, respon-

Table 1Predicted an d Actual Completion Times byPredictionInstruction: Study1

Measure

Predicted daysActual daysDifferenceAbsolute differenceSubjects com pleted in predicted

time (%)R: Predicted and actual days

Prediction instruction

Best

33.955.5

-21.622.6

29.7.77

Optimistic

27.455.5

-28.128.2

10.8.73

Pessimistic

48.655.5-6 .923.2

48.7.72

dents who predicted that they would take more time to finishactually did take more time. Predictions can be informativeeven in the presence of a marked prediction bias.

We also examined respondents' optimistic and pessimisticpredictions. When they assumed tha t "everything went as wellas it possibly could," students offered predictions that were al-most30 days earlier than the actual completion times(M = 21 Adays vs. 55.5 days), f(32) =4.11, p < .001; only 10.8% of therespondents finished their theses by the optimistic date. Inter-estingly, fewer than half ofthe respondents (48.7%) finished bythe time they had predicted assuming that "everything went aspoorly as it possibly could." Although the difference was notsignificant, respondents tended to underestimate their actualcompletion times even when they m ade pessimistic p redictions(M = 48.6 days vs. 55.5 days),t(32) = 1.03, ns . Respondents'optimistic and pessimistic predictionswere both strongly corre-lated with their actual completion times (rs = .73 and .72, re-spectively;ps < .01).

In terms of absolute accuracy, respondents' pessimistic pre-dictions fared no better than their best estimates. The absolutedifference between p redicted and actual completion times wasequivalent whether respondents were instructed to make pessi-mistic predictions or accurate predictions (Ms = 23.2 days and22.6 days, respectively),t(32) < 1} Similarly, respondents' pes-simistic predictions were no more informative than their bestestimates at the correlational level (see Table 1). Although theinstructions to make a pessimistic prediction decreased the op-timistic bias in prediction, it did not increase the accuracy ofrespondents' forecasts.

Study 2

One aim of the second study was to replicate the pattern ofresults obtained in the initial study. We also instituted severalchanges to the procedure to address some alternative inter-pretations of the optimistic bias revealed in Study 1. One possi-bility was that the optimistic bias was due to some atypical as-pects of the target task. The Honors Thesis course, althoughvery similar to many school projects, was novel because of its

Note. Means are based on 33 subjects.

1 In each of the studies reported, there were no significant sex differ-ences in the relation between predicted and actual completion times.The analyses are collapsed across sex.

2 Note that if the d istributions of both sets of difference scores (bestguesses minus actual completion times and worst-case predictions mi-nus actual completion times) were symmetrical around their means(e.g., in perfectly normal or rectangular distributions) then the mean

of the absolute difference scores would necessarily be smaller for thepredictions that had the smaller signed difference. However, the distri-bution of difference scores from the best-guess predictions were m ark-edly skewed, with a long tail on the optimistic side ofzero,a c luster ofscores within 5 or10 days ofzero,and virtually no scores on the pessi-mistic side ofzero. In contrast, the differences from the worst-case pre-dictionswere noticeably more symmetric aroundzero,with the num berof markedly pessimistic predictions balancing the num ber of extremelyoptimistic predictions. These differences in shape imply that the devia-tion of pessimistic predictions from actual completion times shouldhave greater variance than the deviation of best-guess predictions fromactual completion times. Although this pattern was observed, thedifference in varianceswas far from significant.


5/16

370 R. BUEHLER, D. GRIFFIN, AND M. ROSS

magnitude and lack of external deadlines. In Study 2, partici-pants made predictions about familiar everyday activities, in-cluding school assignments and tasks around the home. Antici-pating that there would be external deadlines for many of theacademic assignments, we sought to examine the relationsamong deadlines, predictions, and actual performance.

A second possibility was that participan ts in Study1 offeredoptimistic completion times out of a desire to present them-selves in a positive light. In the present study, we assessed thepossibility ofa self-presentationbias by varyingthe informationparticipants received about the purpose of their predictions. Ifparticipants are fully aware that the researcher will assess thevalidity of their p redictions, then experimen tal dem ands shouldbe for accurate rather than optim istic predictions. Thus,we ex-plicitly informed some participants that researchers wouldcompare subjects' predicted and actual completion times. Theremaining participants were simply told that researchers wereinterested in the activities of university students. As in Study 1,these latter subjects were not informed that the researcherswould evaluate the accuracy of the predictions. If the optimistic

bias in Study 1 was due to self-presentation , then subjectsshould exhibit less optimism when they are aware that the ac-curacy of their predictions will be assessed.

A related possibilitywas that, despite instructions to the con-trary, participants in Study1 merely reported a time by whichthey hoped to be done. Participants may have suspected thatthey were unlikely to finish by the time of their best estimates.Note that one finding suggests otherwise: Even when partici-pants were instructed to furnish their most pessimistic judg-ments, many continued to underestimate their completiontimes. In Study 2, a measure of participants' subjective confi-dence in their predictions provided a more direct assessment oftheir beliefs regarding the accuracy of their forecasts.

The present study differed from Study1 in another importantway. In the first study,we obtained an objective measure of com-pletion time; in the second study, we examined participants'forecasts for everyday tasks and thus had to rely on self-reportedcompletion times. We took a number of precautions to mini-mize the likelihood th at subjects would rep ort erroneous com-pletiontimes:We obtained participants' completion times aftera relatively short interval and emphasized their confidentialityas well as the importance of accurate recall. Although these pro-cedures cannot guarantee accurate reporting, we believed thatthe increase in external validity achieved by sampling a widerange of familiar activities outweighed the risks.

