Speed/accuracy decisions in task performance: Built-in trade-off or separate strategic concerns? q

Organizational Behavior and Human Decision Processes 90 (2003) 148 164 ORGANIZATIONAL BEHAVIOR AND HUMAN DECISION PROCESSES www.elsevier.com/locate/obhdp Speed/accuracy decisions in task performance: Built-in trade-off or separate strategic concerns? q Jens F orster, a, * E.Tory Higgins, b and Amy Taylor Bianco c a International University Bremen and Universit at W urzburg b Columbia University c Teachers College, Columbia University Abstract In four studies we show that participantsõ regulatory focus influences speed/accuracy decisions in different tasks. According to regulatory focus theory (Higgins, 1997), promotion focus concerns with accomplishments and aspirations produce strategic eagerness whereas prevention focus concerns with safety and responsibilities produce strategic vigilance. Studies 1 3 show faster performance and less accuracy in simple drawing tasks for participants with a chronic or situationally induced promotion focus compared to participants with a prevention focus. These studies also show that as participants move closer to the goal of completing the task, speed increases and accuracy decreases for participants with a promotion focus, whereas speed decreases and accuracy increases for participants with a prevention focus. Study 4 basically replicates these results for situationally induced regulatory focus with a more complex proofreading task. The study found that a promotion focus led to faster proofreading compared to a prevention focus, whereas a prevention focus led to higher accuracy in finding more difficult errors than a promotion focus. Through speed and searching for easy errors, promotion focus participants maximized their proofreading performance. In all four studies, the speed effects were independent of the accuracy effects and vice versa. These results show that speed/accuracy (or quantity/quality) decisions are influenced by the strategic inclinations of participants varying in regulatory focus rather than by a built-in trade-off. Ó 2003 Elsevier Science (USA). All rights reserved. 1. Introduction One of the fundamental questions since the beginning of experimental psychology has been when and why people are fast or accurate (Woodworth, 1899). Across psychological areas, the so-called speed/accuracy tradeoff or quantity/quality conflict has been of major concern. In cognitive psychology, it has inspired theorizing about motor performances (Fitts, 1954; Fitts & Peterson, 1964; Howarth, Beggs, & Bowden, 1971; Keele, q This research was supported by a National Institute of Mental Health Grant, MH 39429, to E. Tory Higgins, by a Heisenbergfellowship from the Deutsche Forschungsgemeinschaft and by a Grant FO 244/6-1 from the Deutsche Forschungsgemeinschaft to Jens F orster. Jens F orster is now at the International University Bremen. * Corresponding author. Present address: International University Bremen, School of Humanities and Social Sciences, P.O. Box 750561, D-28725 Bremen, Germany. E-mail address: j.foerster@iu-bremen.de (J. F orster). 1968; Keele & Posner, 1968; Meyer, Abrams, Kornblum, & Wright, 1988; Meyer, Smith, & Wright, 1982; Woodworth, 1899; Zelaznik, Mone, McCabe, & Thaman, 1988). In developmental psychology it has, for example, been introduced as a diagnosticum for developmental coordination disorder (cf. Maruff, Wilson, Trebilcock, & Currie, 1999). In personality psychology, it has been used as an indicator for concentration and attention (cf. Brickenkamp, 1972), impulsivity and reflexivity (cf. Bush & Dweck, 1975; Dickman, 1985; Dickman & Meyer, 1988; Leung & Connolly, 1997; Salkind & Nelson, 1980), extraversion and neuroticisim (Malhotra, Malhotra, & Jerath, 1989; Socan & Bucik, 1998), anxiety (Revelle & Leon, 1985), intelligence (Phillips & Rabbitt, 1995; Tucker & Warr, 1996), general processes involved in achievement orientation (cf. Miller & Vernon, 1997), information processing (Dunn, Vaughan, Kreuzer, & Kurtzberg, 1999), and specific disabilities and problems (Chabot, Petros, & McCord, 1983; Raesaenen & Ahonen, 1995; Snowling, Hulme, & 0749-5978/03/$ - see front matter Ó 2003 Elsevier Science (USA). All rights reserved. doi:10.1016/s0749-5978(02)00509-5

J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 149 1 Even though the given examples might seem different in various aspects, we believe that there are fundamental principles that drive them. It is our attempt in this article to identify fundamental principles that possibly work across those different domains. Goulandris, 1994). Speed and accuracy have been investigated in the human resources area as well, with respect to supervisory monitoring (Brewer & Ridgway, 1998), selection of planning strategies (Josephs & Hahn, 1995), self-efficacy perceptions on sales performance (Lee & Gillen, 1989), computer menu structures (Seppaelae & Salvendy, 1985), the relationship between personality and faculty research productivity (Taylor, Locke, Lee, & Gist, 1984), and diverse leadership styles (Johnson, 1975). 1 Given the extensive interest in speed/accuracy decisions, it is surprising that the basic processes underlying these decisions are still poorly understood. Why are some people fast and why some accurate? Are those differences due to personality variables, situational variables, or both? Are there circumstances where people are both fast and accurate, thus optimizing task performance? The psychological literature generally treats speed/accuracy decisions as involving a built-in trade-off, people either trade speed for accuracy or vice versa. However, in this paper we want to go beyond the notion of built-in trade-off. We want to propose a selfregulatory account of behavior in speed/accuracy tasks. From our perspective, people can have different self regulatory foci, either a promotion focus or a prevention focus, which involve strategic concerns that influence speed/accuracy decisions. Let us briefly describe regulatory focus theory (e.g., Higgins, 1998) from which the model is derived, and then delineate its implications for performance in speed/accuracy tasks. Regulatory focus theory distinguishes between two kinds of goal pursuit that vary in regulatory focus concerns: concerns with attainment of aspirations and accomplishments (promotion focus), and concerns with attainment of responsibilities and safety (prevention focus). These distinct regulatory concerns can be emphasized either chronically or momentarily. To illustrate, employer employee interaction can chronically emphasize goal pursuit with either promotion focus concerns or prevention focus concerns (see also Higgins, 1989). Employees in employer employee interactions that involve a promotion focus experience pleasure when employers, for example, reward an employee by praising her creativity and encouraging the employee to seek opportunities to engage in rewarding activities. The employee experiences pain, when employers, for example, stop praising or when they ignore her achievements. The pleasure or pain from these interactions are experienced as the presence or absence of positive outcomes, respectively. The employersõ messages are communicated in reference to a state of the employee that does or does not meet promotion concerns, either This is what I ideally like you to do or This is not what I ideally like you to do, respectively. The regulatory focus is one of promotion, a concern with advancement and accomplishment, hopes, and aspirations (ideals). Strategically, individuals with a promotion focus are eager to approach matches to a desired end-state (i.e., pursue all means of advancement). Employees in employer employee interactions that involve a prevention focus experience pleasure when employers, for example, train the employee to be alert to potential dangers or misbehaviors. The employees experience pain when employers, for example, yell at or punish the employee for being irresponsible or careless. Here, the pleasure and pain are experienced as the absence or presence of negative outcomes. The employersõ messages are communicated in reference to a state of the employee that does or does not meet some prevention concerns, either This is what I believe you ought to do or This is not what I believe you ought to do, respectively. The regulatory focus is one of prevention, a concern with protection and safety, duties, and responsibilities (oughts). Strategically, individuals with a prevention focus are vigilant to avoid mismatches to a desired end-state (i.e., careful to avoid mistakes). According to the theory, momentary situations as well as chronic leadership styles can also temporarily induce either a promotion focus or a prevention focus on goal attainment. For example, feedback messages or task instructions can communicate gain/non-gain information (promotion focus) or non-loss/loss information (prevention focus). A promotion focus and a prevention focus involve different motivational orientations. Whereas individuals in a promotion focus with their inclination to approach matches are eager to attain advancements and gains, individuals in a prevention focus with their inclination to avoid mismatches are vigilant to assure safety and non-losses. In signal detection terms (e.g., Green & Swets, 1966; Tanner & Swets, 1954), individuals in a state of eagerness from a promotion focus are motivated to ensure hits and ensure against errors of omission (i.e., a lack of accomplishment). In contrast, individuals in a state of vigilance from a prevention focus are motivated to ensure correct rejections and ensure against errors of commission (i.e., making a mistake). These regulatory differences have been shown to influence performance in signal detection tasks. In recognition memory tasks for example, individuals in a promotion focus want to ensure recognizing a true target (i.e., want many hits ) and ensure against omitting a true target (i.e., want few misses or errors of omission), thereby producing an overall inclination to

