Responsibility for talent?
Does the type of production, effort- versus talent-related, affect allocation decisions?
Msc Thesis written by:
Milan Braak 11302429
Degree: Master Business Economics, track Neuroeconomics Faculty: Economics and Business
Supervisor: Dr. D. R. Amasino Second reader: Dr. J. Hausfeld Universiteit van Amsterdam
Statement of Originality
I, Milan Braak, take full responsibility for the contents of this document and 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.
1. Introduction ... 5
2. Literature Review ... 7
2.1. Dictator Game ... 7
2.3. Dictator Game with production ... 9
2.4. Talent versus effort ... 10
2.5. Drivers of success ... 11
2.6. Entity versus incremental theory ... 12
2.7. Responsibility ... 13
2.8. Attention ... 13
2.9. Neural fairness systems ... 15
3. Hypotheses ... 16
4. Experimental design ... 18
4.1. Experiment set up... 18
4.2. Production Phases ... 19
4.2.1. Slider task ... 19
4.2.2. Math task ... 20
4.2.3. Pay rate and monetary contribution ... 20
4.3. Dictator games ... 20
4.5. Questionnaire... 22
5. Results ... 23
5.1. Participants ... 23
5.2. Descriptive statistics of production phases ... 23
5.3. Descriptive statistics of offers ... 24
5.4. Production type and allocation differences ... 25
5.5. Production type and attention... 28
6. Discussion ... 32
7. Conclusion ... 34
8. Appendix... 36
A. Additional regression(s) ... 36
B. Welcome screen ... 36
C. Instructions ... 37
D. Production Tasks ... 38
E. Allocation Decisions... 39
F. Questionnaire ... 41
9. Acknowledgements ... 43
10. References... 44
In the past decade, extensive research is done on fairness ideals using the dictator game with production phase. It is shown that most people do not hold each other responsible for luck factors in a production phase. But how do individuals perceive talented people, as deserving or lucky? This research disentangled effort and talent to investigate the role of both production factors in allocation decisions and the attention seeking process. An online experiment in which participants played dictator games with two production phases, effort and talent-related, is conducted. The findings showed that the type of production, effort- versus talent-related, did not play a role in the allocation decisions. In talent-related decisions, people tend to pay more attention to the merit information compared to pure luck information than in effort-related decisions. Furthermore, it is shown that an attention shift from luck information to merit information does not affect the allocation decision.
Inequality of opportunity is still a necessary topic in political discussions. There are several approaches to look at this inequality. Former American president Richard Nixon once said:
“The American dream does not come to those who fall asleep”. This means you are able to rise as far as your talents will let you, provided that you put in hard work and it implies you deserve the benefits you receive. According to this approach, you are responsible for your own successes and failures. This responsibility has a positive and a negative side because it could make someone feel more or less worthy, depending on their position on the social ladder. This critique of talent-based meritocracy is known as the tyranny of merit (Sandel, 2021). According to Sandel (2021), the meritocracy is not as just as is often suggested, as it generates hubris amongst the people on top of the social ladder and humiliation and resentment amongst the people on the lower side. This meritocratic competition separates people into classes and, thus, leads to polarization in the society (Carvalho, 2021). High-income countries, like the Netherlands, are also meritocratic and try to close the inequality gap by means of education.
Education equality should lead to equal opportunities to develop your talents. This implies talent plays an important role in achieving successes and climbing the social ladder. Therefore, it is relevant to study how individuals reward talent compared to pure effort. Does someone whose talents are more developed deserves more in the eyes of others?
Talent is defined as a natural ability (to do something), but the way people perceive talent is diverse. According to Dweck (2000), there are two types of mindsets regarding talent, a fixed mindset and a growth mindset. Someone who has a fixed mindset believes that talent is constant over time and it is not in someone’s power to improve this ability. In contrast, according to the growth mindset, talents can change over time by putting in effort. Differing perspectives over the source of talent can have significant consequences in how one perceives benefits that relate to these talents. Someone who believes talent is constant over time and therefore more due to luck might think that gains related to talent should not only benefit the talented ones and the benefits should be redistributed. Supporters of the growth mindset, on the contrary, can believe that talents arise because of effort and the gains that come with this talent should benefit the talented alone. The fixed and growth mindsets differ substantially from each other, but the expression of talent probably arises from both genes and effort and these are difficult to fully disentangle (Suhay, Klasnja and Rivero, 2021). So, both mindsets are justifiable and may play a role in attitudes toward talent and redistribution.
Many studies have explored attitudes toward fairness more generally. Cappelen, Sørensen and Tungodden (2010), found most people redistribute joint earnings based on the production
of the workers and do not consider external factors in their decisions. Productivity consist of multiple factors among which effort and talent, however in previous research those two components are not disentangled. It seems to be obvious that effort should be rewarded, but how do people perceive talent? Are talented people more deserving or are they just lucky? This is still unclear. This research investigates what people think is fair regarding gains that relate to talent versus effort. In previous research (Konow, 2000; Cappelen et al., 2010), participants play the same tasks as their matched partner. This is done to make a correct comparison between the production of both. Some production phases are more related to talent and others are more related to pure effort, but does this relate to how the allocator distributes the joint earnings?
Previous research to fairness preferences includes attention as an important factor in the decision-making process. Fiedler, Glöckner, Nicklisch, and Dickert (2013), found competitive and prosocial players pay more attention to information related to their partners’ performance and allocate more to their partner. However, the role of information type, effort- versus talent- related, in the attention seeking process, is still unknown.
How do people perceive talented people, as deserving or just lucky? Based on the meritocracy in high-income countries, it can be stated that talented people are more deserving.
However, how do people think about these fairness ideas on a personal level? Do people believe talent should be rewarded? And how does type of information, talent vs. effort, affect the attention seeking process? To investigate this, the research presented here investigates: “How does the type of production, based on talent or effort, influence the choices made in allocation decisions?”
To study this, we conduct an online experiment inspired by Cappelen et al. (2010).
Participants first produce a surplus by completing either a task related to effort or related to talent. Subsequently, participants are matched with a hypothetical partner and allocate the joint earnings between the two players. During the allocation, participants can reveal information about determinants of the joint earnings. Our results indicate that effort- versus talent-related differences in the production phase do not play an essential role in allocation decisions.
