• No results found

The self-serving attribution bias and overestimation : a replication with the intent of debiasing

N/A
N/A
Protected

Academic year: 2021

Share "The self-serving attribution bias and overestimation : a replication with the intent of debiasing"

Copied!
29
0
0

Bezig met laden.... (Bekijk nu de volledige tekst)

Hele tekst

(1)

Joris Amin

10677038

BSc Economie & Bedrijfskunde

Universiteit van Amsterdam

(2)

1 Statement of Originality

This document is written by Student Joris Amin who declares to take full responsibility for the contents of this document.

I declare that the text and the work presented in this document is original and that no sources other than those mentioned in the text and its references have been used in creating it.

The Faculty of Economics and Business is responsible solely for the supervision of completion of the work, not for the contents.

(3)

2

Abstract

Prior research has proven that people are subject to the self-serving attribution bias. This means that people tend to attribute success to internal factors such as skill and effort and failure to external factors such as difficulty and luck. Furthermore, Libby & Rennekamp (2012) argue that this leads people to become overconfident. An experiment was conducted in which participants answered two rounds of trivia questions. Between these rounds, participants indicated their attributions to internal vs. external factors and their confidence in improving in the second round. It was confirmed that people indeed engage in self-serving attribution. Additionally, this study aimed at finding a way of debiasing. Sufficient evidence was found to state that people do not show overconfident behavior regarding an upcoming task when they are able to compare the difficulty of the tasks.

(4)

3

Inhoudsopgave

Introduction ...4

Theorethical Framework ...5

Bounded rationality ...5

The self-serving attribution bias ...6

Overconfidence ...8

The self-serving attribution bias & overestimation ...9

Information about task difficulty ...9

Methods ... 11 Testing H₁ ... 12 Testing H₂... 13 Post-hoc testing ... 13 Results ... 14 Descriptive statistics ... 14 Support for H₁ ... 15 Support for H₂ ... 16 Post-hoc analysis ... 18

Discussion & Conclusion ... 18

Conclusion ... 18

Limitations ... 19

Suggestions ... 19

Appendix ... 20

(5)

4

Introduction

Economic science, like science in general, uses a wide variety of models. Models are idealizations that are used to represent reality for specified purposes (Giere, 2004). Some of these models are created to value stocks, others try to identify the perfect style of leadership. In order to function and give a good representation of reality, many models rely on assumptions. A variable in many economic models is human behavior. To account for human behavior in models in economic science, John Stuart Mill proposed the theory of the ‘economic man’ (Persky, 1995). This theory assumes that the economic man’s goal is to maximize his utility. He therefore always acts rational, considers all options in decision-making, is not biased towards certain alternatives and does not satisfice (Simon, 1955).

Yet this economic man has endured quite some criticism concerning his rationality. Research has uncovered consistent deviations from the predictions of this theory (Henrich et al., 2016). To enhance this theory and give a more realistic representation of the behavior of economic agents, psychologists, sociologists and anthropologists help improve the characterizations of economic behavior (Thaler, 2016). Due to the notion of bounded rationality, economists have a more realistic view concerning the behavior of economic agents. “The notion of bounded rationality was proposed in the mid- 1950s to connect, rather than to oppose, the rational and the psychological” (Gigerenzer & Selten, 2002, p. 1). In economic terms, bounded rationality refers to the manager’s limits in their ability to process and interpret a large volume of information in their decision-making activities (Simon, 1979). In general, this implies that people have cognitive limitations and other constraints that hinder them from being rational. Since people are subject to these cognitive limitations, it is argued that people are biased. This means that human cognition portrays a version of reality that is systematically distorted compared to some aspects of the objective reality (Haselton, Nettle, & Andrews, 2005).

One of the cognitive limitations that people are subject to is the self-serving attribution bias. The self-serving attribution bias is a phenomenon that has received much attention in psychology and economic literature. This bias causes people to attribute their success to their own dispositions and failure to external forces (Miller & Ross, 1975). The implications of the self-serving attribution bias are widespread. It is argued that the bias affects employees in their attributions as to why they did, or did not, get the promotion (Shepperd, Malone, & Sweeny, 2008). Further, Shepperd et al. (2008) argue that athletes are the victim of this bias in their attribution to their performance and drivers are biased in their explanations of accidents. Often, examples can be found of the bias affecting the issuance of management forecasts (Libby & Rennekamp, 2012), organizational planning

(6)

5 (Larwood & Whittaker, 1977), relationships (Sedikides, Campbell, Reeder, & Elliot, 1998), self-esteem (Schlenker, Weigold, & Hallam, 1990) and many more cases in human decision making (Evans, 1989).

According to Libby & Rennekamp (2012), the self-serving attribution bias is a cause for overconfidence. Research literature defines the manifestation of overconfidence in three ways: overestimation of one’s actual performance, overplacement of one’s performance relative to others and excessive precision in one’s beliefs (Moore & Healy, 2008). The consequences of overconfidence are important considering that the bias causes one to overestimate the extent to which their internal characteristics contribute to a better performance. It is argued that overconfidence affects for example income prospects and business failure (Camerer & Lovallo, 2016), clinical psychologists’ judgement (Oskamp, 1965), investment decisions and corporate investment distortions (Lambert, Bessière, & N’Goala, 2012; Malmendier & Tate, 2005), driving ability (Svenson, 1981).

Numerous studies address the causes and consequences of both the self-serving attribution bias and overconfidence. Yet little research has been devoted to successfully reducing the effects of these biases, or debiasing. This research focuses on the relationship between the self-serving attribution bias and overconfidence. Furthermore, it aims to find a way of debiasing. The first part of this research aims to replicate the study conducted by Libby & Rennekamp (2012) to test whether performance is explained through the self-serving attribution bias and whether this indeed leads to overconfidence. The purpose of the second part is to test whether information about the difficulty of a task leads to debiasing.

The following section consists of a theoretical framework that discusses existing literature concerning this topic. Subsequently, a step-by-step description of the conducted experiment is given. Afterwards, the results of this experiment are stated. Conclusively, the limitations and conclusions are discussed.

