14.160  Experimental & Behavioral Econ

Hi Class--

Here's a quick apology--Thursday and Friday did not go as I had planned at 11 AM PST when I sent the announcement. I will get you feedback on both the group and individual proposals as soon as I finish corresponding with Ernst.

Sorry for the confusion!

--Randall

Hi Class--

Just a reminder that your group experiment write-ups are due on Sunday night @ 9pm EST (December 6, 2009). Please submit the reports on time to avoid a late penalty.

Also, you should receive feedback from me on your individual proposals by the end of today. Let me know if you have any questions.

Thanks and Congratulations on the execution of your experiments! I hope it all went smoothly!

--Randall

Hi Class--

Since I have not heard from any groups yet (I hope I didn't miss someone's email), I have selected the following people to be the treasurers for each group's funds:

Group 1: Jennifer Peck           jpeck1" AT "mit.edu
Group 2: Jan Kozak               jkozak" AT "fas.harvard.edu
Group 3: Anushree Subramaniam    anushree" AT "mit.edu
Group 4: Johanna Mollerstrom     jmollers" AT "fas.harvard.edu

These people will receive a check from Jessica Colon for $500 to be used in the group experiments. Please email her at <jcolon@mit.edu> to ensure that she has the right information and that you will know when to pick up your check.

Remember that you will be responsible to obtain the payment information I have mentioned earlier and to return any unused funds with the expense listing (see previous announcements).

Please let me know if you have any questions or concerns.

--Randall

Hi Class--

Here is a list of everything that I could think of regarding the finalization of your group experiments. Please email me or call me if you have any questions about these or other issues.

--Randall

Group Experiment Implementation:

SLOAN GROUPS DETAILS:

#1. If you are using the Sloan Trading Room or Economics Grad Lab to run your experiment need choose a "Team Captain" to contact me and be the agent responsible for your use of and access to the lab. Please call me to finalize this: 617..671..8294.

#2. If you are using the Sloan Trading Room, I can reserve time for you in the Sloan Trading Room. I need to know what times you would like to run your experiment. The more flexible you are, the better for everyone. In particular, daytime experiments are very good for tapping into the MIT student population that may have an extra hour between classes.

#3. If you are using the Sloan Trading Room, you have about 16 computers that you can book. You need to keep the back row of computers available for Sloan students during the experiments as they are generously letting us use the computers.

#4. If you are using the Sloan Trading Room, you should test your software on z-Tree in the lab before your first session so you can fix any technical bugs. You want the software to work flawlessly. Please call me if you are going to need to test the software before November 28th (as I will be back in Boston for a few days). The software is on all of the virtual machines, but currently one specific machine can act as the server or "Tree," while the other clients or "Leaves" will communicate with the server machine. This machine is on the inner front ring of 4 or 5 machines, the farthest to the left when standing in the front of the room, facing the computers. Windows+r --> "cmd" --> Enter --> "ipconfig" should give you the IP address of a given machine. This should line up with the zleaf shortcut's /server IPaddress for all of the leaves.

ALL GROUPS DETAILS:

#5. Recruiting subjects requires creativity and effort. You should start recruiting early so you have a good list of potential subjects to draw from.

#6. Include the dates and times of the experiment when you post flyers or send announcements so that potential subjects know enough to not miss it due to other conflicts. One way to recruit is with poster flyers with an email address tab people can pull off--help them make it easy to remember.

#7. Create a list of email addresses or other contact information as people sign-up so you can remind them the day before and the day of about their participation in the experiment. Try to assign subjects randomly to treatments. If they have 100% say over which date/treatment they are part of, your results might be correlated with temporally varying events (say, a Thursday versus Friday night experiment will likely compete differently with athletic events, concerts, school, etc. so your samples might be heterogeneous due to such unobservables.

#8. Overbook slightly as some people will email back saying that they can't make it anymore, even though they thought they could.

#9. Make sure your software (and/or budget) is flexible enough to handle variability in the number of participants. If too few people show up, you want to be able to run something. If too many people show up, you want to still have some budget to pay them at least $5 for showing up (even though they don't get to participate). Latecomers (>5 minutes late) certainly shouldn't get anything to be turned away.

#10. Don't be "innovative" by posting flyers where there aren't any already. You might get fined (see 2008). Be considerate about posting to email lists as well.

#11. Keep in mind that the MIT/Harvard IRB's should be sending out approvals to proposals that pass shortly following November 20th, and you need approval before running your experiment.

#12. Jessica Colon <jcolon""at""mit.edu> will help you get the money you need for the experiment ($500 per group). You can get her to write you a check beforehand or reimburse you following the experiment. You should decide upfront as a group who is going to be responsible for the money to avoid any accounting issues later on.

#13. In order to get money for the experiment, you must collect (and give to Jessica Colon) the following information about each participant for accounting purposes, separately from experimental information: Name, Citizenship, Address, and Payment Amount.

#14. Do the math beforehand on how much you expect to pay per person so that you don't go over budget and have to fork out additional cash from your own wallets to pay subjects (unless, of course, you'd like to). You should be paying people ~$15 per hour that you will have them there. Recalibrate your experimental payment parameters if necessary.

#15. You should make sure to have enough change (you can get this at the bank very easily--they can even prepare it for you if you tell them beforehand or when you show up) to pay everyone exactly what they earned (trying to figure out how to break a $20 at the end of an experiment can be very awkward). Make sure to have sufficient $5's, $1's, and quarters, nickels, dimes, and pennies (if relevant). Remember to take into account any show-up fees.

Hi Class--

Sorry for the confusion about the date of Sunday (11/8/09, NOT 11/7/09). I was running on very little sleep when I wrote the incorrect date for Sunday. Your individual proposals are due this evening. For those of you who recently sent me a draft to look at, I will review it shortly and will send you feedback.

--Randall