Furthermore, it is shown that the type of production influences the attention seeking process.
Participants pay more attention to the production information in the talent-related allocation decisions than in the effort-related allocation decisions. The effect of attention on the allocation is ambiguous. Increased attention of the participant to production information in general, the participant’s production and the recipient’s production, does not affect the allocation. However, an increase in the participant’s attention to own production increases the share allocated to the
participants. Whereas paying relatively more attention to the recipient’s production than own production, increases the share allocated to the recipient.
The remainder of this paper consists of the following sections. Section 2 contains related literature and Section 3 outlines the hypotheses. Section 4 describes the experimental design.
Subsequently, Section 5 gives an overview of the empirical results. Section 6 includes the discussion and Section 7 concludes the paper.
2. Literature Review
2.1. Dictator Game
The dictator game is often used to study fairness ideals. This game gives more information about personal preferences in money division, which can help us understand social norms and fairness perceptions (Krupka & Weber, 2008). The dictator game is an experiment introduced by Kahneman in 1986 (Engel, 2011). In this game, two players are matched with each other.
One of the players is the dictator, and this person has a certain amount of money to split between them self and the other player, the recipient (Engel, 2011). There is no negotiation about the proposal done by the dictator. According to review by Camerer (2003), dictators offer around 15% of the total amount to the recipient. This offer contradicts the 28.3% that, according to Engel (2011), is given to the recipient. In the meta-analysis of Engel (2011), the dictator games include social control features. These features are the most likely reason for the difference between the fraction given to the recipient. On an individual level, one factor that increases the fraction given to the recipient is non-anonymity of the dictator. If the recipient knows who the dictator is, and the dictator is aware of this, the fraction given to the other player increases by a significant percentage (Engel, 2011). A non-anonymous dictator is an example of including social control in the Dictator Game. Another way of including social control can be showing a cue of 3 dots in a watching eye configuration as Rigdon, Ishii, Watabe, and Kitayama (2009) do, finding that 25.42% of the participants in the watching face treatment choose to keep all the money to themselves. By contrast, 40% of the participants choose to keep the whole amount in the neutral configuration treatment. The watching face treatment significantly decreases the number of dictators keeping the whole stake to themselves compared to the neutral configuration treatment. The results mentioned by Rigdon et al. (2009) and Engel (2011) compared to the review by Camerer (2003) show that social norms are an essential factor in the fairness ideals of the dictator and therefore have an effect on the proposals.
The perceived deservingness of a recipient also plays a role in the fraction given to this recipient (Eckel & Grossman, 1996). A deserving recipient increases the average share given
to the other player and decreases the percentage of dictators that keep the whole stake to themselves. Eckel and Grossman (1996) found that a dictator game played with a charity, the American Red Cross, as the recipient leads to an increase in the fraction allocated from 10.6%
to 31%. Moreover, the share of dictators keeping the entire stake to themselves decreases from 62.5% to 27.1% in the experiment done by Eckel and Grossman (1996). Eckel and Grossman (1996) use a design where the dictators know with certainty the recipient is a charity. Kappes, Nussberger, Faber, Kahane, Savulescu and Crockett (2018) investigate how uncertainty affects generosity. Including two types of uncertainty, outcome- and impact uncertainty, in dictator decisions causes contradicting changes in prosocial behavior per type of uncertainty, compared to open information dictator games. Outcome uncertainty decreases prosocial decisions, whereas uncertainty about the impact on other’s well-being increases prosocial decisions. This implies uncertain deservingness increases prosocial behavior. In contrast to Kappes et al.
(2018), who have studied how uncertainty about how the recipient’s well-being is impacted by a decision, Chen, Zhu and Chen (2013) investigated how well-being of the dictator affects prosocial behavior. Chen et al. (2013) found a negative correlation between family income and prosocial decisions, which implies that wealthier children keep more to themselves and low- income children perceive recipients more deserving in general. Four out of five dictators seek information about the deservingness of the recipient (Thunström, Cherry, McEvoy & Shogren, 2016). Thunstrom et al. (2016) found that dictators, who seek for deservingness information, give more to the deserving recipients compared to undeserving.
This study uses the dictator game to investigate the perception towards talent and effort related to fairness ideals. It is investigated how talent and effort affect the deservingness of the recipients and how information seeking plays a role in this process.
2.2. Fairness ideals
There are three main fairness principles discussed in the literature. In the strict egalitarianism principle, people are not held responsible for all factors that play a role in production (Cappelen et al., 2010). Supporters of this principle believe all monetary benefits of country should be equally divided over the citizens. Individuals who use this principle to divide money between themselves and another split the total amount equally. If the money is divided in a meritocratic way, the production of both workers should be considered. Meritocrats believe that people are responsible for all their traits, but are not responsible for external drivers of the production or their success (Bowles, Durlauf & Arrow, 2000). In that case, individuals divide the total stake in percentages of the production contribution and do not consider external factors in the
division. According to the libertarianism principle, people are responsible for all factors that play a role in their success, including production. Besides the trait factors that are of importance to meritocrats, libertarians believe people are responsible for external factors as well.
Supporters of this theory believe people should get paid according to the worth of production, including external factors of production. This means everyone should receive their monetary contribution to society.
2.3. Dictator Game with production
The production phase is added to make the participants more responsible for the value they distribute, which makes it more realistic compared to a random endowment (Cappelen et al., 2010). In a dictator game with production, a production phase precedes the dictator decision.
Cappelen et al. (2010) suggest the dictator game with production can investigate for which drivers of the produced stake people hold their matched players responsible in the allocation decision. In their experiment, the participants had to rewrite a text on the computer. The value each participant contributed to the pie was the task performance, words rewritten correctly, and a randomly-assigned high or low pay rate per correct word. The participants worked either 10 minutes or 30 minutes. After the production phase, every participant was matched with other participants and split the total stake between the two of them. An important finding is that dictators, on average, give 38.5% of the jointly-produced stake to the recipient. This fraction is much higher than the average of 15% that dictators give in the standard dictator game (Camerer, 2003). This difference could imply that people decide based on their fairness ideals rather than pure profit maximization or social norms in the dictator games with a production phase, since the endowment in these dictator games is not random. However, due to production of the participants, the dictator is incentivized to base the decision on what both participants deserve rather than on social norms or altruistic behavior. Konow (2000) found that participants in his dictator game with production experiment made decisions in line with the results of Cappelen et al. (2010). Konow (2000) found allocation by a third party increases the equality of division.