150 J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 say Yes (a risky bias). Individuals in a prevention focus want to ensure rejecting a false distractor (i.e., want many correct rejections ) and ensure against failing to reject a false distractor (i.e., want few false alarms or errors of commission), thereby producing an overall inclination to say No (a conservative bias). A study by Crowe and Higgins (1997) tested these predictions. Participants were told that they would first perform a recognition memory task and then would be assigned a second, final task. A liked and disliked activities have been selected earlier for each participant to serve as the final task. The participants were told that which of the alternative final tasks they would work on at the end of the session depended on their performance on the initial recognition task. The relation between the initial memory task and the final task was described as contingent for everyone, but the framing varied as a function of both regulatory focus (promotion versus prevention) and valence (self-regulation) succeeding (pleasure) versus self-regulation failing (pain). Valence was included to test whether regulatory focus influences decision-making independent of participantsõ imagining pleasant versus painful outcomes (i.e., independent of regulatory anticipation, see Higgins, 1997). More specifically, the contingency framing was as follows: (a) Promotion Succeeding If you do well on the word recognition memory task, you will get to do the [liked task] instead of the other task; (b) Promotion Failing If you donõt do well on the recognition memory task, you wonõt get to do the [liked] task but will have to do another task instead; (c) Prevention Succeeding As long as you donõt do poorly on the word recognition memory task, you wonõt have to do the [disliked task] but will have to do the other task instead; and (d) Prevention Failing If you do poorly on the word recognition memory task, you will have to do the [disliked task] instead of the other task. The study found, as predicted, that participants in the promotion focus condition had a risky bias of saying Yes in the recognition memory task, whereas participants in the prevention focus condition had a conservative bias of saying No. Moreover, these regulatory focus effects were independent of the valence of the framing (i.e., success versus failure framing), which itself had no significant effects. Using the same paradigm, Crowe and Higgins (1997) found in a second study that when individuals work on a task where generating any number of alternatives is correct, those in a promotion focus generate more distinct alternatives (insuring hits) whereas those in a prevention focus are more repetitive (insuring against errors of commission). The results of additional studies provide substantial evidence for eagerness motivation in a promotion focus versus vigilance motivation in a prevention focus (F orster, Higgins, & Idson, 1998; Liberman, Idson, Camacho, & Higgins, 1999; Liberman, Molden, Idson, & Higgins, 2001; Roney, Higgins, & Shah, 1995; Shah, Higgins, & Friedman, 1998). To sum up, a promotion focus emphasis on strategic eagerness should lead to a more risky processing style that is concerned with getting hits, whereas a prevention focus emphasis on strategic vigilance should lead to a more careful processing style concerned with avoiding mistakes. Thus, generally one would expect participants in a promotion focus to be faster at the expense of accuracy and participants in a prevention focus to be more accurate at the expense of speed. This difference is due to strategic concerns with gains and non-gains (promotion) versus losses and non-losses (prevention) rather than to a built-in trade-off. That is people do not necessarily experience a built-in conflict between being accurate or fast. Rather, when they have strong promotion focus concerns they naturally are eager for hits (speed) and when they have strong prevention focus concerns they naturally are vigilant against mistakes (accuracy). If this is true, more specific predictions can be derived for motivation in speed/accuracy tasks than is possible from the notion of a built-in trade-off. The first is that if strategic motivation increases the closer a person is to goal completion (the classic goal looms larger effect), than promotion focus eagerness and speed should increase the closer one is to goal completion, and prevention focus vigilance and accuracy should increase the closer one is to goal completion. Let us consider the background for this prediction. There is evidence that regulatory focus influences motivational strength as reflected in the goal looms larger effect (F orster et al., 1998). This effect refers to the fact that motivation increases as the distance to the goal decreases (see Brown, 1948; Hearst, 1960, 1962; Lewin, 1951; Losco & Epstein, 1977; Miller, 1944, 1959; Miller & Murray, 1955). The value of each successive step toward a goal increases as its contribution to final goal attainment increases (F orster et al., 1998; see also Brendl & Higgins, 1995). Each successive step reduces a higher proportion of the remaining discrepancy. If the goal is to solve each of 10 anagrams, for example, solving the first reduces 10% of the remaining discrepancy whereas solving the last reduces 100% of the remaining discrepancy. As the value of each successive step increases, the motivation to take the step and reach the goal increases, the goal looms larger effect. The strategic motivations, however, are different for promotion and prevention. As the goal looms larger, an increase in strategic approach motivation (increasing eagerness) should be more evident for people in a promotion than a prevention focus, whereas an increase in strategic avoidance motivation (increasing vigilance) should be more evident for people in a prevention than a promotion focus. To test these hypotheses, F orster et al. (1998; Studies 1 & 2) used arm pressure as an on-line measure of

J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 151 motivational strength. Arm flexion (in which the direction of force is toward the self) has been shown to be more associated with consumption or approach, whereas arm extension (in which the direction of force is away from the self) is more associated with rejection or avoidance (see Cacioppo, Priester, & Bernston, 1993; Chen & Bargh, 1999; F orster et al., 1998; F orster & Strack, 1997; F orster & Strack, 1998; Priester, Cacioppo, & Petty, 1996; Solarz, 1960). Each participant solved two sets of seven solvable anagrams. While solving one set, they pressed on the flat surface of a machine on the bottom of a table inducing arm flexion (i.e., approach), and while solving the other set they pressed the machine on top of the table inducing arm extension (i.e., avoidance). Promotion versus prevention focus was either a chronic individual difference (Study 1) or an experimental variable manipulated by framing (Study 2). Both studies found that the approach gradient was more positive for participants with a promotion than a prevention focus, and the avoidance gradient was more positive for participants with a prevention than a promotion focus. These effects were independent of participantsõ expectancies, and they were replicated in a third study that used persistence rather than arm pressure as the measure of motivational strength (see also F orster et al., in press). Accordingly, if strategic concerns drive the processes in performing speed/accuracy tasks, we would predict speed and accuracy to follow the same motivational goal looms larger principle. In the promotion focus motivational state of eagerness, speed in speed/accuracy tasks should increase the closer one is to a goal whereas accuracy should not. In the prevention focus motivational state of vigilance on the other hand, accuracy should increase the closer one is to goal completion whereas speed should not. This specific hypothesis is tested in Studies 1 to 3. Study 4 tests another specific prediction of our strategic concern perspective that is not derivable from the notion of a built-in trade-off. The question here is under which circumstances would people be both fast and accurate? Efficient self-regulation should overcome the speed/accuracy trade-off under certain conditions, so that greater speed need not sacrifice accuracy. There are tasks, for instance, where the goal itself is to be as accurate as much as possible. A common example of such a task is proofreading. The purpose of proofreading is to check a manuscript for accuracy. The goal is to make the manuscript as accurate as possible. The signal is the actual presence of an error and every error found is a hit. Individuals in a promotion focus should eagerly pursue as many hits as possible. They should be motivated to maximize the number of hits they find during the time they have to search. To do so, they need to be efficient. They cannot afford to waste time searching for difficult errors when they can attain more hits during the limited time they have by searching for easy errors. We would predict, then, that individuals in a promotion focus would find more errors in a given time period than individuals in a prevention focus, that is, speed would be greater for a promotion than a prevention focus, and this would be accomplished largely by promotion focus individuals being especially accurate in searching for easy errors. Thus, for individuals in a promotion focus, accuracy would not be sacrificed for speed with respect to easy errors. Rather, speed and accuracy in finding easy errors would be maximized. In contrast, individuals in a prevention focus are not concerned with maximization or efficiency. They are concerned with being vigilant against mistakes. They want to ensure against failing to correct errors. Because failing to correct difficult errors is more likely than failing to correct easy errors, individuals in a prevention focus should concentrate on being vigilant against difficult errors. For difficult errors, then, individuals in a prevention focus should be more accurate than individuals in a promotion focus. Correcting difficult errors takes more time, however. Thus, individuals in a prevention focus could be characterized as sacrificing overall speed for the sake of accuracy. Because participants in a prevention focus are vigilantly looking for difficult errors, they might miss at the same time easy errors. This does not mean that there is a built-in speed/accuracy trade-off, however, because no similar sacrifice would be evident for individuals in a promotion focus. Rather, both individuals in a promotion focus and individuals in a prevention focus would be simply functioning in terms of their specific strategic concerns. 2. Overview of the studies In Studies 1 to 3, participants were asked to draw a picture by connecting numbered dots within a given time. It was predicted that participants with a promotion focus would be faster but less accurate than participants in a prevention focus. To investigate the motivational nature of the quantity/quality decision, however, these differences were assumed to get larger the closer participants worked toward the goal of completing the task, thus reflecting a goal looms larger effect. More specifically, it was predicted that when working on a set of four pictures, promotion focus participants would be generally faster than prevention focus participants but especially as they moved closer to the goal of completing the task. Prevention focus participants on the other hand, would be generally more accurate than promotion focus participants but especially as they moved closer to the goal of task completion. In Study 4 a more complex task (i.e., proofreading) was used and