Theoretical framework

Bounded rationality

“The notion of a bounded rationality was proposed in the mid- 1950s to connect, rather than to oppose, the rational and the psychological” (Gigerenzer & Selten, 2002, p. 1). This was done by Herbert Simon in 1955 as a reaction to the theory of homo economicus. He wrote:

Broadly stated, the task is to replace the global rationality of economic man with a kind of rational behavior that is compatible with the access to information and the computational capacities that are actually possessed by organisms, including man, in the kinds of

(7)

6 environments in which such organisms exist. One is tempted to turn to the literature of psychology for the answer. (Simon, 1955 pp. 99-100)

Herbert Simon stated in his book, published in 1982, that the term ‘bounded rationality’ is used to designate rational choice that takes into account the cognitive limitations of the decision maker. He made a distinction between limitations of both knowledge and computational capacity (Simon, 1982). After many years of research into bounded rationality, many examples of cognitive biases can be found in scientific literature. Carter, Kauffman & Lust (2007) constructed a list of 76 cognitive biases in decision-making alone. For instance, the ‘endowment effect’ describes the tendency of people to value something higher as soon as they own it. Another example is the ‘confirmation bias’ which describes the tendency of decision-makers to seek confirmatory information and exclude the search for disconfirming information.

The self-serving attribution bias

All of these biases, including the self-serving attribution bias, limit humans in their capacity for rationality while making decisions. Since this bias has been given different labels, such as ‘fundamental attribution error’ and ‘self-serving attribution bias’ it shall from now on, for writing purposes, be termed the SSAB. Ross (1977) defines the SSAB as the tendency for attributers to underestimate the impact of situational factors and to overestimate the role of dispositional, or internal factors in controlling behavior. This implies that people overestimate to what extent they have personally influenced a situation and underestimate the role of factors they could not influence. Furthermore, according to Schlenker et al. (1990), people tend to take greater personal responsibility for successes than failures. People tend to attribute their successes to identity-relevant factors such as ability and effort, yet attribute failures to identity-irrelevant factors, such as bad luck and task difficulty.

Plenty of attention has been devoted to researching the SSAB and there has been quite some discussion regarding this topic. By reviewing evidence for and against the proposition of the SSAB, Miller & Ross (1975) argued that there is little support that individuals engage in self-protective attributions under conditions of failure, such as attributing the outcome to bad luck. They reviewed multiple ‘achievement task’ studies in which subjects were asked to perform a task and aim for the maximum outcome. These tasks would vary from pattern recognition, number estimation, simulation games and solving anagrams (Chaikin & Darley, 1973; Davis & Davis, 1972; Feather, 1969; Fitch, 1970; Streufert & Streufert, 1969). After completing the task, the participants perspectives as to why they succeeded or failed was assessed. The scientists who conducted these studies proclaimed to have found evidence that subjects indeed attributed success to their identity-relevant factors such as

(8)

7 ability and effort and failure to external factors. Another study, conducted by Johnson, Feigenbaum & Weiby (1964), was ubiquitously cited as a demonstration of the SSAB. In this ‘interpersonal influence’ study, subjects had to teach two children to multiply by respectively 10 and 20. When multiplying by 10, child A had done well and child B had done poorly. When multiplying by 20, child A had continued to do well and child B had either continued to do poorly or he/she improved. When the attributions of the subjects as to why the children performed well or poorly were assessed, they found that subjects took credit for improved performance and blamed the child for bad

performance. However, Miller & Ross (1975) found that due to incautious examination and design faults in the method of assesment, the evidence indicating that individuals attribute failure to external factors was inconclusive. Amongst other things, they argued that all subjects taught a child who performed well and consistently (child A) and could therefore rationally assume that their teaching methods were sufficient. Consequently, they attributed a bad performance by child B to other, external, factors. Yet, Miller & Ross (1975) acknowledge that individuals engage in self-enhancing attributions under conditions of success.

This expression of criticism evoked a reaction. Bradley (1978) states:

Miller and Ross’s (1975) provocative review highlighted the equivocal nature and methodological inadequacies of some of the studies often cited as support for self-serving attributional biases. However, a considerable amount of new research has emerged since their review appeared, and much of this research provides more conclusive evidence regarding the influence of motivational biases in the causal inference process. (Bradley, 1978, p. 57)

Bradley’s reexamination (1978) draws on many ‘interpersonal influence’ and ‘skill-oriented task performance’ studies and found strong support for self-serving biases. “Individuals tend to accept responsibility for positive behavioral outcomes and to deny responsibility for negative behavioral outcomes,” according to Bradley (1978, p. 68).

In recent years there have been no expressions of doubt or criticism concerning the SSAB. Moreover, the focus of research regarding the SSAB shifted towards its moderators and

consequences which yielded interesting results. For instance, a moderator that was discovered by Clapham & Schwenk (1991) is gender. Clapham & Schwenk found that men were more likely to display the SSAB than woman. Arguably because men tend to have higher success expectancies of upcoming tasks and more ego-involvement, which makes them more susceptible to the bias (Rosenfield & Stephan, 1978). Furthermore, relationships between people seem to affect the SSAB too. In relationships, people do not differ in terms of personal responsibility they take for success or

(9)

8 failure. In addition, closely related individuals engage less in self-enhancement. This means that individuals take less credit for success and display a lesser tendency to blame others for failure (Sedikides et al., 1998). Regarding the SSAB’s consequences, Clapham & Schwenk (1991) analyzed the letters to shareholders in annual reports of electric and gas utilities. They found that management takes credit for good performance, yet do not take responsibility for bad outcomes. Moreover, they argue that self-serving attributions in annual reports are negatively related to future firm

performance. One of the possible explanations they provide is that biased patterns of attributions reduce the effectiveness of strategic decision-making. While this consequence of the SSAB is detrimental, other research provides empirical evidence of positive manifestations of the SSAB. It is argued that unrealistically positive self-evaluations promote happiness as well as other aspects of mental health (Taylor & Brown, 1988).

Overconfidence

Libby & Rennekamp (2012) argue that since individuals tend to overestimate the extent to which their internal characteristics are contributing to a better performance, people become overconfident. The notion of overconfidence already occurred in 1965 when Oskamp conducted a study with 32 judges. He found that, as judges study a case, the confidence about their decision regarding the case increases steadily but the accuracy of their conclusions quickly reaches a ceiling. From these 32 judges, all but two displayed this overconfident behavior. As Plous stresses: “No problem in judgement and decision making is more prevalent and more potentially catastrophic than overconfidence” (1993, p. 217). Moore & Healy (2008) examined 183 studies concerning

overconfidence and provide three distinct ways in which this phenomenon was researched. They argue that one of three manifestations of overconfidence is that people overestimate one’s actual performance. Another way overconfidence manifests itself is the better-than-average effect through which the majority of people believe themselves to be better than others. This is statistically

impossible. The third way of displaying overconfidence is what Moore & Healy term overprecision. Results show that when people are asked to estimate a 90% confidence interval around their answer in for example a number estimation game, their confidence intervals are too narrow. This suggests that people are too certain that they gave the correct answer.