This is not researched in the study by Cappelen et al. (2010). Furthermore, Capellen et al. (2010) shows that 47% of the participants base their decisions on in-control factors, working time, and correct answers. In-control factors are subjected to choices and productivity. These participants do not consider the randomly assigned pay rate, out-control factor, in the allocation decision.
This is in line with meritocratic principles (Bowles, Durlauf & Arrow, 2000). As shown in the research by Cappelen et al. (2010). In this research, an extra element is included to the research to investigate fairness ideals. In the research done by Cappelen et al. (2010), all participants
played the same task and made dictator game decisions related to this task. Therefore, the results could identify the fairness ideals of the participants related to that specific task. In this research, two different production phases will be compared to see how the different types of production influence the allocation decisions. The production phases either relate to pure effort or talent.
Adding a talent-related production phase to the design, gives the opportunity to investigate talent vs. effort in allocation decisions.
The percentage of the total stake given to the recipient differs per individual, and they even differ per dictator game. Rodriguez-Lara and Moreno-Garrido (2011) explored these differences by categorized participants’ fairness types based on their divisions in a dictator game with production. In their categorization, an egalitarian participant would always send half of the total contribution to the recipient. Meritocrats would divide the total stake based on the production effort of both workers, and libertarians make their allocations decisions based on monetary contributions, including randomly-assigned pay rate. However, Rodriguez-Lara and Moreno-Garrido (2011) state that different principles may be chosen, depending on the worker’s specific production and pay rate, to yield the highest outcome for themselves. This research gives more information about why people make choices and how choices change when the perspective of a decision changes. Being in a more favorable position, getting assigned the high pay rate, increases the chance the dictator allocates the joint-produced stake according to the libertarian principles. Whereas a less favorable position, low-pay rate treatment, decreases the chance of allocation based on libertarianism.
Production phases based on two different aspects, effort and talent, gives new insights compared to the previous studied dictator games with production phase. Cappelen et al. (2010) show that in-control factors are commonly used in the allocation decision. This research will investigate the extent to which effort and talent are seen as equivalent in-control factors.
2.4. Talent versus effort
In this research, both dispositional causes—talent (intelligence) and effort—were tested against each other. According to Lockhart, Keil and Aw (2013), children and adults preferred to interact with persons that always had a likable trait over someone who learned this skill. This is called the natural bias, in which people believe that acquired traits vanish over time, whereas natural traits are permanent. Another study found that honors college students believe that their talent was a more critical factor in their successes than their effort (Siegle, Rubenstein, Pollard &
Romey, 2009). This could imply that these students support the fixed mindset theory, in which natural-born talents are fixed over time, and effort has less influence on performance. Fairness
ideals concerning this finding will be tested in this research to investigate if natural ability should be rewarded according to the participants.
However, Leung, Kim and Tse (2020) found that firms that attribute their service employees’ competent performance to dedicated effort over natural talent are more likely to enjoy word-of-mouth advertisement. Moreover, consumers expect a communal-oriented relationship over an exchange-oriented relationship when a firm attributes the competence of its employees to committed effort (Leung et al., 2020). This is contradicting with the results that are found in the articles as mentioned earlier (Lockhart et al. 2013; Siegle et al., 2009).
According to Brown, Troy, Jobson and Link (2018), the context is an essential factor of the preference for talented or strivers (those who put in the dedicated effort). Brown et al. (2018) state that a person’s own experience leads to what is preferred, a natural or a striver. If the rater believes that putting in effort is more important in a specific context, the striver is preferred over the natural and vice versa. In our case, this implies participants born as talented mathematicians, are likely to believe talent is a more important factor in being a good mathematician compared to effort. How this affects dictator proposals is still unclear.
2.5. Drivers of success
Since this study is focused on distinguishing fairness norms around talent vs. effort, it is essential to know what the drivers of success are and how people view talent versus effort in general. Suhay et al. (2021), investigated how very affluent people look at drivers of success compared to ordinary people. They distinguished two causes of success, dispositional and situational. Dispositional causes are related to intelligence and hard work, and the situational causes correspond with family background and luck in life. Suhay et al. (2021) found that highly affluent people, the wealthiest 1% of American citizens, believe that dispositional causes of success are more relevant compared to less wealthy Americans. This finding means that intelligence and hard work are even more important according to the top 1% than to the rest of the citizens. This is in line with the previous findings of Kluegel and Smith (1986) who found that people in higher social classes believe individualistic causes of success are more important than structural or situational causes. Overall, more successful or more wealthy people believe that personal factors are more important drivers of success than external factors. This implies that less wealthy people attach more importance to situational factors compared to wealthy people.
The way that drive for success (effort) and IQ (talent) comes about according to the different social classes is also investigated by Suhay et al. (2021). According to their article,
choice, genes and environment are more important to the more affluent people than to the less affluent. In particular, the choices a person makes are more important to the affluent people than the rest. This again is in line with previous research (Kleugel and Smith, 1986).
Interestingly, all groups in this experiment think that environmental factors are the most important for IQ and drive to succeed (Suhay et al., 2021). Overall, previous research has shown that talent and effort are closely related in terms being successful. This research disentangles these two components of production.
Furthermore, it has been researched that luck in life is an essential indicator of success, which is generally underestimated (Pluchino, Biondo & Rapisarda, 2018). This is contradicting to the result of Suhay et al. (2021), who found people believe environmental factors are the most important ingredients of IQ and drive to succeed. The importance of the factors is measured in a survey apart from a associated success. Someone’s success, in contrast, is always measured afterward; this could lead to a misinterpretation of how this success came about. An explanation for success often has to do with merit, but this is an overestimated cause of success (Pluchino et al., 2018). Pluchino et al. (2018) stated that luck is a more critical factor for success.
