Charles J. Romeo and Barry Sopher, "Learning and Decision Costs in One
Person Games", Journal of Applied Econometrics, Vol. 14, No. 4, 1999,
pp. 335-357. 

Readme file for dataset "games". 

Dataset size: 11,400 rows and 7 columns.

The data are in the file games.dat, which is in DOS format and is zipped in
the file rs-data.zip.

Column Names: trial decision playershock cost prize computershock earnings.

trial--a sequence of integers going from 1-100 for each player.
decision--an integer in [0,100] 

playershock--a uniform random number on [-40,40] that is applied to the
player's decision.

cost--a value that is function of the players decision, and the game.

prize-gross return from making a particular decision.  

computershock--a uniform random number on [-40,40] that is applied to the
computer's choice.

earnings-net return from making a particular decision: earnings = prize -
cost.

There are 100 records of data for each of the 114 players.  The data for
real games are stacked on top of the data for hypothetical games, and the
data for both real and hypothetical games are ordered Game 1 followed by
Game 2 followed by Game 3.

Ordering of the data:

1. Game 1R, 12 players.

2. Game 2R, 16 players.

3. Game 3R, 12 players.

4. Game 1H, 22 players.

5. Game 2H, 17 players.

6. Game 3H, 35 players.

Note: The computer programs used in the paper are available from either
author.


Chuck Romeo
Rutgers University
(732) 932-8260
romeo@economics.Rutgers.edu