This study’s focus is on only one of the three displays of overconfidence described by Moore & Healy (2008), the overestimation bias. Regarding this topic, Moore & Healy examined 168 studies that investigate behavior of overestimation. For example, a student overestimated himself when he took a 10-item quiz and he believed that he answered five of the questions correctly when the reality was that he only gave three correct answers.

(10)

9

The SSAB & overestimation

Research of Libby & Rennekamp (2012) provides evidence of the relationship between the SSAB and overestimation. They argue that a positive performance leads individuals to engage in self-serving attribution. In turn, this self-self-serving attribution is strongly associated with confidence in a better performance. In their experiment, a participant first answers 25 trivia questions on either high or low difficulty. Participants in the low difficulty group received a mix of fifteen ‘easy’ questions, 5 ‘medium’ questions and 5 ‘hard’ questions. The mix in the ‘high’ difficulty group consisted of five ‘easy’, five ‘medium’, and fifteen ‘hard’ questions. Participants were not aware of the mix of difficulty levels of the questions. Libby & Rennekamp did not inform the participants of this mix so as to intentionally create ambiguity in the complexity of the task which, in turn, they argued, would lead people to engage in self-serving attribution. Participants then learn their score and answer questions designed to measure the SSAB. Subsequently, participants are informed of the actual mix of question difficulty in the second round. It could be argued that this information is of less value since

participants are not able to compare this with the first round. Participants then indicate their confidence in improving and provide a forecast of their expected performance in the second round. Finally, participants choose whether to commit to improving, which would alter their payoff and play a second round. Libby & Rennekamp (2012) found that participants who performed well in the first round attributed this to internal factors, in this case, skill and effort. They subsequently found evidence that participants who did well and thus attribute this to internal factors, became significantly more confident of their ability to perform better in the next round. This increase in confidence, stemming from a past positive result, is not surprising. Yet it could be argued that it is not entirely rational either since the difficulty of the first round is ambiguous.

Since this study partly aims to replicate the study conducted by Libby & Rennekamp (2012) to test whether performance is explained through the SSAB a similar hypothesis is constructed:

H₁: High (low) ratings of first-round performance will be explained by the participants through the self-serving attribution bias, and attributed to internal (external) factors.

It is expected that participants who experience success in the first round engage in self-enhancing attributions. It is also expected that participants who believe that they did not perform well in the first round explain this through the SSAB and attribute this to external factors.

Information about task difficulty

Simultaneously, this study tries to find a way of debiasing. It aims to do so by testing whether enabling people to compare information about the task difficulty causes people to be less susceptible

(11)

10 to biases. Lipman stresses that it is evident that one of the main causes of humans’ bounded

rationality is a lack of information (1995). Similarly, the criticism expressed by Simon (1955) on the theory of homo economicus was, amongst other things, directed at its access to information. People are forced to make decisions subject to cognitive limitations and numerous information problems (Doucouliagos, 1994). Therefore, it could be argued that if a person acquires more information, he/she is able to more accurately analyze a situation. This should result in better estimates, decisions and less biased views on a situation. In micro-economics and game theory, much research has been conducted with regard to decision-making in circumstances in which there is incomplete information. As Potter & Anderson (1980) state, people make more adequate decisions with accurate information. However, they also state, that when accurate information is not available, any inference is doubtful. Since a lack of information or ambiguity in a situation creates biased views, it is assumed that more information allows for more rational decision-making.

It could be argued that when this information contains the difficulty of the task, it enables people to more rationally estimate how well they will perform a task since they know more about the task than when they did not know the difficulty. If this is the case, then people would become less biased by learning the task difficulty. Yet not much research into debiasing using information about task difficulty can be found in research literature. However, Mowen, Keith, Brown & Jackson (1985) found a relationship between these concepts. In their research, conducted with sales managers who evaluate the performance of a salesperson, they found a bias in the performance evaluations of these salespersons. Their results show a tendency of sales managers assigning greater weight to internal factors of a salesperson than situational factors that might have influenced the salesperson’s performance. Moreover, they conclude by stating that sales managers exclusively use information about effort to evaluate a salesperson. Mowen et al. argue that this pattern is a

manifestation of the SSAB. However, this manifestation of the SSAB is not the same as the SSAB described earlier. This manifestation is concerned with evaluations by others instead of self-evaluations. Mowen et al. propose a solution to hinder this bias by increasing the use of task difficulty information. This way, a person’s actual performance can be compared with the expected performance. Similarly, Fischhoff (1982) proposes a strategy to cope with the bias of overconfidence. He states that a strategy to cope with a person’s overconfidence was proposed to assessors of a person. These assessors could make use of a correction factor that indicates the person’s certainty (Lichtenstein & Fischhoff, 1977). For example, if a person would say that he/she is certain, the assessor could read it as an 85% chance of them being correct. Unfortunately, it is impossible to perfectly estimate this correction factor since, according to Fischhoff, it depends on the difficulty of the task at hand. A very difficult task should accordingly yield a low probability of certainty

(12)

11 (Lichtenstein & Fischhoff, 1977). Yet this technique for accounting for overconfidence is, similar to the proposed solution by Mowen et al., designed for an evaluator. In existing scientific literature, it is still not proven that a person is able to become less biased in assessing their own confidence in performing a task by knowing the task difficulty.

However, participants in the study of Libby & Rennekamp (2012) were informed of the difficulty level of the second round and still showed a bias when asked to indicate their confidence in improving. It could therefore be argued that existing scientific literature answers the question whether knowing the task difficulty has a debiasing effect negatively. Yet, as stated, participants were not able to compare the difficulties of both rounds. Therefore, participants did not really know the difficulty. Fifteen easy questions, five medium questions and 5 hard questions do not imply that this is an ‘easy’ round for all participants. A participant has to experience performing a task of similar difficulty so that he/she is able to compare both tasks. It could be argued that, in this scenario, a person can accurately adjust their expectations about a task and estimate their confidence in improving in this task more rationally.

Research literature, as described, states that the more information people acquire, the more this enables them to provide better estimates and be more rational in decision-making. Furthermore, research literature clearly indicates that, to evaluate a person’s performance, the task difficulty has to be known. It does however not indicate whether a person becomes less biased when they acquire knowledge of the task difficulty. As proven by Libby & Rennekamp (2012), merely labeling the difficulty of a task does not have a debiasing effect. It could be argued that people become less biased when they acquire information about a task’s difficulty when they have already performed a task of similar difficulty. Therefore, a second hypothesis is constructed.