2.6. Entity versus incremental theory
There are different theories of how ability and effort are related to each other. According to the entity theory, your intelligence is fixed over time. People who believe attributes like intelligence are fixed over time push off their achievements and failures to an external factor they cannot influence (Dweck, Chiu & Hong, 1995). This implies that intelligence is not changeable and effort cannot increase someone’s abilities. On the contrary, people who support the incremental theory believe that achievements and failures are due to their (lack of) effort and that intelligence is changeable over time. The incremental way of thinking makes you more responsible for your abilities since it is in your power to increase them. Dweck (1986) stated that students with a growth mindset, supporters of the incremental theory, outperformed students with a fixed mindset, which is an advantage of the incremental theory. However, less talented supporters of the growth mindset could reach their proverbial ceiling, while others advance to the higher ranks on the social ladder. This could cause the hubris among higher ranks of the social ladder and humiliation and resentment among the citizens at the bottom of the social ladder (Sandel, 2021), which is a disadvantage of the incremental theory.
However, how do these theories relate to the research question? Intuitively, according to the entity theory, people who perform better in a talent task are not responsible for their (good
or bad) performance. The talented should compensate people with less talent or talent should not be rewarded in the first place. Therefore, the allocations in an effort task will be more in line with the production contributions of the workers. However, looking at the incremental theory, people are responsible for their talents. This responsibility implies that talented people deserve higher earnings related to their talent. Therefore, the allocations in the talent-related task will be more in line with the production contributions of the workers. Thus, the mindset of the participants, fixed or growth, could play a critical role in the decision-making process.
Responsibility measures for which factors individuals should be rewarded. According to Cappelen et al. (2010), most of the participants in their study choose to share the total stake in a meritocratic way. This means that the participants hold each other responsible for their production and their working time, but for luck components. In research done by Schildberg- Hörisch, Trieu and Willrodt (2020), affirmative action policies were used to investigate what policies were perceived as fair. In this way, Schildberg-Hörisch et al. (2020) could investigate whether individuals hold each other responsible for bad luck, talent or effort. All three affirmative action policies were considered fairer than the control treatment, with no affirmative action policy. According to Schildberg-Hörisch et al. (2020), the compensation for the participants in the bad luck treatment is seen as the fairest compensation. This implies that the participants believe that it is unfair to hold someone responsible for out-control factors, which is in line with the meritocratic principle. However, the compensation for working time is evaluated as fairer than the compensation for low productivity. This strokes with the principles of meritocrats. Since working time is in control of the individuals and the productivity differs among participants due to talent. This implies that the participants hold each other less responsible for an in-control factor, working time, than for an out-control factor, productivity.
Attention plays a vital role in the decision-making process and plays an important role in this research. Krauzlis, Wang, Yu, and Katz (2021) define attention as “the set of evolved brain processes that leads to adaptive and effective behavioral selection." One of the best ways to measure attention is eye-tracking, which is done by measuring how long and where the eye gaze is focused on. Eye-tracking gives precise information about a subject's attention; this could give researchers more information about the decision-making process. In our case, the adaptive
and effective behavioral selection is measured in time participants look at information, which from now on will be referred to as attention. One model that is often used to explain the link between the eye gaze and choice behavior, is the attentional-drift-diffusion model (aDDM) (Krajbich, Armel & Rangel, 2010). The aDMM that predicts the outcome of a binary or trinary choice, based on visual attention paid to the options. According to Krajbich et al. (2010), attention can be used to predict the decisions of an individual. In this research Krajbich et al.
(2010) measured the attention of the participants during the binary choice between two food items, this measurement strongly predicted the choice of the participants. Ghaffari and Fiedler (2018) argue that paying attention to a specific option increases the likelihood of choosing this option. In the research by Fiedler et al. (2013), participants made allocations decisions while their eye-gaze was measured. The allocations consisted of binary choices of money division.
Fiedler et al. (2013) found that prosocial players and competitive players pay more attention information related to the outcomes of others and allocate more to their partner. In contrast, players that wanted to maximize their profit did pay less attention to this information and focused on their own. This is in line with the research by Bieleke, Dohmen and Gollwitzer (2020).
Visual attention involved different types of attention among which endogenous and exogenous (Orquin & Mueller-Loose, 2013). Endogenous attention is usually defined as goal- driven attention, whereas exogenous attention is defined as stimulus driven attention. In this research, we focus on endogenous attention, to ensure participants make decisions on their goal- driven attention. This should increase the chance participants base their allocation on well- considered choice.
In the research done by Amasino, Pace and van der Weele (2021), two participants generated money, after which one was assigned to divide the total joint-earnings. In one of the attention treatments, the dictator was forced to pay attention to the production information of both players. It was found that participants who were forced to focus on merit related information information were more likely to allocate the money based on the production of both players.
This again implies that attention is an essential factor in the decision-making process. Amasino et al. (2021) used MouselabWEB (Willemsen and Johnson, 2019) to measure the attention of their participants. MouselabWEB records the time spent on boxes that reveal information; this shows the researcher what the participant is paying attention to. MouselabWEB is also used in this research to track the attention of the participants.
2.9. Neural fairness systems
It is important to understand how diverse fairness ideals lead to different allocation decisions and increasingly research is done on this topic. More research, however, is needed to the underlying neural mechanisms involved in fairness decisions. Cappelen, Eichele, Hugdahl, Specht, Sorensen, and Tungodden (2014), claim to be the first researchers investigating neural responses to income inequality in situations in which individuals contributed to the income in terms of effort. In their experiment participants produced a surplus by providing correct responses to a pure effort task. The only difference per participant was the working time. All participants initially received the same hourly pay of 500 NOK (47,91 Euro). In the scanning phase after the production phase, participants were matched in duos and possible distributions of the total joint-earnings were shown. The working time of both workers was common knowledge, based on that participants indicated whether they rated the proposed distributions.
Cappelen et al. (2014) found that a marginal increase in proposed income led to a stronger neural response among participants who worked longer. It is shown that proposals related to production causes neural activity related to production proportions. This is in line with the behavioral results of Cappelen et al. (2010), since most participants base their allocation on the production.
The investigated brain regions in the study by Cappelen et al. (2014) are the left and right caudate nucleus, which are part of the striatum. The striatum is an important component of the reward system (Delgado, 2007). This system is activated when a reward in all kinds of forms is received. Other important brain regions involved in this system are the prefrontal cortex (Thut, Schultz, Roelcke, Nienhusmeier, Missimer, Maguire & Leenders, 1997) and the ventral tegmental area (D'Ardenne, McClure, Nystrom & Cohen, 2008). Those regions, however, show no activation compared to control trials of the experiment (Cappelen et al., 2014). This implies those regions are not involved in the underlying neural systems of fairness ideals. Besides the reward system norm compliance is involved in allocation decisions as well.