152 J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 the difficulty of errors within a text was manipulated. It was predicted that in a promotion focus participants would be faster than participants in a prevention focus. It was also predicted that in a promotion focus participants would find more easy errors than participants in a prevention focus but participants in a prevention focus would be more accurate in finding difficult errors than participants in a promotion focus. 3. Study 1 3.1. Participants Fifty undergraduates, 25 males and 25 females from the University of W urzburg were recruited for a battery containing several unrelated experiments. The battery lasted about 2 h and for participation either course credit or DM 25, (at the time approximately US $ 14,) were given. Participants were tested in groups from 1 to 3. They were separated by screens so that they could not see each other. There were no significant main effects or interactions involving gender in any of the analyses below. More specifically, gender was added as a variable in all the analyses below with no effect, therefore in this study, data were collapsed over gender. 3.1.1. Stimulus material Four pictures from a childrenõs drawingbook were taken that involved connecting consequentially numbered dots in order to draw each picture. When correctly completed, all pictures depicted cartoon animals whose names were printed on the bottom of the page (e.g., Nilpferd [hippopotamus]). This was done to control for higher-order cognitive mediation, such as ability to imagine the animal behind the dots. 3.1.2. Strength of regulatory focus The computer questionnaire is an idiographic measure that asks participants to list attributes describing different self-representations from their own standpoint (see Higgins, Shah, & Friedman, 1997; Shah & Higgins, 1997). The questionnaire first defines ideal self-guides and ought self-guides. The ideal self-guide is defined as the type of person participants would ideally like to be, the type of person they hope or aspire to be, whereas the ought self-guide is defined as the type of person participants believe they ought to be, the type of person they believe it is their duty or responsibility to be. The participants are then asked to provide attributes describing their ideal and ought selves as quickly and as accurately as they can. Additionally, they are told that the ideal and ought attributes should be different from one another. The participants were asked to provide three ideal self attributes and three ought self attributes in a seemingly random order. After each ideal attribute, they had to rate the extent to which they ideally would like to possess it as well as the extent to which they actually possessed it. After each ought attribute, they had to rate the extent to which they believed they ought to possess it as well as the extent to which they believed they actually possessed it. All rating scales went from 1 (slightly) to 4 (extremely). Thus, six measures were obtained for each individual participant: (a) the response time to produce an ideal self-guide attribute; (b) the response time to make an ideal self-guide extent rating; (c) the response time for an actual self extent rating in relation to an ideal selfguide; (d) the response time to produce an ought selfguide attribute; (e) the response time to make an ought self-guide extent rating; and (f) the response time to make an actual self extent rating in relation to an ought self-guide. These response times were transformed by a logarithmic transformation. An overall ideal response time for each participant was calculated by adding together the three response times related to each ideal self-guide attribute provided by the participant and averaging across the three ideal attributes. The faster the overall response time for the ideal selfguide, the stronger the ideal self-guide and its promotion focus (see Higgins et al., 1997; Shah & Higgins, 1997). Similarly, an overall ought response time for each participant was calculated. An overall response time for each participant was calculated by adding together the three response times related to each ought self-guide attribute provided by the participant and averaging across the three ought attributes. The faster the overall response time for the ought self-guide, the stronger the ought self-guide and its prevention focus (see Higgins et al., 1997; Shah & Higgins, 1997). This personality measure has been commonly used in research on regulatory focus and has proven to be a valid measure of the strength of individualõs chronic concerns with promotion (ideal accessibility) or prevention (ought accessibility) (e.g., F orster et al., 1998; F orster, Higgins, & Strack, 2000; Higgins et al., 1997; Shah & Higgins, 1997). 3.1.3. Procedure Upon arrival, the participants started with filling out the regulatory strength measure and then completed tasks unrelated to the present experiments for about 50 min. Then, they were asked to participate in research on the development of a measure of motoric speed and accuracy in kids. To validate the task we would need an adult control sample. Participants were instructed to connect the dots of four subsequent pictures as fast and as thoroughly as they could. They were asked not to miss a dot (reflecting accuracy) and to get as much of the drawing done as possible within a period of 30 s for each of the four pictures (reflecting

J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 153 speed). 2 Participants were repeatedly told that they should be both accurate and fast. Pretesting indicated that no one completed a picture within the time period. The experimenter started and stopped the time for each picture with a stopwatch. The pictures were always presented in the same order. Afterwards, other unrelated tasks were performed for about 1 h. At the end of the experiment participants were asked whether they thought that some tasks from the battery were related to each other. Nobody mentioned a relation between the strength measure and the connecting-the-dots task. Participants were then debriefed, thanked, and paid. To divide participants in terms of chronic regulatory focus, the mean reaction time for the ought strength measure was subtracted from the mean reaction time for the ideal strength measure, so that higher values indicate a more predominant promotion focus and lower values indicate a more predominant prevention focus. The experimental design then involved Regulatory Focus (Predominant Promotion; Predominant Prevention) [a between-participants variable based on this median split] and Picture Order (1st; 2nd; 3rd; 4th) as a withinparticipants variable. Thus our design made it possible to observe change over time within participants as a function of regulatory focus strength. 3.2. Results 3.2.1. The relation between speed and accuracy The number of the dots for each picture where a participant ended after the time period was added across the four pictures. This sumscore was the dependent 2 An anonymous reviewer asked whether our instructions implied whether speed or accuracy is the more appropriate criterion of performance. We were very careful to make sure that the participants understood that both speed and accuracy were important. They were asked not to miss a dot (reflecting accuracy) and to get as much of the drawing done as possible within a period of 30 s for each of the four pictures (reflecting speed). Participants were repeatedly told that they should be both accurate and fast. Given that these task instructions for all participants explicitly state that both speed and accuracy are important, we do not believe that the instructions implied that either speed or accuracy were more important. However in order to be sure, the same instructions were given to 50 University of W urzburg students. The students majored in different disciplines and were asked to help us evaluate some stimulus materials. For participation they got a chocolate bar. Participants were asked to read the instructions and decide whether they implied speed, accuracy or both of them at the same time. Forty-three (86%) out of 50 indicated that the instructions implied both speed and accuracy, 3 (6%) decided that speed was more pronounced and 4 (8%) that accuracy was more pronounced. Participants were also asked whether the instructions could be improved in order to make sure that other participants understand the instructions as implying both at he same time. Here, only 6 participants indicated some suggestions. Thus, in general we are sure that our participants thought that both was important: speed and accuracy. measure of speed. The numbers of dots for each picture that a participant missed (i.e., that were not connected) up to the number ended after the time period was added across the four pictures. This sumscore was the dependent measure of inaccuracy. These speed and accuracy measures were positively but non-significantly correlated with each other, r ¼ :14; p >:30, indicating that when participants were faster they tended to make more mistakes. 3.2.2. Speed The dependent measure of speed as described above was entered into a multiple regression analysis with ideal and ought strength as the independent variables. As predicted, ideal strength and speed were positively correlated, B ¼ 2:46; p <:0001; whereas ought strength and speed were negatively correlated, B ¼ 1:85; p <:001, reflecting the fact that, as predicted, the higher participantsõ ideal strength the faster they were, and the higher participantsõ ought strength the slower they were. These correlations remained significant when the total amount of mistakes as a measure of inaccuracy was also entered into the model. To examine whether speed increased in a promotion focus and decreased in a prevention focus, additional repeated measures ANOVAS with Picture Order (within participants) and Regulatory Focus (between participants) as factors were conducted. The results are depicted in Fig. 1. The analyses revealed a main effect of Regulatory Focus, F ð1; 48Þ ¼28:88, p ¼ :001, indicating that participants in a promotion focus were generally faster (M ¼ 47:1) than participants in a prevention focus (M ¼ 37:5), and a main effect of Picture Order, F ð3; 144Þ ¼ 4:32, p ¼ :01, indicating that in general participants became faster as they moved closer to task completion (M1st ¼ 38:3, M2nd ¼ 43:1, M3rd ¼ 42:5, M4th ¼ 45:2). These two effects were qualified by a significant two-way interaction between the factors, F ð3; 144Þ ¼9:43; p <:001, reflecting the fact that there was a clear increase in speed for participants with a Fig. 1. Study 1. Speed as a function of Regulatory Focus and Picture Order.

154 J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 predominant promotion focus, whereas speed for participants with a prevention focus actually decreased slightly. Contrast analyses revealed no significant difference for the first picture, F < 1, but a significant difference for the second, F ð1; 144Þ ¼4:34, p <:01, the third, F ð1; 144Þ ¼17:00, p <:001, and fourth picture, F ð1; 144Þ ¼54:00, p <:0001. To see whether accuracy mediated these speed effects, accuracy for the pictures was entered as a covariate. All effects remained significant. 3.2.3. Accuracy The dependent measure of inaccuracy (as described above) was entered in a multiple regression analyses with ideal and ought strength as the independent variables. Here, there was a positive correlation between ideal strength and mistakes, B ¼ :50; p <:10, and a negative correlation between ought strength and mistakes, B ¼ :68; p <:05, indicating that, as predicted, the higher participantsõ ideal strength was the more mistakes they made, and the higher was participantsõ ought strength was the fewer mistakes they made. When speed was entered into the model as a covariate independent variable, the relation between ought strength and fewer mistakes remained significant, whereas the significance of the relation between ideal strength and more mistakes decreased, B ¼ :51; p ¼ :17. To examine, whether accuracy was a function of regulatory focus and goal distance, a repeated measures ANOVA with Regulatory Focus (based on a median split as described above) and Picture Order was conducted for mean number of mistakes per picture (see Fig. 2). There was a main effect of Regulatory Focus, F ð1; 48Þ ¼14:07, p <:001, showing that in general participants with a promotion focus committed more errors (M ¼ 7:7) than participants with a prevention focus (M ¼ 4:0), and a main effect of Picture Order, F ð3; 144Þ ¼5:10, p ¼ :01, showing that errors increased as participants moved closer to task completion Fig. 2. Study 1. Mistakes (as a measure of accuracy) as a function of Regulatory Focus and Picture Order. (M1st ¼ 4:7, M2nd ¼ 4:6, M3rd ¼ 6:8, M4th ¼ 7:2). These two main effects were qualified by a significant two-way interaction between the factors, F ð3; 144Þ ¼ 6:72, p <:001, showing that it was for participants with a promotion focus that there was an increase in errors, whereas for participants with a prevention focus errors actually decreased slightly. Again, contrast analyses were conducted, revealing that the difference in accuracy between promotion and prevention focus was not significant for the second picture, F < 1, but was significant for the first picture, F ð1; 144Þ ¼2:96, p <:05, the third, F ð1; 144Þ ¼9:07, p <:001 and the fourth picture, F ð1; 144Þ ¼45:12, p <:0001. Including speed as covariate, the effects for the third and fourth pictures remained significant. 3 In sum, the results of Study 1 showed that promotion focus eagerness produces a decrease in accuracy and an increase in speed the closer one is to the goal of task completion, whereas prevention focus vigilance produces a slight increase in accuracy and decrease in speed. The two effects of greater speed for promotion than prevention and greater accuracy for prevention than promotion were independent of one another. In Study 2 we investigated whether a situationally induced regulatory focus would yield similar effects. 4. Study 2 4.1. Method Because Study 2 was similar to Study 1, only the differences between the two are described. 3 We also conducted multiple linear regression analyses to test more thoroughly the correlations among speed, accuracy and promotion (ideal) strength and prevention (ought) strength, with the relation between each pair of variables being tested while controlling for the remaining variables. In a first analysis, total speed was the dependent variable, and total amount of mistakes and ideal and ought strength were the continuous independent variables. This analysis revealed a strong positive correlation between speed and ideal strength (controlling for ought strength and accuracy), B ¼ 2:41, tð46þ ¼4:66, p <:001, and a strong negative correlation between speed and ought strength (controlling for ideal strength and accuracy), B ¼ 1:87, tð46þ ¼ 3:35, p <:01. There was no significant correlation between speed and accuracy (controlling for ideal and ought strength), t < 1. In a second analysis, the total amount of mistakes was the dependent variable, and total speed and ideal and ought strength were the continuous independent variables. This analysis revealed a positive but not significant correlation between number of mistakes and ideal strength (controlling four ought strength and speed), B ¼ 0:51 tð46þ ¼ 1:39, p <:18, and a significant negative correlation between number of mistakes and ought strength (controlling for ideal strength and speed) and accuracy, B ¼ :69, tð46þ ¼ 1:99, p ¼ :052. As above, there was no significant correlation between speed and accuracy (controlling for ideal and ought strength), t < 1. In sum, the regression analyses matched the overall pattern we found with the median split analyses and thus confirmed our hypotheses.