H₂: One is able to more rationally estimate whether one’s second-round performance is better than one’s first-round performance when one knows the difficulty of both tasks.

Methods

As mentioned in the introduction, this study’s focus is partly to replicate the study conducted by Libby & Rennekamp (2012). The other aim of this study is to find a way of debiasing by informing subjects of the difficulty of the task. Libby & Rennekamp used a (1 x 2) between-subjects design so that they were able to compare the data from the low-difficulty group with the high-difficulty group. This data indicated that the low-difficulty group attributed their, usually good, performance

significantly more to internal factors, hence the confirmation of their hypothesis. Due to time constraints and the lack of technology in acquiring the data, this study was carried out with a

(13)

within-12 subjects design. The tasks were, similar to Libby & Rennekamp (2012), to answer trivia questions from the board game Trivial Pursuit. The performance of subjects is measured in the total questions answered correctly. The sequence of the tasks subjects had to perform is described here in

chronologic order.

First, subjects answered a round of 20 trivia questions. Hasbro’s game Trivial Pursuit, or ‘Triviant’ in Dutch, has questions of three different difficulty levels. The questions are either easy, medium or hard to answer. Opposite to Libby & Rennekamp’s (2012) study, all subjects received the same mix of questions. This mix of questions was designed to have such complexity that

approximately half of the participants would score high and half of the participants would score low. The first round consisted of ten easy questions and ten medium questions. Questions were solely selected from the ‘Science & Nature’, ‘Arts & Literature’, ‘Geography’ and ‘History’ categories to decrease the likelihood of an unequal challenge for subjects of different genders or ages. Subjects were aware of the mix, unlike Libby & Rennekamp’s experiment where subjects did not know the mix in difficulty in the first round. This difference in design is crucial since it will determine whether this treatment has a debiasing effect or not. The questions asked in the first round can be found in the appendix (Figure 1). This question sheet is the first sheet that the subjects received.

Testing H₁

Upon completing the first round, subjects were informed of the amount of questions they answered correctly. Subsequently, subjects’ relative attribution of internal versus external factors was assessed. Similar to Libby & Rennekamp who assessed this with a method used by Scapinello (1989), subjects provided ratings on four 9-point bipolar scales. These scales were ‘Skill vs. Luck’, ‘Skill vs. Difficulty’, ‘Effort vs. Luck’ and ‘Effort vs. Difficulty’. The attribution to internal factors by subjects are represented by ‘Skill’ and ‘Effort’. The attribution to external factors by subjects are represented by ‘Difficulty’ and ‘Luck’. If a subject attributed a good performance relatively more to their skills or to the effort put into the task, then this implies that they attribute their performance to internal factors.

Furthermore, subjects were asked to indicate how well they thought they had performed in the first round. This was done on a 9-point Likert scale ranging from ‘strongly disagree I did well’ to ‘strongly agree I did well’. At this point, subjects knew the amount of questions they had answered correctly. However, fifteen correct questions might be considered a success by one subject and a failure by another. Therefore, this measure was used to see to what extent a subjects believed he or she performed well. If a subject answered this question with an ‘8’, it is assumed that this person is firmly convinced that he or she performed well. After this data is acquired, statistical tests can be

(14)

13 conducted to indicate whether or not the first hypothesis is accepted. It is expected that subjects who believe they did well attribute this to internal factors. Furthermore, it is expected that subjects who believe that they did not do well attribute this to external factors. This would prove that subjects explain their first-round performance through the SSAB.

Testing H₂

After the subjects answered these questions, they learned that the second round is identical in difficulty to the first round. The second round also consisted of ten easy questions and ten

medium questions. Subjects were asked to indicate their confidence in improving their performance in the second round. It was asked whether subjects, similar to Libby & Rennekamp (2012), thought they would do better or worse in the second round. Again, a 9-point Likert scale was used to

estimate to what extent subjects believed they would improve. The Likert scale ranged from ‘1 = I am 100% certain that I will do worse in the second round than I did in the first round’ to ‘9 = I am 100% certain that I will do better in the second round than I did in the first round’. If a subject answered with a ‘7’, then this person is 50% certain of improving their performance in the second round.

Libby & Rennekamp found in their experiment that subjects who believed that they performed well, and who attributed this to internal factors, showed a bigger tendency to think that they would improve in the next round and become overconfident. Yet these subjects did not know the difficulty of the first round. The ambiguity of not knowing, or not being able to compare, the difficulty of the tasks is deliberately absent in this study. This way, it will become apparent whether having information about the task difficulty enables a subject to be less susceptible to biases. It is expected that although the susceptibility to overconfidence remains, this bias manifests itself to a lesser extent. That is, a subjects is able to more rationally estimate their confidence in improving their performance of the second. Therefore, the data should show an increase in confidence but to a lesser extent than it did in the results of the study conducted by Libby & Rennekamp. If this is the case, then it could be argued that this decrease in ambiguity has a debiasing effect. By comparing the data generated by this question with the results from Libby & Rennekamp, a decision can be made concerning the second hypothesis. The questions concerning the attributions of subjects, their belief in their performance and their confidence in improving can be found in the appendix (Figure 2). It was printed on the backside of the sheet with the questions of round 1.

Post-hoc analysis

After subjects answered the question that asked them to indicate to what extent they believed that they would improve, subjects started the second round of trivia. As stated, this round contained exactly the same amount and mix of difficulty of the questions. This data allows for a post

(15)

14 hoc test. For instance, if a subject answered fifteen questions correctly in the first round, and later stated that he/she is 100% sure of improving, then this person might be correct. He or she might answer eighteen questions correctly in the second round and has therefore not overestimated him/herself. Therefore, true overestimation is shown when the cases in which participants were correct about their confidence in improving are omitted in the analysis. The questions of the second round can be found in the appendix (figure 3).

Results

Descriptive statistics

Since four subscales were used in the experiment to measure whether participants attributed internally or externally, Cronbach’s alpha was calculated to indicate the reliability of this measure. As can be seen in the appendix (Figure 2.1), Cronbach’s alpha has a value of 0,571. It could be argued that this is a poor value of Cronbach’s alpha. Nunnely (1978), who is ubiquitously cited for his view on an acceptable level of Cronbach’ alpha, states that its value should at least be higher than 0.7. Further analysis of Cronbach’s alpha indicated that, although all items are positively correlated (Figure 2.2), the scale ‘Luck vs Skill’ is the least correlated with the other scales in this measure. The correlation of ‘Luck vs Skill’ with the total measure of attribution if ‘Luck vs Skill’ was left out is only 0.280, whereas this value ranges from 0.349 to 0.404 for the other scales (Figure 2.3). However, in the last column of figure 2.3 can be seen that Cronbach’s alpha cannot be inflated by deleting this item since this would result in a Cronbach’s alpha of 0.567.