The prefrontal cortex is the main region involved in this system (Ruff, Ugazio & Fehr, 2013).
This is in line with previous research showing that the prefrontal cortex is important in social behavior (Roberts, Robbins & Weiskrantz, 1998). Spritzer et al. (2007) conducted a research studying neural activity related to norm compliance in dictator games with punishment, in which recipient could punish the dictator. Spritzer et al. (2007) found that prefrontal cortex activation increased in the punishment treatment compared to the no punishment treatment.
Schaefer, Kühnel, Rumpel and Gärtner (2021) found that prosocial decisions in the dictator games correlated with activation in the right temporoparietal junction. This part of the brain is
involved in analyzing self-produced actions as well as actions of others (Decety, & Lamm, 2007).
This research does not investigate neural activation. Future research, however, could build on the findings of Schaefer et al. (2021) and Spritzer et al. (2007). It could be tested whether the right temporoparietal junction shows more activation in effort or talent related allocation decisions. Furthermore, fMRI analyses could lead to more insights in the role of the prefrontal cortex in social decision-decision making and fairness perceptions.
2.10. Literature gap
Fairness ideals have been studied extensively, still the role of talent vs. effort is vague in this context. Several researchers (Eckel & Grossman, 1996; Kappes et al., 2018; Thunström, 2016) found deservingness is ramified with allocation decisions. Still unclear is the role of talent and effort in deservingness. Talent and effort lump together in studies investigating drivers of success and responsibility (Suhay et al., 2021; Cappelen et al., 2010), since no clear distinction between the importance of both factors is made. Dweck (1995) supported the idea that effort drives talent. This would imply talent should be rewarded, but should talent be rewarded more than pure effort?
This paper investigates this question by disentangling talent and effort in a dictator game with production phase. Furthermore, it is widely studied how attention impacts the decision- making process (Fiedler et al., 2013; Amasino et al., 2021; Krauzlis et al., 2021). The aspects of talent vs. effort in attention seeking are not clear. This research investigates how talent and effort relate to each other in the attention seeking process and how attention drives the allocation decision.
This section outlines three hypotheses. These hypotheses are based on previous literature and intuition because of the lack of literature on that topic.
The first hypothesis answers the main research question: “How does the type of production, based on talent or effort, influence the choices made in allocation decisions?”. We investigate how allocations change if the type of production changes, effortful versus talent-related. Since there is very little literature about this topic, the hypothesis is based on intuition and personal interpretation of literature that tangentially relates to this topic. The expectation is that dictators allocate the joint earnings more in line with the production proportions of the workers in the talent task than in the effortful task. This hypothesis is based on the idea that participants believe
that talent, in this case, their mathematical skills, needs to be rewarded. According to Schildberg-Hörisch et al. (2020), participants hold each other less responsible for an in-control factor, namely working time, than talent which may have a luck component. Those results could imply that the participants hold each other more responsible for their ability than for effort in this study. The decisions could also be subject to the vision regarding talent growth. Supporters of the incremental theory (Dweck et al., 1995), might believe talent should be rewarded more since it relates to previous efforts and current effort. Therefore, participants will base their allocations more on production in the talent-related decisions than in the effort-related decisions.
Hypothesis 1. In the dictator decisions based on the talent task, the dictators offer a proportion to the recipient more in line with the production contributions than in the dictator decisions based on the effortful task.
The second hypothesis is related to the participant's attention and especially to the role of talent and effort in the attention seeking process. Fiedler et al. (2013) and Amasino et al.
(2021), found attention to a specific option increases the chance of choosing this option. The results by Fiedler et al. (2013) and Amasino et al. (2021) imply that attention in general drives choice. However, it is unclear how attention is driven by the type of information, talent- or effort related. The talent task production consists of two factors, both talent and effort, and the effort task only relates to effort. Therefore, it is expected participants pay more attention to production information compared to monetary contributions in talent-related allocation decisions. This would be in line with the incremental theory (Dweck et al., 1995). Since effort drives talent, production information in talent-related allocation decisions reveals information about current effort and previous efforts. Whereas, production information in effort-related allocation decisions only shows the current level of effort. Another reason for the increase in attention to production information in the talent-related allocation decisions could be that people are simply more curious about the production in the talent task compared to the effort task, because the talent task is more demanding. Thus, it is expected participants pay more attention to production information in talent-related allocation decisions than in effort-related allocation decisions.
Hypothesis 2. Dictators pay more attention to production information compared to monetary contribution information in talent-related allocation decisions than in effort-related allocation decisions.
The third hypothesis is related to the participant's attention and especially to the attention related to the production. It is expected that participants who pay more attention to the production information compared to the monetary contribution information will base their allocation decisions more on the production. This is in line with Ghaffari and Fiedler (2018) and Amasino et al. (2021), who stated that paying attention to a specific option increases the chance of choosing this option. The results by Ghaffari and Fiedler (2018) and Amasino et al.
(2021) imply that attention in general drives choice; it is expected that attention to production contributions results in an allocation more in line with production.
Hypothesis 3. An attention shift from monetary contribution information to production information leads to an allocation more in line with production proportions.
4. Experimental design
4.1. Experiment set up
Every participant received a participation number at the beginning of the experiment to ensure that their answers were stored anonymously. The experiment started with a welcome page and is followed up by the instruction page (see Appendix C). This experiment consisted of 5 phases:
a production phase, allocation decisions, a production phase, allocation decisions, and a questionnaire. The entire experiment is conducted via MouselabWeb (Willemsen & Johnson, 2019). Participants read the instructions and answered one comprehension question about the upcoming production phase (see Appendix C). The comprehension question needed to be answered correctly to proceed to the task. During the production phase, the participants completed a task, which is either related to effort or talent. In these tasks, the participants were instructed to give as many correct answers as possible. After the participants finished two rounds of the task, they made six allocation decisions related to the task. As soon as all six decisions were made, the participants proceeded to the second production phase. Again, this task was repeated twice, after which participants made six allocation decisions related to the last task. When both production phases and allocation decision phases were completed, the participants filled in a questionnaire. All individual parts of the experiment will be explained in the next section.
