PEM105 ASSESSMENT

Sample Coursework, adapted from a real submission


PART 1: Scenario description and analysis

Personnel management

This scenario assumes a real-life situation in which a company has to decide whether to invest
money to train its staff or not, based on situations I have seen in my corporate role.

A business has the possibility to invest a certain amount of money (say $6,000) to train a member
of its staff. The productivity of the trained worker would benefit the company by say $19,000
increasing the productivity of the business itself. Nevertheless, the latter, should now pay their
trained member an increased salary of say $4,000 as the worker is more specialized. However, the
individual could potentially change their job after the training, getting a higher salary from a
competitor who offers a $9,000 increase. If this is the case, the $6,000 investment made by the
company would be lost.

The business has the option not to invest any money in training the staff member, in which case no
expense is incurred, and everything remains as it is. Even in this situation, the worker could still
change job but they will have much more difficulties in finding a new position without the extra
training. Yet, if they decide to leave, the company would incur a small loss due the extra time and
expense of recruiting and onboarding replacement.

To create a simplified version of the scenario for analysis, let’s assume the set of choices available
to the company are the following:
• Train the staff
• Not train the staff

The set of choices available to the worker are:
• Keep the job and stay in the company
• Change the job and leave the company

This scenario could be treated as at least two forms of sequential game. One would have the
employer moving first, with the employee waiting to see whether the employer offers training
before deciding whether to leave or stay. Conversely the employer to hold off making a decision
until the employee either makes a commitment to stay or hands in their notice. However, an
alternative, perhaps more realistic conceptualization would treat the scenario as a simultaneous
game, in which both employer and employee are making long-term plans without knowing what
the other party intends to do. As such, the pay-off matrix of the one-shot game will be as follows:



1\2 TRAIN NOT TRAIN
LEAVE (9, -6) (1, -1)
STAY (4, 10) (0, 0)

• Player 1= the worker
• Player 2= the company

The values reflect deviations from the status quo (no training, employee stays). Clearly the best
outcome for the employee is to receive training and then leave for a more lucrative job. The best
outcome of the employer is to give the training and then reap the benefits of the new skills the
employee has developed as they choose to stay.

By solving the game, we are assuming that players are rational, which means that they are self-
interested and try to maximize their expected utility.

It is clear that the worker has a dominant strategy of leaving. In fact, no matter what the company’s
decision is, the individual will always have an incentive to leave since it ensures then a better pay-
off. Thus, the company will choose to train the worker only if they believe that the employee will
stay. However, under the assumption of rationality, it does not look reasonable that player 1 would
ever pick that strategy since it is a strictly dominated one.

Hence, the scenario is characterized by a unique Nash Equilibrium of (1, -1) in which each player
plays mutual best responses to each other (player 1 decides to leave and player 2 decides not to
train). Neither can improve their payoff by unilaterally changing their choice. However, there is
another combination of strategies that would lead to an improvement in at least one player’s
outcome without a deterioration in the other player’s one. Consequently, the equilibrium solution
is not Pareto optimal. In fact, the latter is Pareto dominated by the (4, 10) outcome in which player
1 chooses to stay in the company and player 2 to train the worker.

This is the simple one-shot analysis of the game. However, in the real world, the scenario is much
more complex and the optimal decision may not be that straightforward. The business often has
the possibility to make a binding agreement before the game so that the worker cannot change job
after the training. One issue to consider is the sense of affiliation of an individual to its company, a
variable which is not known by the business but could potentially affect the outcome of the game.
Employees with high affiliation would choose to stay if their employers invested in their training;
those with low affiliation would leave.


Sense of corporate affiliation

Suppose we have two types of workers: the first one characterized by a low sense of corporate
affiliation and a second one with a high level of it. We can try to incorporate this difference into the
theoretical model by making a re-specification of individual payoffs.

If we assume that high affiliation creates anticipated guilt which reduces the attractiveness for an
employee to take training and then leave and makes leaving at all less attractive, we can assume
that the payoff matrixes of the two types of individuals are:



1\2 TRAIN NOT TRAIN
LEAVE (9, -6) (1, -1)
STAY (4, 10) (0, 0)


Now, workers characterized by a high sense of belonging have a dominant strategy of staying, and
the Nash Equilibrium and pareto-optimal solution is Stay-Train.

If we now assume that the proportion of workers with a high sense of corporate affiliation is equal
to p, the company gains:

• 10 + (−6) (1 − ) = 16 − 6 → if it decides to train its staff
• 0 + (−1)(1 − ) = − 1 → if it decides not to train its staff

So, the threshold for choosing whether to invest money on training for a company will be where:
16 − 6 = − 1 → =
1
3
≈ 33%
In conclusion, as long as the proportion of workers with a high sense of belonging is greater than
33%, the company will be better off training its staff rather then not doing it.