J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 155 4.1.1. Participants 38 undergraduate students majoring in different disciplines at the University of W urzburg participated in a similar battery as described earlier. According to the framing conditions that will be described below, they were recruited for DM 28 or DM 27. All of them were debriefed and received DM 28 at the end of the experiment.. Participants were tested individually. There were no significant main effects or interactions involving gender in any of the analyses below. 4.1.2. Procedure Participants were given the same pictures as described above in a fixed order. Instead of measuring participantsõ regulatory focus strength, regulatory focus was experimentally manipulated through framing. All participants were instructed to be both fast and accurate. More specifically, they were told that a speed/accuracy score would be computed that reflects both speed and accuracy. This score would vary from 0 to 100 where zero reflects both low speed and low accuracy and 100 reflects both perfect accuracy and high speed. Because pretests have shown that W urzburg undergraduates would usually attain speed/accuracy values of 70, they were asked to perform above this level. It was emphasized that they would not meet this criterion if they were only fast or if they were only accurate. The criterion could only be met by being both fast and accurate. These instructions were given to all participants. In the experimental conditions of promotion or prevention framing additional instructions were given. For the promotion framing group that was recruited for DM 27 participants were told that they would gain an extra DM if they performed above the 70 level (gain) and if they did not perform above the 70 level they would not gain an extra DM (non gain). Order of gain and non gain instructions was counterbalanced and order had no effects. The prevention framing group was recruited for 28 DM. They were told that they would loose one DM if they did not perform above the 70 level (loss), but if they did perform above the 70 level they would not lose one DM (non loss). Order of loss and non loss instructions was counterbalanced and had no effects. Thus, the experimental design was a 3 Regulatory Focus Framing (promotion; prevention; control) 4 Picture Order (1st Position 2nd Position; 3rd Position; 4th Position), with the first factor as a between-participants variable and the second factor as a within-participants variable. 4.2.2. Speed The mean number of the dot where participants ended for each condition is depicted in Fig. 3. These numbers were entered in a repeated measures ANOVA with Picture Order and Regulatory Focus Framing as the independent variables. As seen in Fig. 3, participants with promotion framing were faster (M ¼ 47:0) than participants in the control condition (M ¼ 41:7) followed by participants in the prevention framing condition (M ¼ 37:1), F ð2; 35Þ ¼ 14:32; p <:0001, for the predicted main effect of Regulatory Focus Framing. There was also a main effect of Picture Order, F ð3; 105Þ ¼32:05; p <:0001, indicating that overall from the first to the last picture, participants became faster (M1st ¼ 38:4; M2nd ¼ 35:1; M3rd ¼ 45:8; M4th ¼ 58:3). As predicted, these main effects were qualified by a significant two-way interaction, F ð6; 105Þ ¼3:47, p <:01, reflecting the fact that the increase in speed as participants moved closer to task completion was highest for participants in the promotion framing condition, whereas for the control group this increase was much lower and for the prevention group basically no increase was observed. When total amount of mistakes was included in the analysis as a covariate, the predicted interaction effect remained significant. Contrast analyses revealed that the difference between the prevention framing group and the control group was not significant for the first two pictures, F ð1; 105Þ ¼1:09 for the first, and F ð1; 105Þ ¼1:15, for the second, respectively, but was marginally significant for the third picture, F ð1; 105Þ ¼2:05, p <:10; and was significant for the fourth picture, F ð1; 105Þ ¼12:82, p <:001. The difference between the promotion framing group and the control group was not significant for the first, F ð1; 105Þ ¼1:12 or for the second picture, F ð1; 105Þ ¼1:45, but was significant for the third picture, F ð1; 105Þ ¼2:89, p <:05, and the fourth picture, F ð1; 105Þ ¼17:32, p <:0001. When accuracy was in- 4.2. Results 4.2.1. The relation between speed and accuracy There was a positive correlation between speed and inaccuracy (sum scores as described in Study 1), r ¼ :34; p <:05. That is, the faster were participants, the more mistakes they made. Fig. 3. Study 2. Speed as a function of Regulatory Focus Framing and Picture Order.

156 J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 cluded as a covariate, all significant effects remained significant. Fig. 4. Study 2. Mistakes (as a measure of accuracy) as a function of Regulatory Focus Framing and Picture Order. 4.2.3. Accuracy The mean number of mistakes per picture as a function of Picture Order and Regulatory Focus Framing for each condition is depicted in Fig. 4. A repeated measures ANOVA on these numbers revealed no significant main effect of Picture Order, F ð3; 105Þ ¼1:8; p >:15. As predicted, there was a main effect of Regulatory Focus Framing F ð2; 35Þ ¼4:92; p <:05, indicating that participants with promotion framing (M ¼ 32:0) made more errors than participants in the control group (M ¼ 22:3) or those with prevention framing (M ¼ 10:4). As predicted, there was also a significant two-way interaction, F ð6; 105Þ ¼2:62; p <:05, showing that whereas with a prevention framing the number of mistakes decreased as participants moved closer to task completion, with promotion framing the number of mistakes increased. For the no framing control condition, the number of mistakes was stable. Contrast analyses showed that the differences between promotion framing and the control group was significant for the third picture, F ð1; 105Þ ¼5:63; p <:01 and for the fourth picture, F ð1; 105Þ ¼10:11; p <:001, but not for the first or the second pictures, F s < 1:32. The difference between the prevention framing group and control group was significant for the third picture, F ð1; 105Þ ¼11:24; p <:001, and the fourth picture, F ð1; 105Þ ¼12:07; p <:001, but not for the second or first pictures, F s < 1. After entering speed as a covariate, all significant effects remained significant. In general, the results of Study 2 replicate those of Study 1 with situationally induced rather than chronic regulatory focus. Note that the incentive for all participants was the same and still regulatory focus produced differences in speed and accuracy. In addition, note that the control condition was always between promotion framing and prevention framing for both speed and accuracy, indicating that both prevention and promotion framing influence speed (controlling for accuracy) and accuracy (controlling for speed). One might wonder whether the results are confounded with order in the two studies above, because we did not counterbalance the order of the four tasks. This question has to do with there being a series of four pictures that permitted a goal looms larger effect to be examined in addition to the basic prediction that a promotion focus would strategically relate to speed concerns and prevention would strategically relate to accuracy concerns. It should be noted that this basic prediction was supported in both studies. Our hypothesis about the goal looms larger effect was over and above this basic prediction. It was not essential to the basic prediction of the studies. Thus, even if there were an order effect, it would not be a problem for the basic prediction. Nonetheless, it is important to note that the pattern of findings in these studies cannot be explained in terms of an order effect. All of the participants received the same order. Thus, any practice effect or difference in picture ease was the same for everyone. A simple order effect, therefore, cannot predict why, for the same order of the same pictures, promotion focus participants increased their speed performance whereas prevention focus participants did not, and prevention focus participants increased their accuracy performance and promotion focus participants did not, and these two interactions that go in opposite directions were statistically independent of one another. One might also wonder whether participants become tired over time and because of waning resources decided to be either fast or accurate. The results of the control condition in Study 2 do not support such an explanation because, if anything, the participants performed slightly better over time. In a third study, both of these questions were addressed in another way. If it is the psychological distance to the goal that drives motivation (as we hypothesize) rather than just the physical length of a task, then manipulating the psychological distance while keeping the physical distance constant should produce the same effects as found in Study 2. In Study 3, some participants were told that they had to work on four pictures whereas others were told that they had to work on eight pictures. All participants eventually worked on four pictures. For the former group who think they have to solve four pictures, the goal should loom larger sooner while moving through the four pictures than for participants who think they have to solve eight pictures for whom completion of the first four pictures would only be halfway through the task. We expected to replicate the findings of Study 2 for the former four picture group but not for the latter eight picture group. Because all participants have to work on the same actual number of pictures, fatigue from working through the four pictures could not account for this predicted pattern of results.