There were two possible orders in which participants participate in this experiment. In situation 1, the participants first engaged in an effort task during the production phase, and made allocation decisions based on the effort task. As soon as the participants finished the first set of allocation decisions, they started with the second production phase. In the second production phase of situation 1, participants played the talent task (figure 1a). In situation 2, the participants played the talent task in the first production phase and the effort task in the second production phase, (figure 1b). This order, situation 1 or situation 2, was randomly chosen to reduce order biases.
Figure 1a and 1b: Orders of experiment
4.2. Production Phases
We chose this within -subject design, because it disentangled our variables of interest, talent and effort. The within-subject design measured the effect of task type on the level of an individual.
4.2.1. Slider task
During the slider task, participants were tasked with setting sliders to predetermined locations.
The sliders range from 0 to 200, and the predetermined location was a random value between those values, different for each slider. In total, there were 40 sliders; this number of sliders was chosen to ensure that the participants were not able to move all the sliders to the correct location in time. This was done to measure the maximum production of the participant and made sure that the participant putted in effort the entire timespan of the task. This task was chosen because it is previously used as real effort task (Amasino et al., 2021) and it is not talent related.
Therefore, the slider task was used to effort task1.
1 External factors, for example the use of a mousepad instead of physical mouse, could interfere production.
These factors are not related to talent and, therefore, we do not consider these external factors as an effort measurement threat.
4.2.2. Math task
During the math task, participants solved as many calculations as possible. The calculations consisted of the following combinations: multiplication or a division with a summation, and the answers were always integers. In total, there were 30 calculations. This number of calculations was chosen to ensure that the participants were not able to solve all the calculations in time.
Which was done to measure the maximum production of the participant and make sure that the participants put in effort the entire timespan of the task. This task was chosen because it is related to talent and not only based on effort, since the calculations required mathematical skills.
Mathematical skills development differs based on giftedness (Denton & West, 2002).
Therefore, the calculation task was chosen to distinguish the talent task from the pure effort task.
4.2.3. Pay rate and monetary contribution
Participants with different pay rates were included to check whether participants hold another responsible for pure luck features of the production. The high pay rate was €1,50 per correct answer and the low pay rate was €0,50 per correct answer. These pay rates were chosen due to keep payments reasonable and because of evidence for a self-serving bias related to differences in pay rates (Amasino et al., 2021). The pay rate was the between subject component of this design. All recipients in the dictator games were assigned the opposite pay rate of the participant. The monetary contribution of the participant and the partner was calculated by multiplying the correct answers of a round with the pay rate. The monetary contributions of the participant and the hypothetical matched player were added together to determine the joint earnings.
4.3. Dictator games
The participant completed two rounds of one of the two tasks, and made six dictator decisions related to the most recent task. Three decisions related to the production of the first round and three related to the production of the second. During the dictator games, participants saw their pay rate and could reveal information about their production, the production of their hypothetical matched partner, their monetary contribution, and their partners’ monetary contribution. This information was revealed when the participant hovered the mouse over an information box (see figure 2). The participants could take as much time they need to look at the available information and made their allocation decision using a slider. The participants
could only confirm their decision if they had moved the slider at least one time to ensure that the participants did not accidentally skip a decision.
Figure 2 Decision screen
Notes: The confirm button is not available, because the decision slider is not moved.
These dictator games consisted of six production combinations of the participant and the matched hypothetical player, denoted as other (see Table 1). The strategy method was used to represent all possible combinations of production rate differences between two workers.
Furthermore, it could be studied if participants make different decisions when confronted with a slower or faster worker. In this research, the slower worker had a production that equals 60%
of the participant’s production. The faster worker had a production that is equal to 140% of the production of the participant. These rates were chosen to ensure substantial production differences.
Combination 1 2 3 4 5 6
Participant x x x y y y
Other x*0,6 x x*1,4 y*0,6 y y*1,4
Table 1 production combinations
Notes: x denotes the participant’s production in the first round of the task and y the production of the participant in the second round of the task.
The participant was confronted with the dictator games in a randomized order. This was done for two reasons. Firstly, since the participants were unaware of their production because of randomization, the box that revealed their information is still of interest. If the production was the same in every round, participants potentially skipped this information since they were already aware of their production. This design, therefore, ensured that participants paid attention to the information that was of importance to them. In this way, the information to which participants paid attention is more valuable. This gave more insights into the allocation decision. Secondly, randomizing the combination order created a more realistic situation;
therefore, this design was chosen.
4.4. Incentive structure
Among the participants, four were randomly chosen and matched in duos. The individuals in these duos got paid based on one randomly chosen decision made by one of the individuals of each duo. This incentive scheme was chosen due to financial reasons and because the researcher still wanted to include an incentive to make realistic decisions.
The last part of the experiment consisted of a questionnaire in which participants answered questions related to fairness, responsibility, and general demographics. All questions can be found in Appendix F.
In this research it was investigated whether offers are more in line with production under certain conditions. We used the share allocated to the recipient to measure whether an allocation is more in line with the production contributions. This is done to make the variable more intuitive.
According to previous research (Cappelen et al., 2011; Konow, 2000; Rodriguez-Lara and Moreno-Garrido 2011) and our own findings, dictators in general allocated more to themselves than to the recipient. Therefore, we stated that an increase in the share allocated to the recipient, implies an allocation that is more in line with production.
In the regressions in this paper several variables were added to control for factors that could be involved in the decision-making process. In table 2, these variables are introduced.
Furthermore, the questionnaire answers were added in some regression to control for background variables, this is denoted by a “yes” after “control” at the bottom of the regressions.
Variable name List of variables
Allocation % Allocated to recipient
Talent Dummy for decisions based on talent task
No attention Dummy for decisions where no information box was opened Production other % production of the recipient
High pay rate Dummy for decisions where participant had a high pay rate Participant production Correct answers of the participant in related production task
% Attention to X Female
Talent reward Effort reward
% Attention to information box(es) X
Dummy for decisions where participants was a female Rating of talent should be rewarded statement
Rating of effort should be rewarded statement Table 2 Variable names
In total, 129 participants participated in the experiment. After removing 10 participants, who dropped out due to the non-completion of the experiment, we analyzed a total of 119 participants. The participants were allocated either to the high or low pay rate treatment. Table 3 summarizes the characteristics of the participants per treatment. In total, 71 males participated and are equally divided over the treatments. The average age of the participants was 27. The demographics characteristics do not differ per pay rate treatment, so the results are comparable.