After the analysis of Cronbach’s alpha, the four subscales designed to measure the

participant’s attribution were combined. They were first recoded in such a way that a value higher than five indicates internal attribution and a value lower than five indicates external attribution, where 1 and 9 are possibly the most extreme values. This variable is called ‘MeanATT’. Furthermore, the ‘SCORE’ variable indicates the amount of correct answers in the first round. The ‘DIDWELL’ variable indicates whether a participant believed he/she performed well. A value from one to five means a bad performance according to the participant, and a value from 5 to 9 implies a good performance. The ‘CONFIDENCE’ variable expresses a participant’s belief in either performing worse (values of one to five), or better (values of five to nine). Lastly, the variable ‘SCORE2ro’ indicates a participant’s score in the second round. See table below for descriptive statistics.

(16)

15 None of the variables presented in the table is significantly skewed. Neither is the kurtosis of any variable significant. Furthermore, the central limit theorem states that when a sample is large enough (n=30), the distribution is normal (Field, 2013). Since this sample contains 60 participants, the sample can be considered large enough to be normally distributed, as can be seen in the

histogram in the appendix (Figure 3). These statistics are quite similar to the statistics in the study of Libby & Rennekamp (2012). Participants in this study averaged 12 correct answers out of 20

questions, that is a score of 60%. The participants from the study of Libby & Rennekamp averaged 14,149 correct answers from 25 questions, which implies that they correctly answered 57% of the questions.

Support for H₁

To examine whether high (low) ratings of first-round performance will be explained by the participants through the SSAB, and attributed to internal (external) factors, a linear regression was used. Similar to Libby & Rennekamp (2012), the dependent variable in the regression analysis was the mean of the attribution variable and the independent variable is represented by ‘DIDWELL’. Before the regression analysis, the data has been checked for outliers. One outlier was found in the attribution variable. However, it was not deleted since the outlier was not caused by erroneous data, nor did it influence the analysis significantly. The regression analysis showed a significant relationship between ‘DIDWELL’ and ‘MeanATT’ with a t-ratio of 3,492 and a p-value of 0,001. Therefore,

sufficient evidence is found to support the first hypothesis. Results of the regression analysis can be found in the table below.

Descriptive Statistics

SCORE MeanATT DIDWELL CONFIDENCE SCORE2ro

N Valid 60 60 60 60 58 Missing 0 0 0 0 2 Mean 12,07 4,6167 5,12 5,27 11,78 Median 12,00 4,7500 6,00 5,00 12,00 Mode 13 5,25 6 5 12 Std. Deviation 2,564 1,18381 2,067 1,582 2,492 Skewness -,089 -,078 -,066 ,207 -,038 Std. Error of Skewness ,309 ,309 ,309 ,309 ,314 Kurtosis -,357 ,001 -,981 ,872 ,311 Std. Error of Kurtosis ,608 ,608 ,608 ,608 ,618 Minimum 6 2,25 1 1 5 Maximum 18 7,75 9 9 18

(17)

16 Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig. 95,0% Confidence Interval for B

B Std. Error Beta Lower Bound Upper Bound

1 (Constant) 3,396 ,377 9,013 ,000 2,641 4,150

DIDWELL ,239 ,068 ,417 3,492 ,001 ,102 ,375

a. Dependent Variable: MeanATT

Participants attribute high ratings of the round performance internally and low ratings of first-round performance externally. As shown in the appendix (Figure 4.1 & figure 4.2), there is a significant (p = 0,001, F = 12,193) positive correlation (R = 0,417) between the participant’s belief about how well they did and their attribution to internal factors. This implies that R² equals 0,174. Therefore, 17,4% of the variance in attribution can be accounted for by the participants belief as to how well he/she performed.

Since the regression method of least squares was used, there are several criteria that must be accounted for. First, the errors in the model cannot be related to each other. Since all participants filled the questions in separately, this can be assumed. Furthermore the estimates of the parameters in the model are only optimal when homogeneity of variances is assumed (Field, 2013). To test for homoscedasticity, the participants were grouped per value of ‘DIDWELL’. Then, Levene’s test was used to see whether the variance in attribution in these 8 groups (the group with value ‘1’ was omitted since it contained one participant) was significantly different. As shown in the appendix (Figure 4.3), there is no sufficient evidence to prove that there is a difference in the variances and homoscedasticity is proven. Thus it can be safely assumed that there is enough evidence to support the first hypothesis

Support for H₂

It was hypothesized that one is able to more rationally estimate whether one’s second-round performance is better than one’s first round performance when one knows the difficulty of both tasks. This was tested by comparing the data from this study to the data from the study of Libby & Rennekamp (2012). Again, a linear regression analysis was used. Before the analysis, Levene’s test for homoscedasticity was conducted. The result of this test can be found in the appendix (Figure 5). According to Levene’s test, homogeneity of variances can be assumed. The results from the regression analysis are presented in the table below.

(18)

17 As the p-value of the ‘MeanATT’ variable indicates, there is no significant relationship between a participant’s attribution and his/her confidence in improving in the second-round. Neither is there a significant relationship between ‘DIDWELL’ and a participant’s confidence in improving. These results are contradictory to the results of Libby & Rennekamp (2012). They did find a significant relationship between a participant’s attribution and his/her confidence and explained this by stating that the self-serving attribution leads people to be too confident in improving.

The data from this study proves that it was not the self-serving attribution that led to the increase in confidence in their study, yet it was the ambiguity of the task difficulty. Almost all of the variables in this study and the study of Libby & Rennekamp (2012) are equal. All participants were students. The same questions were used to measure the variables ‘MeanATT’, DIDWELL’ and ‘CONFIDENCE’. Furthermore, the task participants were asked to do was equal, answering trivia questions from the same levels of difficulty. In contrast to this study, participants in the study of Libby & Rennekamp received either an easy or a hard mix of questions, while all participants in this study received the same mix. However, as stated, the percentage of the average amount of

questions answered correctly is similar, 57% vs. 60%. It can therefore be assumed that the difficulty of the questions did not influence the results. Also, this study did not ask participants whether to commit or not to improving in the second round. Yet, this question was asked in Libby &

Rennekamp’s study after participants declared their confidence in improving. It can therefore be argued that this too did not influence these results. The main difference between both studies is the

Coefficientsa Model Unstandardized Coefficients Standardized Coefficients t Sig.