J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 157 Study 3, then, was explicitly designed to test for the goal looms larger contribution to our findings controlling for practice effects or differences in picture ease. Once again, everyone received the same four pictures in the same order. Half of the participants, as in Studies 1 and 2, were told there were four pictures. But the other half of the participants were told that there were eight pictures. Practice over the four pictures or effects of differences in picture ease are the same in both conditions. However, the goal looms larger effect as participants moved from one to four pictures would be experienced by those who thought there were four pictures but not by those who thought there were eight pictures. We predicted that there would be an interaction with the pictures in the former condition but not in the latter condition. 5. Study 3 5.1. Method Because Study 3 was basically a replication of Study 2, only the differences will be described. 5.1.1. Participants Sixty-eight undergraduate students majoring in different disciplines at the University of W urzburg participated in a battery containing several unrelated tasks as described above. Depending on the experimental conditions, participants were offered DM 28 or DM 27 and all of them received DM 28 at the end of the experiment. Participants were tested individually. There were no significant main effects or interactions involving gender in any of the analyses below. 5.1.2. Procedure The procedure was exactly the same as described in Study 2 except that only the experimental conditions of prevention and promotion framing were included. In addition, half of the participants were told that they had to work on four pictures whereas the other half were told that there would be eight pictures. The latter group was informed after the fourth task that due to time limits they would not have to do another four pictures. Thus, all participants actually worked on four pictures. Again, because all participants received the same pictures, the predicted interactions cannot be explained in terms of a simple order effect. Accordingly, the experimental design consisted of the following factors: 2 Regulatory Focus Framing (promotion; prevention), 2 Psychological Distance (expect four pictures; expect eight pictures) and 4 Picture Order (1st; 2nd; 3rd; 4th). The first two factors were each between-participants, whereas the last factor was within-participants. 5.1.3. The relation between speed and accuracy There was a positive but non significant correlation between speed and inaccuracy, r ¼ :18; p ¼ :15, indicating that the faster participants were, the more mistakes they tended to make. 5.1.4. Speed Fig. 5 depicts the mean number of the dot where participants ended for each condition. Except for the main effect of Psychological Distance, F < 1, all other effects were significant. First, there was a main effect of Regulatory Focus Framing, F ð1; 64Þ ¼11:9; p <:01, indicating that participants with a promotion focus were faster (M ¼ 41:9) than participants with a prevention focus (M ¼ 36:0). The main effect of Picture Order, F ð3; 192Þ ¼16:90; p <:0001, indicates that participants became faster as they moved closer to goal completion (M1 ¼ 33:7, M2 ¼ 37:9, M3 ¼ 39:1, M4 ¼ 45:9). The significant two-way interaction between Psychological Distance and Picture Order, F ð3; 192Þ ¼2:70; p <:05, reflects the fact that there was more of an increase in speed for participants who thought they had to do four pictures (M1 ¼ 33:4, M2 ¼ 34:9, M3 ¼ 41:4, M4 ¼ 47:1) than for participants who thought they had to do eight pictures (M1 ¼ 34:0, M2 ¼ 39:6, M3 ¼ 36:7, M4 ¼ 44:7). The significant two-way interaction between Regulatory Focus Framing and Psychological Distance, F ð1; 64Þ ¼9:39; p <:01, indicates that participants who thought they had to do four pictures were faster when they were in a promotion focus (M ¼ 44:8) than a prevention focus (M ¼ 33:6), whereas participantsõ speed did not substantially differ by focus when they thought they had to do eight pictures (Mpromotion ¼ 39:1; Mprevention ¼ 38:4). The significant two-way interaction between Regulatory Focus Framing and Picture Order, F ð3; 192Þ ¼12:78; p <:001, reflects the fact that there was an increase in speed for participants with a promotion focus (M1st ¼ 32:8, M2nd ¼ 37:6, M3rd ¼ 42:3, M4th ¼ 55:0) but not for participants with a prevention focus (M1st ¼ 34:5, M2nd ¼ 36:9, M3rd ¼ 35:9, M4th ¼ 36:7). Fig. 5. Study 3. Speed as a function of Regulatory Focus Framing, Psychological Distance, and Picture Order.

158 J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 Most important, these effects were qualified by a significant three-way interaction among all the factors, F ð3; 192Þ ¼13:98, p <:0001, reflecting the fact that, as predicted, the greater increase in speed for participants in a promotion than a prevention focus was stronger when they thought that they had to do four pictures than when they thought they had to do eight pictures. As Fig. 5 shows, all the other gradients were basically flat. Contrast analyses investigated when the psychological distance effect was significant within the promotion and prevention framing conditions. In the promotion framing condition, the greater speed of the four than the eight group was not significant in the first and the second picture, F s < 1:03, respectively, but the difference was significant for the third picture, F ð1; 192Þ ¼4:22; p <:01 and the fourth picture, F ð1; 192Þ ¼36:98; p <:001. In the prevention framing condition, the lesser speed of the four than the eight group was not significant for the first and third pictures, F s < 1 but was significant for the second picture, F ð1; 192Þ ¼2:90; p <:05 and the fourth picture, F ð1; 192Þ ¼22:00; p <:0001. Analyses with accuracy as a covariate revealed that both the promotion and prevention effects for the fourth picture (as well as the promotion effect for the third picture) remained significant. 5.1.5. Accuracy For the dependent measure of accuracy (see Fig. 6), the analyses revealed that except for the main effect of Psychological Distance, F < 1, all main effects and interactions were significant. The main effect of Regulatory Focus Framing, F ð1; 64Þ ¼9:96; p <:01 showed again that participants with a promotion focus (M ¼ 6:3) committed more mistakes than participants with a prevention focus (M ¼ 4:1). There was also an increase of mistakes as participants moved closer to goal completion (M1st ¼ 3:9, M2nd ¼ 4:5, M3rd ¼ 5:5, M4th ¼ 6:9), as reflected in a main effect of Picture Order, F ð3; 192Þ ¼7:1; p <:001. The significant twoway interaction between Psychological Distance and Picture Order, F ð3; 192Þ ¼2:88; p <:05 reflects the fact that mistakes especially increased when participants thought they had to solve four pictures (M1st ¼ 3:5, M2nd ¼ 4:4, M3rd ¼ 5:7, M4th ¼ 8:3) rather than eight pictures (M1st ¼ 4:3, M2nd ¼ 4:6, M3rd ¼ 5:2, M4th ¼ 5:4). The significant interaction between Regulatory Focus Framing and Psychological Distance, F ð1; 64Þ ¼7:77; p <:01 demonstrates that participants who thought they had to solve four pictures committed more mistakes when in a promotion focus (M ¼ 7:6) than a prevention focus (M ¼ 3:4), whereas when they thought they had to solve eight pictures this difference was slight (Mpromotion ¼ 5:0, Mprevention ¼ 4:8). Additionally, the interaction between Regulatory Focus Framing and Picture Order, F ð3; 192Þ ¼6:34; p <:001 reflects the fact that, as predicted, in a promotion focus the amount of mistakes increased with movement toward goal completion (M1st ¼ 4:1, M2nd ¼ 4:9, M3rd ¼ 6:5, M4th ¼ 9:7), whereas this was not the case for participants in a prevention focus (M1st ¼ 3:6, M2nd ¼ 4:2, M3rd ¼ 4:4, M4th ¼ 4:0). Most important, there was a significant three-way interaction among all the factors, F ð3; 192Þ ¼4:75; p <:01. This interaction reflects the fact that, as predicted, the only increase in mistakes with movement toward goal completion was observed for participants in a promotion focus who thought that they would solve four pictures. For all other conditions this gradient was almost flat. Again, separate contrast analyses were conducted within the promotion and prevention framing conditions to test for the psychological distance effect. For the promotion framing condition, the difference between the four and the eight groups was only significant for the fourth picture, F ð1; 192Þ ¼34:50; p <:0001 (for the first three pictures, all F s <:1:44). For the prevention framing condition there was a marginal effect for the fourth picture, F ð1; 192Þ ¼2:52, p <:10 (for the first three pictures all F s < 1:14). The psychological distance effect for the promotion framing condition on the fourth picture remained significant when speed was entered as a covariate. 6. Discussion of studies 1 3 Fig. 6. Study 3. Mistakes (as a measure of accuracy) as a function of Regulatory Focus Framing, Psychological Distance, and Picture Order. All three studies showed that participants with a promotion focus are faster but less accurate than participants in a prevention focus. The studies also showed that as participants moved closer to the goal of task completion, speed increased and accuracy decreased. For participants with a promotion focus whereas speed decreased and accuracy increased for participants with a prevention focus. The effects on speed and accuracy were independent from one another as revealed in covariance analyses. Study 3 also found that it is