Pay rate N Males (%) Age (SD)
Low pay rate 62 35 (56.45) 28.71 (±11.85)
High pay rate 57 36 (63.16) 25.77 (±8.26)
Total 119 71 (59.66) 27.30 (±10.35)
Table 3 Descriptive statistics participants per treatment 5.2. Descriptive statistics of production phases
In table 4, a summary of the production phases is given per production type2. Participants on average produced significantly less in the effort task than in the talent task, respectively 10.44 (SD = 2.77) and 12.22 (4.51), t (393) = -5.18; p < 0.001. This means more calculations were solved correctly than sliders moved to the right place. Table 4 also shows that the standard deviation for the production in the calculations task was 4.51, and in the slider task, the standard deviation equaled 2.77. These standard deviations were used to test whether the production in the talent task was more distributed than the production in the effort task. The standard deviation of the production in the calculations task was higher than the standard
2 The amount of production rounds equals the number of allocation decisions divided by three because every production round was used in three different allocation decisions.
deviation in the slider task at a 1% significance level (F (1,474) = 26.81; p < 0.001). We can conclude that the production was more distributed in the talent task than in the effort task, which implies mathematical skills different among participant. Thus, this test justifies our choice for the calculations task and slider task to investigate talent vs. effort in the production phases.
Task type N Correct
Effort task 238 10.44 (2.77) 1 18
Talent task 238 12.22 (4.51) 1 24
Total 476 11.33 (3.84) 1 24
Table 4 Production characteristics
5.3. Descriptive statistics of offers
In total 1428 allocation decisions are made, in 19.19%3 of the allocation decisions no information was revealed. These allocation decisions were used in the analysis, but robustness checks are done to test their impact on the hypotheses.
Table 5 of average allocation decisions
Table 5 summarizes the characteristics of the average allocation decisions per task type.
Participants allocated 43.92% of the joint earnings to the recipient in the talent task and 45.17%
in the effort task. Furthermore, table 5 shows that the production and monetary contribution were close to 50%, aligning with the experiment setup because the slower and faster worker treatments cancel each other out. The minor deviations from 50% in the production and contributions were due to rounding and, in the contributions, the higher percentage of
3 In 274 of the 1428 allocations participants did not reveal any information about the Characteristics production or monetary contribution of themselves or of the recipient.
N Allocation, (% allocated to recipient)
Production, (% produced recipient)
Contribution, (% contribution recipient)
Share offering nothing
119 0.4517 (0.1507)
119 0.4392 (0.1540)
Total 119 0.4454 (0.1473)
participants in the low pay rate treatment. Table 5 shows that the entire stake was kept by the participants in 5.11% of the 1428 allocations decisions. There were 6.02% allocation decisions in the talent task where nothing was allocated to the recipient and 4.20% in the effort task.
Lastly, table 5 shows the share allocated to the recipient was more than half of the joint earnings in 31.51% of the decisions related the effort task and in 28.99% decisions in the talent treatment.
5.4. Production type and allocation differences
The first hypothesis related to the type of production, based on talent or effort, and the allocation to the recipient. On average, dictators allocated more to themselves (see table 5). An increase in the share allocated to the recipient, therefore, implied an allocation more in line with the production. It was hypothesized that offers were more in line with production in the talent task than in the effort task. This implied a higher allocation to the recipient in the talent task than in the effort task.
Figure 3 Average production and allocation
Notes: The error bars represent the standard deviations.
Figure 3 shows the averages per treatment of the recipient’s production and the share allocated to the recipient. It is shown that the production did not differ per treatment, which was due to the set-up of the study, as previously mentioned. Furthermore, this figure shows us that the share allocated to the recipient was smaller in the decisions based on the talent task (M = 0.4392; SD = 0.1540) than the decisions based on the effort task (M= 0.4517; SD = 0.1507).
This difference of 1.24% was significant (t (118) = 1.75; p = 0.0415). According to our hypothesis, the allocation in the talent-related decisions should be higher than the allocation in the effort-related decisions. However, this figure shows us the allocation was higher in the talent-related decisions. Hypothesis 1 cannot be confirmed based on these statistics, since allocations in the effort-related task are higher, thus, more in line with production.
To control for more variables, two regression analyses were done—the regressions controlled for multiple variables that could be involved in the decision-making process of our participants. The allocation to the recipient, was the dependent variable. An increase of this variable, implied the total stake is divided more according to the production contributions. In the first regression, the talent dummy was added to investigate the allocation differences per treatment, effort versus talent. Also, the dummy related to attention was added to the first regression; this was done to check whether not paying attention to any of the information boxes influenced the participants' decisions. Lastly, the percentage production of the recipient was included to control for production differences between the dictator and the recipient. In regression 2, the control variables high pay rate and participant production quantity were added to the first regression. The pay rate controlled for monetary contribution differences due to luck.
Participant production quantity controlled for absolute production differences among the participants and their effect on allocations. Regression 2 also controlled for background variables.
Hypothesis 1 investigated whether participants base their allocations decisions more on production proportions in talent-related allocation decisions than in effort-related allocation decisions. According to regression 1 in table 6, participants allocated less to the recipient in the talent-related allocation decision than in effort-related decisions, which implied allocations were less in line with production proportions in the talent-related decisions. In regression 1, the talent coefficient is -0.011. This coefficient implied that participants allocated 1.1 percentage point less to the recipient in allocation related to the talent task, though not significant (t (118)
= -1.57, p = 0.120). In regression 2, the coefficient was hard to interpret since it was not close to significant. However, the sign of the coefficient was negative, which was not in line with our hypothesis. Using these regressions and the graph in this section, we concluded that talent- related allocations were not more in line with production than effort related allocation decisions.
Thus, hypothesis 1 could not be confirmed based on these statistics.
The first regression in table 6 controls for no attention to the information boxes.