95,0% Confidence Interval for B

B Std. Error Beta Lower Bound Upper Bound

1 (Constant) 4,836 ,854 5,664 ,000 3,126 6,546 DIDWELL ,138 ,110 ,180 1,251 ,216 -,083 ,358 MeanATT -,059 ,192 -,044 -,308 ,759 -,444 ,325

(19)

18 information participants receive about the difficulty of the task. Libby & Rennekamp did not supply this information in the first round to lead people to engage in self-serving attribution. They did supply this information for the second round. However, participants were never able to compare the difficulty of both rounds. Participants were able to compare the difficulty of both rounds in this study and therefore truly knew the difficulty of the second round.

Since almost all variables in both studies are similar except the ambiguity of the task difficulty, it could be argued that this ambiguity led participants to become overconfident. In other words, supplying participants with information about the difficulty of an upcoming task, when this task is similar to one that they already performed and received feedback on, has a debiasing effect. Merely supplying participants with the difficulty levels of the next task does not have a debiasing effect. A participant must be able to compare the difficulty of an upcoming task to one that they have already performed. It could be argued that by comparing the difficulties, a participant knows that the next round will be more or less the same. Since a participant doesn’t learn anything that will help him/her in the next round to perform better, he/she realizes that he/she will probably score the same. These results are sufficient evidence to support the second hypothesis. One is able to more rationally estimate whether one’s second round performance is better than one’s first round

performance when one knows the difficulty of both tasks. Conclusively, it can be assumed that it was not the self-serving attribution that led participants in the study of Libby & Rennekamp to become overconfident, it was the ambiguity of the task.

Post-hoc analysis

As stated, participants were asked to answer a second round of trivia questions. This was necessary to identify true overestimation. A participant could state that he/she thought that the second round’s score would be higher without overestimating him/herself. However, this post-hoc analysis was not deemed necessary since no significant overestimation was found.

Discussion & Conclusion

Limitations

This study was subject to several limitations. Opposite to Libby & Rennekamp (2012), no psychometric tests were conducted. Since overconfidence can be an individual trait, Libby & Rennekamp conducted several psychometric tests that indicated a participant’s disposition and individual traits. This enabled them to control for this variable. However, when they did, no significant relationships were found. Furthermore, it could be argued that this knowledge is not widely applicable since the difficulties of most tasks are not known. Moreover, the difficulty of a task

(20)

19 depends on the person. Two persons might excel at very different tasks. Therefore, the difficulty of one task might be considered easy by one person and simultaneously very hard by the other. This makes it very hard to apply this knowledge in real-life situations by, for example, managers.

Suggestions

Many questions emerged in the time that this study was conducted. For example, it would be interesting to research whether the same results are acquired when the tasks are of different

difficulty levels. Similarly, a variation in the nature of the tasks could also yield interesting results. These results would be more widely applicable since tasks are often not the same in real life. It would also be interesting to know if people still engage in self-serving attributions when they receive misleading feedback. For instance, when a person believes that he/she did not do well in a task but is told that he/she performed great. If this person would attribute this performance internally, then it could be argued that this person’s attributions might have shifted from external to internal. This would stress the importance of correct feedback and show the consequences of incorrect feedback from a different angle than existing scientific literature does.

Conclusion

This study was conducted with two intentions. The first was to confirm that people are subject to the SSAB by replicating the experiment conducted by Libby & Rennekamp (2012). The second intention was to infer a debiasing effect on the participants. The results of the experiment indicated that people do engage in self-serving attribution. When experiencing success, people attribute this success to internal factors such as skill and effort. When experiencing failure, people attribute this to external factors such as difficulty and luck. Furthermore, Libby & Rennekamp claimed to have found a relation between the SSAB and overconfidence. They state that people who experience success and who attribute this internally become overconfident. This study has proven that it was not the self-serving attribution that led people to become overconfident. In Libby & Rennekamp’s study, it was the ambiguity of the task difficulty that led people to become

overconfident. This implies that the SSAB does not cause a person to overestimate himself/herself. It furthermore implies that supplying people with information about the difficulty of a task, when these people already performed a task of similar difficulty, has a debiasing effect. People are thus able to more rationally estimate their confidence in a task when they acquire this information.

(21)

20

Appendix

Figure 1.1

Thank you for your participation in this experiment. The experiment has 2 rounds and begins with a first round of 20 trivia questions from the board game Trivial Pursuit (Triviant). After you are done with the first round, you are kindly asked to fill in four questions concerning some of your beliefs about the first round. To finish the experiment, you are asked to play the second round of 20 trivia questions. This experiment takes approximately 10 minutes to finish. Good luck!

Round 1 (This round has 10 questions from difficulty level ‘easy’ and 10 questions from difficulty level ‘medium’)

1. Where is the smallest bone in the body located?

A Ear B Nose C Finger

2. What is the biggest country with only one time zone?

A Russia B Turkey C China

3. What did the 7 dwarves in the tale of Snow-white do for a job?

A Construction workers B Miners C Fishers 4. Who invented the telephone?

A Alexander Graham Bell B Nikola Tesla C Thomas Alfa Edison 5. In what year did the Spanish civil war end?

A 1939 B 1937 C 1945

6. What is the first letter on a typewriter?

A Z B A C Q

7. ‘The dude’ in the film The Big Lebowski is obsessed by what sport?

A Basketball B Bowling C Golf

8. Which city lies in the northern hemisphere?

A New Delhi B Rio de Janeiro C Kaapstad 9. What is the name of the Hindu hygiene that is based on holistic principles?

A Yoga B Ayurveda C Nirwana

10. What is the difference between ‘Jansen’ and ‘Janssen’ in Kuifje?

A Their headgear B Their mustache C Their cane (wandelstok) 11. Which country does not have a dragon on their flag?

A Wales B Bhutan C Mongolia

12. Which country’s secret service is called Mossad?

A Israel B Afghanistan C Greece

13. If a boat’s speed is 5 knots, how many km/h does this boat travel?

A 9 km/h B 14 km/h C 21 km/h

14. Which of these paintings is not painted by Rembrandt van Rijn?

A De Nachtwacht B Het Melkmeisje C De Staalmeesters 15. What unit of measurement is used to indicate the intensity of earthquakes?