J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 159 psychological distance rather than physical distance from the goal that drives these latter effect. In addition, the results of these three studies cannot be explained in terms of an order effect because all participants received the same order of tasks in each study. Because they all received the same order, the obtained interactions cannot be explained in terms of practice, fatigue or differential ease of different tasks. Study 3 specifically rules out such possibilities because we found the goal looms larger interactions only when participants thought they would be close to goal fulfillment (i.e., when they thought they had only four tasks to solve) but not when they thought they were only halfway through the task (i.e., when they thought they had eight tasks to solve but were interrupted after the fourth task). Those results cannot be explained in terms of simple order effects. 4 In Study 4, we aimed to replicate the main findings of Studies 1 3 with a more complex task. More specifically, participants were asked to proofread a text and it was predicted that individuals in a promotion focus would be eager to maximize the number of hits. Because they need to be efficient, they would be especially motivated to find easy errors. Thus, we predicted that individuals in a promotion focus would find more easy errors in a given time period than individuals in a prevention focus. Individuals in a prevention focus are not concerned with maximizing hits. They are concerned with being vigilant against mistakes. They want to ensure against failing to correct errors. Because failing to correct difficult errors is more likely than failing to correct easy errors, they should be especially vigilant against difficult errors. This would produce higher accuracy for difficult errors when individuals are in a prevention focus than a promotion focus. Because finding difficult errors takes more time, this higher accuracy for difficult errors should be at the expense of speed for individuals in a prevention focus. 7. Study 4 7.1. Method 7.1.1. Participants Twenty-nine Columbia undergraduate students participated, 15 in the promotion framing condition, and 14 4 In Studies 1 3, the same order of the pictures was fixed for all participants. This raises the possibility that the particular content of the picture in each order position might have contributed to the results. It could not account for the differences found between promotion and prevention, but it could contribute to the goal looms larger effects. Subsequent research by Seibt and F orster (2001) using the same pictures and similar participants as Studies 1 3 found that these pictures do not differ in difficulty, importance or fun. in the prevention framing condition 15 male students and 14 female students participated in the study. Depending on the framing condition (see below), they were recruited for $ 3 or $4. Participants were tested individually. There were no significant main effects or interactions involving gender in any of the analyses below. 7.1.2. Stimulus material: The proofreading task The proofreading task contained sentences from psychology textbooks consisting of a 31-line passage about how theories of attraction developed from attitude similarity research. The passage contained a total of 46 errors, broken down into four different types of errors that appeared in the passage in an intermixed order. In order to ascertain performance on a number of different types of proofreading measures relevant to situations in the real-world, the four different types of errors were developed based on two dimensions. One distinction between the types of errors was made based on a taxonomy of tasks developed by Revelle (1986)and Revelle, Anderson, and Humphreys (1987), and modified by Revelle (1987). Proofreading for non-contextual or surface errors was determined to require less shortterm memory and was therefore less difficult than proofreading for contextual errors (see also Weinstein, 1977). Proofreading for these two types of errors did not differ along RevelleÕs, 1987 dimension of long-term memory. A separate dimension, level of complexity, was employed in the current study. Type 1 simple surface errors (9 errors) such as the difference between peple and people were the least difficult of all four types of errors. The type 2 complex surface errors were next in difficulty level, and included errors that were similar to type 1 errors, but involved mainly misspellings of longer words, such as affliation versus affiliation (13 errors). More difficult yet were contextually inappropriate errors that also involved incorrect punctuation, such as from time to time we here reports. These were called type 3 simple contextual errors (16 errors). Finally, the most difficult errors were type 4 complex contextual errors (8 errors) that involved mistakes in subject verb agreement, such as Our attributions about the causes of other peopleõs behavior seems to be... To sum up, the proofreading task contained 46 errors. The effectiveness of our error difficulty manipulation is demonstrated by Study 4 finding significant main effects of error difficulty on performance (reported later in the Results section). 7.1.3. Procedure Upon arrival participants read the instructions for the proofreading task. They were told to circle the errors in the text passage they would be given, but not to actually correct them. The instructions repeatedly emphasized that they should work as quickly and as accurately as possible. They were told that an individual

160 J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 performance score would be computed based on whether they did well on both speed and accuracy. They were told these scores varied from 0 (extremely bad) to 100 (extremely good), and that usually Columbia students scored around 60. In the promotion framing condition the participants were told that they would be paid $3, but there was a possibility of gaining a dollar. If they scored above 60 in the speed/accuracy score, they would gain a dollar, but if they failed to score above 60 they would not gain a dollar. In the prevention framing condition they were told that they would be paid $4, but there was a possibility of losing a dollar. If they scored above 60, they would not lose a dollar, but if they failed to score above 60, they would lose a dollar. After reading the instructions, they were asked about their expectancies for their performance (What score do you expect to attain on this task?) on a 11-point rating scale (from 0 to 100). Afterwards, they were asked to turn to the next page to start the proofreading task. They were stopped by the experimenter after 4 min. Participants were debriefed, paid $4 and thanked by the experimenter. The experimental design was a 2 Regulatory Focus Framing (promotion; prevention) by 4 Error Difficulty (Type 1; Type 2; Type 3; Type 4), with the first factor as a between-participants variable and the second factor as a within-participants variable. 7.2. Results 7.2.1. Dependent measures Speed was defined as the number of errors found by a participant in the given time. Accuracy was defined as the percentage of errors found by a participant among existing errors for the lines completed when the participant stopped. 7.2.2. Expectancies Regulatory Focus Framing did not produce different levels of expectancies (Mpromotion ¼ 71:1%; Mprevention ¼ 68:0%), t < 1. 7.2.3. The relation between speed and accuracy Bivariate correlation analyses indicated a positive correlation between speed and accuracy, r ¼ :45; p <:02, reflecting the fact that participants who were faster found a higher percentage of errors while proofreading. 7.2.4. Speed (percentage of errors of each type found within 4 minutes) For each type of error a percentage was calculated (e.g., number of type 1 errors found within 4 min/number of type 1 errors overall). A 2 (Regulatory Focus Framing) 4 (Error Difficulty) ANOVA for mixed factorial designs was conducted with overall percentage Table 1 Speed (percentage of errors found within 4 min) as a function of difficulty and regulatory focus framing in Study 4 Errors Type 1 Type 2 Type 3 Type 4 Framing Promotion (n ¼ 15) 78.5% 72.8% 36.3% 26.7% Prevention (n ¼ 14) 57.9% 40.1% 36.2% 41.1% of errors found as the dependent measure (see Table 1), revealing significant main effects for Error Difficulty, F ð3; 81Þ ¼31:70, p <:0001, and for Regulatory Focus Framing, F ð1; 27Þ ¼4:23; p <:05. There was also the predicted significant two-way interaction between the two factors, F ð3; 81Þ ¼12:81; p <:0001. The Regulatory Focus Framing effects and the two-way interactions remained significant when accuracy was entered as a covariate into the analysis. In general, type 1 errors had the highest percentage circled (M ¼ 68:6), followed by type 2 errors (M ¼ 57:0), followed by type 3 errors (M ¼ 36:2), followed by type 4 errors (M ¼ 33:6), reflecting the fact that the participants were faster to find the simpler errors. More germane to the hypotheses, participants with a promotion focus were faster (M ¼ 53:2) than participants with a prevention focus (M ¼ 42:4). In addition, as Table 1 indicates, participants with a promotion focus found generally more easy errors (type 1 and type 2) than participants with a prevention focus, whereas participants with a prevention focus found more difficult errors (type 3 and type 4) than participants with a promotion focus. Contrast analyses were conducted to see whether the differences between promotion and prevention framing were significant for the different types of errors. There were statistically significant differences for type 1, F ð1; 81Þ ¼12:31, p <:001, type 2, F ð1; 81Þ ¼31:12, p <:0001, and type 4 errors, F ð1; 81Þ ¼6:03, p <:01, whereas for type 3 errors no differences were found, F < 1. When accuracy was entered as a covariate into the contrast analyses, the speed effects for the type 1 and type 2 errors remained significant. 5 7.2.5. Accuracy (number of errors found/number of errors present in lines completed) A 2 (Regulatory Focus Framing) 4 (Error Difficulty) ANOVA for mixed factorial designs was conducted with the percentage of errors found up to the line where they stopped. The means are presented in Table 2. The analysis revealed a main effect of Error Difficulty, 5 For these data, log transformations were also conducted that are appropriate for analyses on percentages. These new analyses found the same results as described for the raw percentages. Because the results were the same for both kinds of analyses, we report the raw percentages since they are more descriptive.