Method

Subjects and procedure. Subjects were 104 undergradua te psychol-ogy students (54 men and 50 women) who received course credit fortheir participation. Recruited by telephone for a study of the activitiesof university students, subjects participated individually or in smallgroups in two questionnaire sessions scheduled 1 week apart. At theinitial session, the experimenter provided verbal instructions tha t werevaried to create two information conditions. Subjects assigned to thecomplete information condition were informed that the primary pur-pose of the study was to assess the accuracy of people's predictions.Theywere told that they would predict completion times for projects inthe first session and repo rt their actual com pletion times in the secondsession. Subjectsin the partial information condition were told only that

the researcherswere studyingthe activities that university students wereengaged in throughout the year.

Following the verbal instructions, subjects received a questionnairecontaining the measures of prediction and confidence. Subjects wereasked to describe two tasks or projects that they intended to completein the next week, one that was school related and one that was not.The order of the two project descriptions was counterbalanced across

subjects. Subjects offered their predictions and judgments for eachproj-ect immediately after they described it. They predicted the date andtime they would finish he project. They also indicatedhow certain theywere that they would finish by the predicted time on a percentage scaleranging from0% {not at all certain) to 100% (completely certain).

Subjects returned for a second session1 week later. They indicatedwhether they had finished the projects and, ifso, the date and time ofcompletion. Subjects also reported whether therewas an external dead-line for completing the academic project and when the deadline was.Almost half of the projects remained unfinished at the time of the sec-ond session, even though subjects were instructed to choose tasks thatthey intended to complete within the week. At the end of the semester,we telephoned subjects whose projects were incomplete at the secondsession; we asked these subjects whether they eventually finished thetasks, and ifso, when.

Results

Predicted versus reported completion times. Because wewere unable to contact 3 subjects for the follow-up interview,the analyses are based on 101 subjects. Of these subjects, 97were able to report an academic project that they intended tocomplete in the next week (e.g., essays, compu ter assignments,and laboratory reports) and 78 were able to report a nonaca-demic project (e.g., "fix my bicycle," "clean my apartment,"and "write a letter to my friend"). Their predicted and reportedcompletion times for the projects were rounded to the nearesthalf hour, converted to number of days from the initial session,and subjected to mixed model analyses of variance (ANOVAs).The within-subjects factor was predicted versus reported com-pletion time; the between-subjects factors were the information(partialvs. complete) subjects received and the order (academicfirst vs. nonacademic first) in which they reported th e projects.The ANOVAs were based on the 91 academic projects and 62nonacademic projects that were finished by the tim e of the fol-low-up interview. Means are presented in Table 2. For bothtypes of projects, subjects took longer to finish than they pre-

Table 2Predicted an d Actual Completion Times byTypeof Task: Study 2

Task

Measure

Predicted daysActual daysDifferenceAbsolute differenceSubjects com pleted in predicted time (%)

R: Predicted and actual days

Academic

5.810.7

-4.95.6

37.1.36

Nonacademic

5.09.2

-4 .25.8

42.5.4 8

Note. Means are based on the 91 academic projects and 62 nonaca-demic projects thatwere finished before the follow-up interview.


6/16


dieted: for academ ic projectsM = 10.7 vs. 5.8 days,F(\, 87) =20.00,p < .001;for nonacademic projectsM = 9.2 vs. 5.0 days,F( 1,58) = 6.65,p < .02. The predictionbias was not moderatedby the between-subjects factors. Subjects' predictions wereequivalently biased regardless of whether they had receivedcomplete information about the purpose ofthe study and had

answered questions about their academic or nonacademicproj-ect first (all interactionFs < 1). Subjects finished37.1% of theacademic projects and 42.5% of the nonacademic projectswithin the predicted time.3

Note, again, that the optimistic bias in prediction does notindicate that subjects' predictions were unrelated to their re-ported completion times. Predicted and reported completiontimes were significantly correlated for both academic and non-academic projects (rs = .36 and .48,respectively;p < .01). Sub-jectswho predicted they would finish early, in com parison withthe other participants, reported that they had finished relativelyearly. However, the bias at the mean level does indica te tha tsub-jects' prediction errors were systematic. Subjects tended to un-derestimate rather than overestimate their completion times.

Confidence. On average, subjects reported feeling 74.1%certain that they would m eet their forecasts for academicproj-ects and 69.9% certain for nonacademic tasks. Subjects' cer-tainty ratings were related to whether they finished he projectswithin the predicted tim e. Point-biserial correlations indicatedthat subjects who reported higher certainty ratings were morelikely to fulfill their p redictions (r =.29, p < .01, for academ icprojects;r = .23,p < .06, for nonacadem ic projects).

The role of deadlines. A subset ofthe subjects ( = 62) re-ported having external deadlines for their academic projects; amajority ofthese subjects (80.6%) finished he projects in timeto meet their deadlines. On average, the projects were due in12.9, days and subjects reported finishing them in 11.0 days.Nevertheless, even subjects with deadlines typically exceededtheir predicted completion times. They predicted that theywould finish in only 5.9 days, on average, well in advance ofboth the deadline and the reported completion time; only38.7%of these subjects finished n the predicted time. Apparently, sub-jects underestimated the importance of deadlines in determin-ing when they would finish their projects. Correlational analy-ses support this interpretation: Although subjects' predictionswere only weakly associated with the deadlines (r = .23,p

Documents

Buehler_planning Fallacy of Planning