(22)

21

16. What is the capital of Turkey?

A Istanbul B Antalya C Ankara

17. What color cannot be found in the logo of the United Nations (VN)?

A Red B White C Blue

18. Bolivia is a country in which continent?

A Asia B South America C Africa

19. The episodes of which of these TV shows start with The one…, such as The one with the Ring?

A How I met your mother B Friends C The big bang theory 20. How many times faster does sound travel in water than it does in air?

A 2 times as fast B 4 times as fast C 8 times as fast

You will now be informed of the amount of answers you have answered correctly. Please ask the experimenter for this information.

You have answered questions correctly.

Figure 1.2

Please estimate to what extent these factors contributed to your performance:

Skill 0 0 0 0 0 0 0 0 0 Luck

Skill 0 0 0 0 0 0 0 0 0 Difficulty

Effort 0 0 0 0 0 0 0 0 0 Luck

Effort 0 0 0 0 0 0 0 0 0 Difficulty

Please indicate to what extent you agree with the following statement:

“I believe I did well in the first round of trivia”

0 0 0 0 0 0 0 0

0

The difficulty of round 2 is identical to round 1. Both rounds have 10 easy questions and 10 medium questions. Please indicate to what extent you are certain that the

second-round performance exceeds the first-round performance. 0 0 0 0 0 0 0 0 0 Strongly disagree Strongly agree

I’m 100% certain that I would do worse in the next round

I’m 100% certain that I would do better in the next round

(23)

22

Figure 1.3

Please fill in the final round of trivia questions. Good luck and thank you for your participation!

1. Which of these islands is not Italian?

A Corsica B Sicily C Sardinia

2. In which French city is the seat of the head of the European Parliament located?

A Strasbourg B Paris C Lyon

3. How many strings do most violins have?

A 8 B 6 C 4

4. What is the first digit in the sequence of Fibonacci?

A 3 B 0 C 1

5. Which of these countries does not share a border with Slovakia?

A Austria B Hungary C Slovenia

6. Which of these animals was responsible for spreading the black Plague (de Pest) in the dark ages?

A Rats B Bats C Cats

7. The painter Edvard Munch was born in which of these countries?

A Germany B Norway C Austria

8. Which of these gasses is used to fill balloons?

A Helium B Nitrogen C Chlorine

9. What is the smallest county in the world?

A Vatican City B Monaco C San Marino 10. Which of these was built first?

A The Colosseum B The Chinese Wall C Stonehenge 11. Origami, the art of folding paper, is originally from which of these countries?

A Finland B India C Japan

12. Which of these distances is the smallest?

A Millimeter B Picometer C Nanometer 13. Which of these countries does not use the dollar as their currency?

A Canada B Fiji C Cuba

14. What wooden animal was used by the Greek to conquer Troy?

A Horse B A Cow C A Dog

15. How many stars form the constellation (sterrenbeeld) Ursa Major (de Grote Beer)?

(24)

23

16. What was the previous name of the Indian city Mumbai?

A Calcutta B Mandrai C Bombay

17. Which country was the first to introduce the decimal metric system?

A France B Japan C USA

18. What is the symbol of the French royal dynasty?

A An Eagle B A Red Rose C A Lily

19. The shoes in the original tale of Cinderella (Assepoester) were made of what material?

A Velvet B Fur C Diamond

20. At what age do humans start shrinking?

A 50 B 30 C 60 Figure 2.1 Reliability Statistics Cronbach's Alpha Cronbach's Alpha Based on Standardized Items N of Items ,571 ,580 4 Figure 2.2

Inter-Item Correlation Matrix

LUvsSK DIFvsEF DIFvsSK LUvsEF

LUvsSK 1,000 ,130 ,283 ,208 DIFvsEF ,130 1,000 ,379 ,359 DIFvsSK ,283 ,379 1,000 ,180 LUvsEF ,208 ,359 ,180 1,000 Figure 2.3 Item-Total Statistics Scale Mean if Item Deleted Scale Variance if Item Deleted Corrected Item-Total Correlation Squared Multiple Correlation Cronbach's Alpha if Item Deleted LUvsSK 13,68 14,288 ,280 ,106 ,567 DIFvsEF 14,15 14,265 ,404 ,232 ,460 DIFvsSK 13,63 14,948 ,399 ,199 ,469 LUvsEF 13,93 14,165 ,349 ,155 ,503

(25)

24

Figure 3

Figure 4.1

Correlations

MeanATT DIDWELL

Pearson Correlation MeanATT 1,000 ,417

DIDWELL ,417 1,000 Sig. (1-tailed) MeanATT . ,000

DIDWELL ,000 .

N MeanATT 60 60

DIDWELL 60 60

Figure 4.2

ANOVAa

Model Sum of Squares df Mean Square F Sig.

1 Regression 14,363 1 14,363 12,193 ,001b

Residual 68,321 58 1,178

Total 82,683 59

a. Dependent Variable: MeanATT b. Predictors: (Constant), DIDWELL

(26)

25

Figure 4.3

Test of Homogeneity of Variances MeanATT

Levene Statistic df1 df2 Sig.

1,598a 7 51 ,157

a. Groups with only one case are ignored in computing the test of homogeneity of variance for MeanATT.

Figure 5

Test of Homogeneity of Variances CONFIDENCE

Levene Statistic df1 df2 Sig.

1,873a 12 41 ,068

a. Groups with only one case are ignored in computing the test of homogeneity of variance for CONFIDENCE.

References

Camerer, C., & Lovallo, D. (1999). Overconfidence and excess entry: An experimental approach. The

American Economic Review, 89(1), 306-318.

Carter, C. R., Kaufmann, L., & Michel, A. (2007). Behavioral supply management: a taxonomy of judgment and decision-making biases. International Journal of Physical Distribution & Logistics

Management, 37(8), 631–669.

Chaikin, A. L., & Darley, J. M. (1973). Victim or perpetrator?: Defensive attribution of responsibility and the need for order and justice. Journal of Personality and Social Psychology, 25(2), 268–275.

Clapham, S. E., & Schwenk, C. R. (1991). Self‐serving attributions, managerial cognition, and company performance. Strategic Management Journal, 12(3), 219-229.

Davis, F. B. (1964). Educational measurements and their interpretation (p. 24). Belmont, CA: Wadsworth.

Doucouliagos, C. (1994). A note on the evolution of Homo Economicus. Journal of Economic Issues,

(27)

26 Davis, W. L., & Davis, D. E. (1972). Internal-external control and attribution of responsibility for

success and failure. Journal of Personality, 40, 123–136.

Evans, J. S. B. (1989). Bias in human reasoning: Causes and consequences. Lawrence Erlbaum Associates, Inc.