J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 161 Table 2 Accuracy (errors found/errors present in lines completed) as a function of difficulty and regulatory focus framing in Study 4 Errors Type 1 Type 2 Type 3 Type 4 Framing Promotion (n ¼ 15) 87.6% 88.7% 46.0% 33.3% Prevention (n ¼ 14) 73.1% 62.7% 58.7% 64.6% F ð3; 81Þ ¼18:43; p <:0001, indicating that type 1 errors were the most likely detected errors (M ¼ 80:6%) followed by type 2 (M ¼ 76:2%), followed by type 3 (M ¼ 52:1%), followed by type 4 (M ¼ 48:4%) errors, reflecting the fact that the participants were more accurate for the simpler errors. More germane to the hypotheses, the two-way interaction was significant as predicted, F ð3; 81Þ ¼12:19; p <:0001, showing, that difficult errors (type 3 and type 4) were better detected by prevention-framed participants than by promotionframed participants, whereas easy error (type 1 and type 2) were better detected by promotion- framed than by prevention-framed participants. The main effect of framing was non significant, F < 1. The interaction remained significant when speed was entered as a covariate into the analysis. Contrast analysis showed that the difference between promotion and prevention framing were significant for type 1, F ð1; 81Þ ¼3:81, p <:05, type 2, F ð1; 81Þ ¼12:18, p <:001, type 3, F ð1; 81Þ ¼2:94, p <:05, and type 4 errors, F ð1; 81Þ ¼17:10, p <:001. When speed was entered as a covariate into these separate analysis, all of these effects remained significant. In sum, the results indicate that participants with a promotion focus are faster than participants with a prevention focus. In general, participants with a promotion focus find more errors than participants with a prevention focus because they eagerly seek out the easy errors, whereas participants with a prevention focus vigilantly seek out the more difficult errors and find them more accurately. 8. Discussion of study 4 The results of Study 4 show that speed/accuracy trade-offs are a function of both regulatory focus and difficulty. Whereas easy errors are quickly found with a promotion focus which enhances speed, difficult errors are more accurately found with a prevention focus. In a proofreading task where the goal is to make the text overall as accurate as possible, participants with a promotion focus maximize task performance by searching for the easy errors. For them, accuracy was not generally sacrificed for speed. Instead, speed and accuracy in finding easy errors were maximized. In contrast, participants with a prevention focus did sacrifice speed for the sake of accurately finding the difficult errors. Overall then, there was no built-in trade-off: promotion focus individuals and prevention focus individuals just naturally functioned in terms of strategic eagerness and strategic vigilance, respectively. It should be noted that expectancy valence did not influence the results. That is, anticipation of positive or negative outcomes had no significant effects. This finding is consistent with previous research on regulatory focus and reflects the fact that eagerness and vigilance produce effects independent from valence of anticipation (see Higgins, 1997). 9. Concluding remarks Returning to our initial questions, our results show why some people are fast and why others are accurate. When eagerly pursuing as many hits as possible in a promotion focus, individuals are fast and efficient. In contrast, when vigilantly ensuring correct rejections in a prevention focus, they are accurate but slow. In all four studies the speed effects were independent of the accuracy effects and vice versa. Thus, regulatory focus concerns can quite naturally produce eager or vigilant strategies, leading to fast or accurate behavior (or both). Our motivational account can predict more specifically that this also varies as a function of the subjective distance to the goal (see Studies 1 3). For individuals with a promotion focus, strength of strategic eagerness is greater the closer they are to the goal and thus they are faster as the goal looms larger. In sharp contrast, for individuals with a prevention focus strength of strategic vigilance is greater the closer they are to the goal and thus they are more accurate as the goal looms larger. Those differences are obtained when regulatory focus is a personality variable and when it is a situation variable. That is chronic inclinations as well as situationallyframed task instructions can produce fast and/or accurate behavior. Because participants with a prevention framing started with a higher amount of money, they might have experienced greater risk aversion (Studies 2 4). One might wonder then, whether our results could be explained by the mechanism of risk aversion. However, even if prevention framing participants did have greater risk aversion, it is not at all clear how this would account for the full pattern of findings. The concept of risk aversion per se would not predict the promotion and prevention differential effects for speed versus accuracy nor their interaction with goals looming larger. Risk aversion would be silent with respect to this complicated set of opposite interactions for these different dependent measures. Moreover, Study 1 did not manipulate regulatory focus using framing instructions that could have

162 J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 manipulated amount of risk aversion. It examined the effects of chronic individual differences in regulatory focus and obtained the same results as the other studies. Thus, the predicted results were obtained independent of any instructional manipulation of risk aversion. One might also wonder how our work is related to other theories in goal orientation. Most research relating motivation to performance has emphasized differences in goal orientation, such as pride in success versus fear of failure (e.g., Atkinson, 1964; McClelland, Atkinson, Clark, & Lowell, 1953), intrinsic versus extrinsic (e.g., Deci & Ryan, 1985), or performance versus learning (e.g., Dweck, 1996). The distinction between a promotion focus with its eagerness strategic inclination and a prevention focus with its vigilance strategic inclination is independent of these goal orientation distinctions. For example, individuals with high pride in success can have either a promotion focus or a prevention focus. Although both have a high achievement orientation, they differ in their eagerness versus vigilance strategic inclination (see Higgins et al., 2000). As another example, both individuals in a promotion focus who use an eagerness strategy to perform a task and individuals in a prevention focus who use a vigilance strategy experience a regulatory fit that increases both their enjoyment of doing the task and the value of the task outcomes (see Freitas & Higgins, 2002; Higgins, 2000). As a final example, a performance goal of demonstrating oneõs ability could be represented as a responsibility (prevention focus) and be pursued with a vigilance strategy or it could be represented as an accomplishment (promotion focus) and be pursued with an eagerness strategy. Any goal can be pursued with either a promotion focus or a prevention focus, and thus regulatory focus is independent of type of goal orientation (see Higgins, 1997). The results of Study 4 also show that people can be both fast and accurate to the extent that the task includes easy problems that serve the needs of a promotion focus to maximize efficiency (i.e., maximize hits). On the basis of these results, it might be interesting to investigate, whether speed can also be improved in a prevention focus, such as when speed signifies an ought or responsibility. That is, in dangerous situations like heavy traffic that induce a prevention focus, vigilance could lead to fast behavior, such as faster breaking. As another example, when there is a competitive necessity to get products out of the door to consumers quickly and thus employees must avoid the mistake of being too slow, faster performances for employees in a prevention focus would be predicted. Similarly, if accuracy tasks are means to pursue a promotion focus goal, such as hitting clay targets in trapshooting, accuracy should increase within a promotion focus. The results show that framing a situation as a promotion or a prevention task can produces different strategic inclinations that can fulfill different functions in a work context. There are situations in which accurate outcomes are crucial, and situations where doing things quickly and efficiently are more important. On the basis of our results, employers can decide how to frame a task for their employees. For example if the task is easy, framing in terms of gains and non-gains might work better than framing in terms of losses and non-losses. However, if the task at hand is difficult and if accuracy is crucial, then loss and non loss framing would be more effective. Acknowledgments We thank Simon Finkeldei, Christine Haupt, Anke Siebers, and Sabine Wolfrath who served as experimenters. Special thanks to Anna Seidle who helped to draw the graphs. Lorraine Chen Idson, Ron Friedman, Nira Liberman, Fritz Strack, and Elke Weber are acknowledged for invaluable suggestions and discussions. References Atkinson, J. W. (1964). An introduction to motivation. Princeton, NJ: D. Van Nostrand. Brendl, C. M., & Higgins, E. T. (1995). Principles of judging valence: What makes events positive or negative? Advances in Experimental Social Psychology, 28, 95 160. Brewer, N., & Ridgway, T. (1998). Effects of supervisory monitoring on productivity and quality of performance. Journal of Experimental Psychology: Applied, 4, 211 227. Brickenkamp, R. (1972). Aufmerksamkeits-Belastungs-Test Handanweisung d-2. [Attention Capacity Test Manual d-2]. Goettingen, Germany: Hogrefe. Brown, J. S. (1948). Gradients of approach and avoidance responses and their relation to motivation. Journal of Comparative and Physiological Psychology, 41, 450 465. Bush, E. S., & Dweck, C. S. (1975). Reflections on conceptual tempo: Relationship between cognitive style and performance as a function of task characteristics. Developmental Pychology, 11, 567 574. Cacioppo, J. T., Priester, J. R., & Bernston, G. G. (1993). Rudimentary determinants of attitudes. II: Arm flexion and extension have differential effects on attitudes. Journal of Personality and Social Psychology, 65, 5 17. Chabot, R. J., Petros, T. V., & McCord, G. (1983). Developmental and reading ability differences in accessing information from semantic memory. Journal of Experimental Child Psychology, 35, 128 142. Chen, M., & Bargh, J. A. (1999). Consequences of automatic evaluation: Immediate behavioral predispositions to approach or avoid the stimulus. Personality and Social Psychology Bulletin, 25, 215 224. Crowe, E., & Higgins, E. T. (1997). Regulatory focus and strategic inclinations: Promotion and prevention in decision-making. Organizational Behavior and Human Decision Processes, 69, 117 132. Deci, E. L., & Ryan, R. M. (1985). Intrinsic motivation and selfdetermination in human behavior. New York: Plenum. Dickman, D. L. (1985). The tactless spouse. Medical Aspects of Human Sexuality, 19, 217 228. Dickman, S. J., & Meyer, D. E. (1988). Impulsivity and speed-accuracy tradeoffs in information processing. Journal of Personality and Social Psychology, 54, 274 290.