In the real world, is not that unrealistic to think that at least 33% of employers develop a strong
sense of corporate affiliation and this might be one of reason why a lot of companies decide to
invest substantial amount of money to train their staff.

PART 2: Behavioral literature review

Literature review

Even though game theory could be a great tool to predict what players should rationally do in
strategic situations, the empirical evidence often shows deviation from the Nash Equilibrium in
many scenarios. This happens because economic predictions rely on the assumption of rationality
in which agents only care about maximizing their own payoff. On the other hand, the literature
suggests that people’s behavior is actually affected by many different factors rather than self-
interest.

1\2 TRAIN NOT TRAIN
LEAVE (3, -6) (-1, -1)
STAY (4, 10) (0, 0)
Affiliation No affiliation
The game theoretical scenario analyzed above can be described as a Prisoner’s Dilemma kind of
situation. Nevertheless, in this particular example, one economic agent is a person and the other is
a company, while in the standard game both parts are individuals.

Economic rationality suggests that in the Prisoner’s Dilemma game, both participants should choose
to defect. However, behavioral evidence indicates that people cooperate at higher levels than
expected if they were maximizing their outcomes.

For instance, in public good games (which work as the Prisoner’s Dilemma one), previous literature
has shown that individuals actually choose to cooperate and generally contribute halfway between
the free-riding and the Pareto optimal levels in one-shot games (Andreoni, 1998). This might happen
because in single-shot games, participants do not learn the incentives. Repeated trials of the same
game allow such learning, and thus experience may bring people to act rationally. However, as
Andreoni (1988) pointed out, the learning hypothesis alone fail to explain the cooperative behavior
since even after learning about the free-riding rewards, participants continue to contribute to the
public good (even if they might give less).

It could also be that human behavior takes into consideration not only the individual gain of
choosing a particular action but also the welfare of others. In this kind of games, it is highlighted the
conflict between individual and mutual interests so that cooperation may be explained by the fact
that people have prosocial preferences. Capraro, Jordan and Rand (2014) report three major forms
of social preferences:

• Efficiency: people may be willing to pay costs (to cooperate) to provide significant benefits
to others if they gain utility from aggregate welfare;
• Inequity aversion: people might want to cooperate in order to minimize the disparity
between their payoffs and the payoffs of others if they receive disutility from unfair
outcomes;
• Reciprocity: people gain utility by cooperating with cooperative people and not cooperating
with uncooperative individuals, so they might be able to pay the price of cooperation if they
expect others to do the same.

These prosocial motivations may be explained by factors such as altruism or empathy and moral
principles. Morality could lead people to cooperate because they feel the pressure to do what is
correct and honest.

To what concerns altruism, McClintock and Liebrand (1988) came to the conclusion that altruistic
motivation does not exhibit in Prisoner’s Dilemma games. However, the empathy-altruism theory
indicates that, while a general tendency toward altruism might be uncommon, empathy for a
specific person in a particular situation could lead to cooperate. For instance, Batson and Moran
(1999) in their study found that empathy-induced-altruism improved cooperation in a single-trial
Prisoner’s Dilemma.

Another important factor that leads people to cooperate in the Prisoner’s Dilemma game is the
feeling of guilt and shame associated with a “transgressive” or “unfair” decision. According to
previous studies, committing a transgression commonly results in more helpful behavior in future
interactions with an unattached third party. Timothy Ketelaar and Wing Tung Au (2003) found that
guilt motivates people to cooperate in repeated social bargaining games such as the Prisoner’s
Dilemma one. In fact, in their study, participants who felt guilty, as opposed to those who didn't,
after adopting a non-cooperative approach in the first round of the game, showed greater levels of
cooperation in the following round of the game. This result suggests that in social bargaining game,
non-cooperative individuals who experience the feeling of guilt might be using this emotion state
as "knowledge" about the potential costs of following an uncooperative strategy.

Finally, demographics and cultural background of participants play an important role in making
strategic decisions. In the scenario described in part one, different cultural background may entail
contrasting choices. For example, the employee retention rate of the trainees coming from a
cultural environment where moral commitment is supporting a sense of corporate affiliation
(typically Japan and far-east countries), may be higher than in other cultural context, such as
Western countries.

PART 3: Experiment overview

Experiment rationale

We have just seen how strategic decisions can be influenced by different behavioral factors. In the
Prisoner’s Dilemma, learning about other’s people optimal strategies, social preferences, altruism,
morality and the feeling of guilt may either induce participants to act rationally or to cooperate.
Additionally, demographics and cultural background can influence people’s actions.

In a business environment, such as the one described in part one, understanding individuals’ cultural
background may be relevant to evaluate possible outcomes of game theoretical scenarios. In fact,
worker’s cultural differences may entail diverse levels of moral motivation and cooperative
attitudes.
The aim of this experiment is to test whether cultural differences in participants lead to different
levels of cooperation in a standard one-shot Prisoner’s Dilemma game. To this extent, the game is
administrated separately to two different groups of participants in which one of them received a
certain stimulus to activate cultural knowledge before playing the game. If the cultural context had
no influence, outcomes of the two groups should not show any significant differences, i.e. the
confidence interval of the mean response of one group should overlap even partially the confidence
interval of the mean response of the other group, showing a p-value greater than 0.05.

