Experimental Economists

Orlando Gomez

September 14, 2016

Homework 2

1. Experimental economists take great care to provide sufficient stakes to participants. Explain why (10 pts)

- Experimental economists try to provide sufficient stakes to participants as an incentive for their participation. The participants have to acknowledge the opportunity cost of their time, and when being provided sufficient incentives for their participation, they are more likely to participate correctly in an experiment.

1. Explain the difference between a within and a between subjects design and tell me the benefits/costs of each approach (10 pts)

- Between subjects design each participant only experiences one of the possible versions of a specific factor. Within subjects design each participant experiences multiple factors within the same session. The benefits of within subject design are that it takes fewer people, but the cost is with the carryover effects such as fatigue decreasing the participant’s performance. A benefit of a between subject design is that it eliminates the carryover effects which is a reason why this design is mostly used for experiments. The cost is the complexity and how this design often requires a big amount of participants.

1. Define internal and external validity. Explain why each is important, and what an experiment should do to ensure their experiment is valid. (10 pts)

-Internal validity refers to determine if a causal relationship exists between one or more independent variables and one or more dependent variables. To ensure the experiment is valid experiments must be well designed, carefully controlled, and meticulously measured. External validity refers to the generalizability of an experiment. In other words it determines if the sample truly represents the whole population. To ensure the experiment is valid it may require realism, separate issue, etc.

1. Why is it important to randomize in an experiment? (10 pts)

-It is important to randomize in an experiment to reduce possible bias in the experiment that may arise due to bias responses.

1. Write down the factors and each treatment necessary to run each of the following experiments: (30 points, 10 each)

1.  I believe that the effect of punishment in the trust and public goods game depends on the anonymity of the punisher

• Factor 1: Game (Trust and Public Goods)
• Factor 2: Central Player
• Treatment 1: Public Goods punishment with and without anonymity.
• Treatment 2: Trust punishment with and without anonymity.

1. A student believes that Caucasians and African Americans react differently to a supply shift.

• Factor 1: Race (Caucasian vs. African American)
• Factor 2: Supply Shift
• Treatment 1: African American’s reaction with and without a supply shift
• Treatment 2: Caucasians reaction with and without a supply shift

1. A student in your class believes that men and women behave differently in markets and that their behavior depends on whether they are interacting with the same or different sexes. Furthermore, they believe that the gender effect may depend on whether they are operating in a pit market or double auction.

• Factor 1: Gender (Male or Female)
• Factor 2: Gender interaction (same or different sexes)
• Factor 3: Type of Market (Pit Market or Double Action)
• Treatment 1:  Gender interaction (same sex, both MM or FF)
• Treatment 2: Gender Interaction (Different sexes, FM or MF)
• Treatment 3: Gender interaction with Markets (Pit market or Double Action)

1. Critique the following experimental designs by identifying issues and explaining how to fix them (note: there may be more than one issue per design) (30 points, 10 each)

1. I am interested in testing for the effect of endowment in the trust game: To do so, I bring a group of participants into the lab and tell them that they are going to play “the Trust Game.” In the first treatment, I tell the participants that they can send any amount of a $10 endowment which will be multiplied by three and sent to a second player who can then choose any amount to send back to the first player. In the second treatment, I repeat the above procedure but endow players with $20 and only multiply the amount sent by 2.

-Naming the experiment the trust game is wrong because it incites that you are testing the trust of the participants, which can lead to bias responses. The best way to solve this is to using a name, which doesn’t interfere with what is really being tested. The reward amount multiplications are mentioned which could affect the participants’ moves. The best way to solve this is to keep this on the low, so participants won’t be bias when choosing which move to make.

1. I want to test how the threat of termination affects effort at work. To do so, I tell participants that they must complete a series of tasks and after each they will have a chance of being fired that is inversely related to the amount of work accomplished. Because of time constraints, I actually roll a six-sided die for each participant and fire them if I roll a 6. I tell each participant that $5 will be sent to the charity of their choice for their participation.

-The incentive of sending money to a charity is not too good, because some people prefer to stay with the money. This could lead to less engaging in the experiment. To solve this, the participant should be given the choice of keeping the money or donating it to the charity of their choice. Mentioning that they’ll have a chance of being fired “inversely related to the amount of work they do” simply incites the participants to not work at all. The way to fix this is not by mentioning which will be the method of firing participants, so their participation won’t be bias and to accurately measure how the threat of termination affects effort at work.