J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 163 Dunn, M., Vaughan, H., Kreuzer, J., & Kurtzberg, D. (1999). Electrophysiologic correlates of semantic classification in autistic and normal children. Developmental Neuropsychology, 16, 79 99. Dweck, C. S. (1996). Implicit theories as organizers of goals and behavior. In P. M. Gollwitzer & J. A. Bargh (Eds.), The psychology of action: Linking cognition and motivation to behavior (pp. 69 90). New York: Guilford. Fitts, P. M. (1954). The information capacity of the human motor system in controlling the amplitude of movement. Journal of Experimental Psychology, 47, 381 391. Fitts, P. M., & Peterson, J. R. (1964). Information capacity of descrete motor responses. Journal of Experimental Psychology, 67, 103 112. F orster, J., Higgins, E. T., & Idson, L. C. (1998). Approach and avoidance strength during goal attainment: Regulatory focus and the goal loom larger effect. Journal of Personality and Social Psychology, 75, 1115 1131. F orster, J., Higgins, E. T., & Strack, F. (2000). When stereotype disconfirmation is personal threat: How prejudice and prevention Focus moderate incongruency effects. Social Cognition, 18, 178 197. F orster, J., & Strack, F. (1997). Motor actions in retrieval of valenced information: A motor congruence effect. Perceptual and Motor Skills, 85, 1419 1427. F orster, J., & Strack, F. (1998). Motor actions in retrieval of valenced information: II. Boundary conditions for motor congruence effects. Perceptueal and Motor Skills, 86, 1423 1426. Freitas, A. L., & Higgins, E. T. (2002). Enjoying goal-directed action: the role of regulatory fit. Psychological Science, 13, 1 6. Green, D. M., & Swets, J. A. (1966). Signal detection theory and psychophysics. New York: Wiley. Hearst, E. (1960). Simultaneous generalization gradients for appetitive and aversive behavior. Science, 132, 1769 1770. Hearst, E. (1962). Concurrent generalization gradients for foodcontrolled and shock-controlled behavior. Journal of the Experimental Analysis of Behavior, 5, 19 31. Higgins, E. T. (1989). Continuities and discontinuities in self-regulatory and self-evaluative Processes: A developmental theory relating self and affect. Journal of Personality, 57, 407 444. Higgins, E. T. (1997). Beyond pleasure and pain. American Psychologist, 52, 1280 1300. Higgins, E. T. (2000). Making a good decision: Value from fit. American Psychologist, 55, 1217 1230. Higgins, E. T., Friedman, R. S., Harlow, R. E., Idson, L. C., Ayduk, O. N., & Taylor, A. (2000). Achievement orientations from subjective histories of success: Promotion pride versus prevention pride. European Journal of Social Psychology, 30, 1 23. Higgins, E. T., Shah, J., & Friedman, R. (1997). Emotional responses to goal attainment: Strength of regulatory focus as moderator. Journal of Personality and Social Psychology, 72, 515 525. Howarth, C. I., Beggs, W. D., & Bowden, J. M. (1971). The relationship between speed and accuracy of movement aimed at a target. Acta Psychologica, 35, 207 218. Josephs, R. A., & Hahn, E. D. (1995). Bias and accuracy in estimates of task duration. Organizational Behavior and Human Decision Processes, 61, 202 213. Johnson, G. A. (1975). The relative efficacy of stimulus versus reinforcement control for obtaining stable performance change. Organizational Behavior and Human Decision Processes, 14, 321 341. Keele, S. W. (1968). Movement control in skilled motor performance. Psychological Bulletin, 70, 387 403. Keele, S. W., & Posner, M. I. (1968). Processing visual feedback in rapid movements. Journal of Experimental Psychology, 77, 155 158. Lee, C., & Gillen, D. J. (1989). Relationship of Type A behavior pattern, self-efficacy perceptions on sales performance. Journal of Organizational Behavior, 10, 75 81. Leung, P. W. L., & Connolly, K. J. (1997). Test of two views of impulsivity in hyperactive and conduct-disordered children. Developmental Medicine and Child Neurology, 39, 574 582. Lewin, K. (1951). Field theory in social science. New York: Harper. Liberman, N., Idson, L. C., Camacho, C. J., & Higgins, E. T. (1999). Promotion and prevention choices between stability and change. Journal of Personality and Social Psychology, 77, 1135 1145. Liberman, N., Molden, D. C., Idson, L. C., & Higgins, E. T. (2001). Promotion and prevention focus on alternative hypotheses: implications for attributional functions. Journal of Personality and Social Psychology, 80, 5 18. Losco, J., & Epstein, S. (1977). Relative steepness of approach and avoidance gradients as a function of magnitude and valence of incentive. Journal of Abnormal Psychology, 86, 360 368. Maruff, P., Wilson, P., Trebilcock, M., & Currie, J. (1999). Abnormalities of imagined motor sequences in children with developmental coordination disorder. Neuropsychologia, 37, 1317 1324. Malhotra, L., Malhotra, D., & Jerath, J. M. (1989). Speed and accuracy in learning as a function of personality. Journal of Personality and Clinical Studies, 5, 5 8. McClelland, D. C., Atkinson, J. W., Clark, R. A., & Lowell, E. L. (1953). The achievement motive. New York: Appleton-Century- Crofts. Meyer, D. E., Abrams, R. A., Kornblum, S., & Wright, C. E. (1988). Optimality in human motor performance: Ideal control of rapid aimed movements. Psychological Review, 95, 340 370. Meyer, D. E., Smith, J. E., & Wright, C. E. (1982). Models for the speed and accuracy of aimed movements. Psychological Review, 89, 449 482. Miller, L., & Vernon, P. A. (1997). Developmental changes in speed of information processing in young children. Developmental Psychology, 33, 549 554. Miller, N. E. (1944). Experimental studies of conflict. In J. M. V. Hunt (Ed.), Personality and the behavior disorders (vol. 1, pp. 431 465). New York: Ronald Press. Miller, N. E. (1959). Liberalization of basic S-R concepts: Extensions to conflict behavior, motivation and social learning. In S. Koch (Ed.), Psychology: A study of a science General systematic formulations, learning and special processes (vol. 2, pp. 196 292). New York: McGraw-Hill. Miller, N. E., & Murray, E. J. (1955). Displacement: Greater generalization of approach than avoidance in a generalized approach-avoidance conflict. Journal of Experimental Psychology, 43, 227 231. Phillips, L. H., & Rabbitt, P. M. A. (1995). Impulsivity and speedaccuracy strategies in intelligence test performance. Intelligence, 21, 13 29. Priester, J. R., Cacioppo, J. T., & Petty, R. E. (1996). The influence of motor processes on attitudes toward novel versus familiar semantic stimuli. Personality and Social Psychology Bulletin, 22, 442 447. Raesaenen, P., & Ahonen, T. (1995). Arithmetic disabilities with and without reading difficulties: A comparison of arithmetic errors. Developmental Neuropsychology, 11, 275 295. Revelle, W. (1986). Motivation and efficiency of cognitive performance. In D. R. Brown & J. Veroff (Eds.), Frontiers of motivational psychology: Essays in honor of John W. Atkinson (pp. 107 131). New York: Springer. Revelle, W. (1987). Personality and motivation: Sources of inefficiency in cognitive performance. Journal of Research in Personality, 21, 436 452. Revelle, W., Anderson, K. J., & Humphreys, M. S. (1987). Empirical tests and theoretical extension of arousal based theories of personality. In J. Strelau & H. J. Eisenck (Eds.), Personality dimensions and arousal: Perspectives on individual differences (pp. 17 33). New York: Plenum Press.

164 J. F orster et al. / Organizational Behavior and Human Decision Processes 90 (2003) 148 164 Revelle, W., & Leon, M. R. (1985). Effects of anxiety on analogical reasoning: A test of three theoretical models. Journal of Personality and Social Psychology, 49, 1302 1315. Roney, C. J. R., Higgins, E. T., & Shah, J. (1995). Goals and framing: How outcome focus influences motivation and emotion. Personality and Social Psychology Bulletin, 21, 1151 1160. Salkind, N. J., & Nelson, C. F. (1980). A note on the developmental nature of reflection impulsivity. Developmental Psychology, 16, 237 238. Seibt, B. & F orster, J. (2001). Risky and careful processing under stereotype threat: how regulatory focus can enhance and deteriorate performance when self stereotype are active. Ubpublished manuscript, Universit at W urzburg, under submission. Seppaelae, P., & Salvendy, G. (1985). Impact of depth of menu hierarchy on performance effectiveness in a supervisory task: Computerized flexible manufacturing system. Human Factors, 27, 713 722. Shah, J., & Higgins, E. T. (1997). Expectancy value effects: Regulatory focus as determinant of magnitude and direction. Journal of Personality and Social Psychology, 73, 447 458. Shah, J., Higgins, E. T., & Friedman, R. (1998). Performance incentives and means: How regulatory focus influences goal attainment. Journal of Personality and Social Psychology, 74, 285 293. Snowling, M., Hulme, C., & Goulandris, N. (1994). Word recognition in developmental dyslexia: A connectionist interpretation. Quarterly Journal of Experimental Psychology: Human Experimental Psychology, 47A, 895 916. Socan, G., & Bucik, V. (1998). Relationship between speed of information processing and two major personality dimensions extraversion and neuroticism. Personality and Individual Differences, 25, 35 48. Solarz, A. K. (1960). Latency of instrumental responses as a function of compatibility with the meaning of eliciting verbal signs. Journal of Experimental Psychology, 59, 239 245. Tanner, W. P., & Swets, J. A. (1954). A decision-making theory of visual detection. Psychological Review, 61, 401 409. Taylor, M. S., Locke, E. A., Lee, C., & Gist, M. E. (1984). Type A behavior and faculty research productivity: What are the mechanisms? Organizational Behavior and Human Decision Processes, 34, 402 418. Tucker, P., & Warr, P. (1996). Intelligence, elementary cognitive components, and cognitive styles as predictors of complex task performance. Personality and Individual Differences, 21, 91 102. Weinstein, N. D. (1977). Noise and intellectual performance: A confirmation and extension. Journal of Applied Psychology, 62, 104 107. Woodworth, R. S. (1899). Accuracy of voluntary movements. Psychological Review, 3, 1 101. Zelaznik, H. N., Mone, S., McCabe, G. P., & Thaman, C. (1988). Role of temporal and spatial precision in determining the nature of the speed-accuracy trade-off in aimed-hand movements. Journal of Experimental Psychology: Human Perception and Performance, 14, 221 230. Received 4 December 2000