Method

Participants. The sample contains 100 Japanese workers from different multi-national companies in
United Kingdom, in London (so that it will be easier to run the experiment). Every bi-cultural
employee, male or female, would be eligible for the study. This choice is made because, as Hong et
al. (2000) pointed out, bi-cultural people have internalized two cultures so that it is possible to
analyze cultural effects stimulating one particular cultural knowledge.

Design. Individuals will be randomly allocated either to the treatment or to the control (placebo)
group, with 50 participants in each condition. Workers randomly assigned to the treatment group
are subjected to a Japanese cultural priming condition (that will be explained later on) while those
in the placebo group do not receive anything. After that, both groups are asked to play
simultaneously the Prisoner’s Dilemma game in pairs. 10$ are given to each participant and they
can choose either to participate or not to a common pot. At the end of the game, individuals get the
money depending on the outcome. The experiment is run with real money since there is evidence
that, under certain circumstances, there is lot of a difference between what people will
hypothetically do and what they may actually do.

Procedure. Participants are asked whether they want to participate to the experiment or not and
are required to come to a specific building on a certain day. Individuals will then be randomly
allocated in the two groups and divided in two different rooms. The treatment group is subjected
to the Japanese primer condition: a Japanese audio content is played in the room. This condition
will spread activation in a network of cultural construct (Hong et al. 2000). In fact, there is evidence
that for bilingual people, the two languages are likely to be associated with diverse cultural system.
Subsequently, all participants are asked to play the game. A moderator will carefully explain the task
to both groups, including the setup, the available strategies and the possible payoffs. Each player
knows that the individual outcome will also depend on the other’s player decision.
A website link will then be provided to each individual which they can access through their phones.
After clicking on it, an algorithm will randomly allocate individuals (belonging to the same group) in
pairs and they can decide whether to cooperate (contribute to the pot) or to defect (not contribute
to the pot). At the end of the experiment, a voucher of the proper value is sent to each participant.

Analysis Plan

The variable of interest will be the number of time in which participants decide to cooperate in each
group. If, as assumed, cultural background matters in this kind of strategic decision, we expect the
mean response of the treatment group to be significantly different from the one in the placebo
group.
References

Andreoni, J. (1988). Why free ride?: Strategies and learning in public goods experiments. Journal of
public Economics, 37(3), 291-304.

Andreoni, J. (1993). An experimental test of the public-goods crowding-out hypothesis. The
American Economic Review, 1317-1327.
Batson, C. D., & Moran, T. (1999). Empathy‐induced altruism in a prisoner's dilemma. European
Journal of Social Psychology, 29(7), 909-924.

Brosig, J. (2002). Identifying cooperative behavior: some experimental results in a prisoner’s
dilemma game. Journal of Economic Behavior & Organization, 47(3), 275-290.

Burton-Chellew, M. N., & West, S. A. (2013). Prosocial preferences do not explain human
cooperation in public-goods games. Proceedings of the National Academy of Sciences, 110(1), 216-
221.

Capraro, V., Jordan, J. J., & Rand, D. G. (2014). Heuristics guide the implementation of social
preferences in one-shot Prisoner's Dilemma experiments. Scientific reports, 4(1), 1-5.

Hong, Y. Y., Morris, M. W., Chiu, C. Y., & Benet-Martinez, V. (2000). Multicultural minds: A dynamic
constructivist approach to culture and cognition. American psychologist, 55(7), 709.

Isaac, R. M., Walker, J. M., & Thomas, S. H. (1984). Divergent evidence on free riding: An
experimental examination of possible explanations. Public choice, 43(2), 113-149.

Ketelaar, T., & Tung Au, W. (2003). The effects of feelings of guilt on the behaviour of uncooperative
individuals in repeated social bargaining games: An affect-as-information interpretation of the role
of emotion in social interaction. Cognition and emotion, 17(3), 429-453.

McClintock, C. G., & Liebrand, W. B. (1988). Role of interdependence structure, individual value
orientation, and another's strategy in social decision making: A transformational analysis. Journal of
personality and social psychology, 55(3), 396.



Overall Evaluation

Analysis: Lit Review: Experiment Design
Distinction

Very nice presentation of the
scenario; compelling
translation into a payoff
matrix and subsequent
analysis. Sophisticated
consideration of mixed
population.
Borderline Merit/Distinction

Sound discussion of relevant
papers.
Borderline Merit/Distinction

Nice idea for an experiment -
I'd want to know what the
results were. Sensible setup;
just a couple of suggestions re
clarifying payoffs and the
implementation.