Feather, N. T. (1969). Attribution of responsibility and valence of success and failure in relation to initial confidence and task performance. Journal of Personality and Social Psychology, 13(2), 129–144.

Field, A. (2013). Discovering Statistics using IBM SPSS Statistics. Discovering Statistics Using IBM SPSS

Statistics, 297–321.

Fischhoff, B. (1982). Debiasing. Judgment Under Uncertainty: Heuristics and Biases.

Fitch, G. (1970). Effects of self-esteem, perceived performance, and choice on causal attributions.

Journal of Personality and Social Psychology, 16(2), 311–315.

Giere, R. N. (2004). How Models Are Used to Represent Reality. Philosophy of Science, 71(5), 742– 752.

Gigerenzer, G., & Selten, R. (2002). Bounded rationality: The adaptive toolbox. MIT Press.

Haselton, M. G., Nettle, D., & Andrews, P. W. (2005). The evolution of cognitive bias. The Handbook

of Evolutionary Psychology.

Henrich, J., Boyd, R., Bowles, S., Camerer, C., Fehr, E., Gintis, H., & McElreath, R. (2001).In search of homo economicus: behavioral experiments in 15 small-scale societies. The American

Economic Review, 91(2), 73-78.

Hoelzl, E., & Rustichini, A. (2005). Overconfident: Do you put your money on it? Economic Journal,

115(503), 305–318.

Johnson, T. J., Feigenbaum, R., & Weiby, M. (1964). Some determinants and consequences of teachers’ perception of causation. Journal of Educational Psychology, 55(5), 237–246.

(28)

27 Lambert, J., Bessière, V., & N’Goala, G. (2012). Does expertise influence the impact of overconfidence on judgment, valuation and investment decision? Journal of Economic Psychology, 33(6), 1115– 1128.

Libby, R., & Rennekamp, K. (2012). Self-serving attribution bias, overconfidence, and the issuance of management forecasts. Journal of Accounting Research, 50(1), 197–231.

Lichtenstein, S., & Fischhoff, B. (1977). Do those who know more also know more about how much they know? Organizational Behavior and Human …, 183(3052), 159–183.

Lipman, B. L. (1995). Information Processing and Bounded Rationality : A Survey, 28(1), 42–67.

Mathis, K., & Shannon, D. (2009). Homo Economicus. Efficiency Instead of Justice?, 7-30.

Miller, D. T., & Ross, M. (1975). Self-serving biases in the attribution of causality: Fact or fiction?

Psychological Bulletin, 82(2), 213–225.

Miller, D. T. (1978). What constitutes a self-serving attributional bias? A reply to Bradley. Journal of

Personality and Social Psychology, 36(11), 1221–1223. doi:10.1037/0022-3514.36.11.1221

Moore, D. A., & Healy, P. J. (2008). The trouble with overconfidence. Psychological review, 115(2), 502.

Nunnally, J. C., & Bernstein, I. (1994). Psychometric Theory. Rdsepiucsforg.

Persky, J. (1995). Retrospectives: the ethology of homo economicus. The Journal of Economic

Perspectives, 9(2), 221–231.

Peterson, R. A. (1994). Meta-analysis of Cronbach’ s Coefficient Alpha. Journal of Consumer Research,

21(2), 381–391.

Plous, S. (1993). The psychology of judgement and decision making. McGraw-Hill , Inc.

Oskamp, S. (1965). Overconfidence in case-study judgments. Journal of Consulting Psychology, 29(3), 261–265.

Rosenfield, D., & Stephan, W. G. (1978). Sex differences in attributions for sex-typed tasks. Journal of

(29)

28 Ross, L. (1977). The Intuitive Psychologist And His Shortcomings: Distortions in the Attribution

Process. Advances in Experimental Social Psychology.

Scapinello, K. F. (1989). Enhancing differences in the achievement attributions of high- and low-motivation groups. Journal of Social Psychology, 129(3), 357–363.

Schlenker, B. R., Weigold, M. F., & Hallam, J. R. (1990). Self-serving attributions in social context: Effects of self-esteem and social pressure. Journal of Personality and Social Psychology, 58(5), 855–863.

Sedikides, C., Campbell, W. K., Reeder, G. D., & Elliot, A. J. (1998). The self-serving bias in relational context. Journal of Personality and Social Psychology, 74(2), 378–386.

Shepperd, J., Malone, W., & Sweeny, K. (2008). Exploring Causes of the Self-serving Bias. Social and

Personality Psychology Compass, 2, 895–908.

Simon, H. A. (1955). A behavioral model of rational choice. The quarterly journal of economics, 99-118.

Simon, H. (1979). Rational Decision Making in Business Organizations. The American

Economic Review, 69(4), 493-513.

Simon, H. A. (1982). Models of bounded rationality: Empirically grounded economic

reason (Vol. 3). MIT press.

Streufert, S., & Streufert, S. C. (1969). Effects of conceptual structure, failure, and success on attribution of causality and interpersonal attitudes. Journal of Personality and Social

Psychology, 11(2), 138–147.

Svenson, O. (1981). Are we all less risky and more skillful than our fellow drivers?. Acta

Psychologica, 47(2), 143-148.

Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: a social psychological perspective on mental health. Psychological Bulletin, 103(2), 193–210.

Thaler, R. H. (2000). From homo economicus to homo sapiens. The Journal of Economic

Referenties

GERELATEERDE DOCUMENTEN

Wanneer er op deze manier naar televisieseries wordt gekeken zou er dus een nog sterkere band tussen de kijker en het mediapersonage kunnen ontstaan, met mogelijk sterkere

- Beide partijen moeten het bestaan van de relatie inzien en deze moet ook wederzijds erkend worden. - De relatie gaat verder dan incidenteel contact en wordt met een

The for- mer is discussed in more detail by Beukeboom (2014), who defines linguistic bias as “a systematic asymmetry in word choice as a function of the social category to which

When one estimates a gravity equation using GDP as a proxy for the mass variables, Baldwin and Taglioni (2014) show that the estimate for the mass coefficients are lower when

To learn about the debt behaviour, table 2 shows summary statistics of gross household income and monthly debt service payments for households with and without self-control

Es evidente que prevenir es mejor que remediar, y que aún falta mucho por mejorar en el acceso equitativo a la educación superior chilena; pero por ahora parece

Die response in tabel 4.22, 4.23 en 4.24 dui op die leierskapstyl wat die hoof openbaar tydens die verskillende fases van die bestuursontwikkelingsprogram, naamlik