According to the coefficient of this variable, participants who did not reveal information during an allocation allocated 3.1 percentage points less to the recipient. Although, this coefficient is
not significant (ß=-0.031; t (118) = -1.31; p = 0.193). The coefficient of production other shows us that an increase of 1% of the production of the recipient resulted in an increase of 0.573 percentage point in allocation to the recipient. This finding was highly significant (ß=0.573; t (118) = 10.86; p < 0.000).
Y = Allocation Regression 1 Y
Regression 2 Y
X Marginal Effect (SE) p-value Marginal Effect (SE) p-value
Talent -0.011 (0.007) 0.120 -0.0044 (0.009) 0.605
No attention -0.031 (0.023) 0.193 -0.040 (0.023) 0.072 Production other 0.573 (0.052) 0.000 0.571 (0.053) 0.000
High pay rate -0.125 (0.022) 0.000
-0.004 (0.003) 0.153
Female 0.052 (0.023) 0.025
Constant 0.179 (0.028) 0.000 0.261 (0.154) 0.093
Control No Yes
N 1428 1428
𝑅2 0.0774 0.2476
Table 6 OLS Regression with clustered errors: Allocations
Notes: Talent = 1 if the is decision is based on the talent task, Talent = 0 if decision is based on effort task; No attention = 1 if no attention was paid to any of the information boxes, No attention = 0 otherwise; Production other denotes the percentage produced by the recipient; High pay rate = 1 if the high pay rate was assigned to the participant, High pay rate = 0 otherwise; participant production equals absolute production of the participant in an allocation decisions; Female = 1 if participant is a female; Female = 0 otherwise.
In regression 2 in table 6 the no attention variable has a coefficient of -0.040, which means that participants on average allocated 4.0 percentage point less to the recipient if no information was revealed (ß=-0.040; t (118) = -1.82; p = 0.072). An 1% increase of the recipients’ production resulted in an 0.571 percentage point increase of the allocation to the recipient (ß=0.571; t (118) = 10.84; p < 0.000). Furthermore, it is shown that participants in the high pay rate treatment allocated 12.5 percentage points more to themselves than participants in the low pay rate treatment (ß=-0.125; t (118) = -5.54; p < 0.000). A one-unit increase of correct answers in the allocation decision related production phase resulted in a decrease of 0.4 percentage point in the allocation to the recipient, though not significant (ß=-0.0037; t (118) =- 1.44; p = 0.153). Lastly, it is shown that females allocated 5.2 percentage points more to the recipient than males (ß=0.052; t (118) =2.27; p = 0.025).
5.5. Production type and attention
Hypotheses 2 related to the production type and the attention of the participants during the allocation decisions. In table 7, a summary of the average relative attention during the
allocation decisions is given. All attention measures used in this research were relative to total time spent on opening information boxes.
Variable N %Time spent on box
Your production 119 24.59 (7.77) 0 100
Other Production 119 23.27 (11.82) 0 100
Your contribution 119 26.36 (13.41) 0 100
Other contribution 119 25.78 (10.33) 0 100
Table 7 Relative attention
Notes: Time spent on box, minimum and maximum are all shown in percentages. Time spent on box represents the time participants spent on average on a certain information box during an allocation decision. Your production and other production are the boxes where the production of the participant and the matched partner are shown, respectively. Your contribution and other contribution are the boxes where the monetary contribution of the participant and the matched partner are shown, respectively.
It was hypothesized that participants pay more attention to production information compared to monetary contribution information in the talent-related allocation decisions than in the effort-related decisions. This hypothesis is tested using all the allocation decisions where the participants opened at least one information box4. For each allocation decision, we calculated the relative production attention, which is the time participants reveal production information divided by the total time information is revealed. The relative production attention was the sum of relative attention to own production and the attention to the production of the recipient. Our hypothesis implied a higher relative production attention in the talent treatment than in the effort treatment. The relative production attention in the talent related task was 52.21% (SD= 11.35) and the relative production attention in the effort task was 45.69%
(SD=24.15), the difference between the relative production attention variables was significant (t (118) = 2.50, p = 0.069). This was in line with our hypothesis. Based on this test, we could conclude that participants paid more attention to production information in allocations related to the talent task than in allocation related to the effort task.
Two regression were computed to estimate the relative production attention (see table 8). In both regression the coefficient of the talent dummy has a positive sign, this implied that
4 In total 1154 were used to test this hypothesis. Since in 274 of the in total 1428 allocation decisions, participants did not reveal information at all.
participants pay relatively more attention to production information in talent-related decisions compared to effort-related decisions. This is in line with our hypothesis and the result of the statistical test in the previous paragraph. According to regression 3, participants paid 4.5 percentage points more attention to production information in the talent-related decisions than in the effort-related decisions (t (112) = 1.79; p = 0.077). After including the control variables in regression 4, there was still a positive relation between decisions related to the talent task and the relative production attention, though not significant (ß=0.038; t (112) = 1.42; p = 0.157). According to the statistics of this section we could confirm hypothesis 2. Thus, participant paid more attention to production information compared to monetary contribution information in the talent-related allocation decisions than in the effort-related allocation decisions. Furthermore, the control variables in regression 4 did not significantly affect the relative attention variable.
Y = Relative production attention
Regression 3 Y
Regression 4 Y
X Marginal Effect (SE) p-value Marginal Effect (SE) p-value
Talent 0.048 (0.028) 0.077 0.038 (0.026) 0.157
Production other -0.016 (0.060) 0.787
0.005 (0.003) 0.397
Female 0.019 (0.028) 0.502
Constant 0.470 (0.024) 0.000 0.036 (0.142) 0.093
Control No Yes
N 1154 1154
𝑅2 0.010 0.062
Table 8 OLS Regression with clustered standard errors: Relative production attention
Notes: Talent = 1 if the is decision is based on the talent task, Talent = 0 if decision is based on effort task;
Production other denotes the percentage produced by the recipient; High pay rate = 1 if the high pay rate was assigned to the participant, High pay rate = 0 otherwise; participant production equals absolute production of the participant in an allocation decisions; Female = 1 if participant is a female; Female = 0 otherwise.
5.5. Attention and allocation
Hypotheses 3 related to the attention and the allocation of the participants. It was investigated whether participants allocated more in line with production if they paid more attention to the production information compared to the monetary contribution information. It was