maarten berg results of a solidarity game experiment
TRANSCRIPT
Maarten Berg
Results of a solidarity game experiment
21/11/2011 Results of a solidarity game experiment
Content
• 1. Research questions and concepts• 2. (Experimental) game theory• 3. Design• 4. First results• 5. Phase 2: introducing identity and
social distance
21/11/2011 Results of a solidarity game experiment
1. Research questions and concepts
Research questions:
1. What are the conditions of solidarity?
2. What is the role of social distance?
21/11/2011 Results of a solidarity game experiment
1. Research questions and conceptsSolidarity:
- Definition: The performance of voluntary acts that benefit others, without the guarantee of receiving (at least) equal benefits in return
- Solidarity refers to behavior- Solidarity refers to voluntary behavior- Solidarity is not defined in terms of
support for public goods.
21/11/2011 Results of a solidarity game experiment
2. (Experimental) game theory
What is an experiment?
- Advantages of the experimental method- Control over the circumstances/conditions
of interest- The ability to draw causal conclusions
- Random assignment- Number of subjects
- Disadvantages of the experimental method- Low external validity
21/11/2011 Results of a solidarity game experiment
2. (Experimental) game theory
- What is a game?- Interactivity
- E.g. The so-called ‘Ultimatum Game’
- Rules of a game: how do ones decisions relate to the decisions of others and chance?- ‘No deception rule’- No social desirable statements
21/11/2011 Results of a solidarity game experiment
3. Design
Basic design of our ‘solidarity game’:
- Experimental groups consist of four players
- Subjects answer ten quiz questions modelling the productive effort in a real economy
- Two ‘winners’ and two ‘losers’- Two winners distribute 20 credits
(reflecting real money!)
21/11/2011 Results of a solidarity game experiment
3. Design
• 2*2 design• Performance-based vs random-
based• One shot vs four shots
One shot Four shots
Performance
Condition 1 Condition 3
Random Condition 2 Condition 4
• 2 studies• Study 1: Laboratory experiment:
156 UvA student subjects• Study 2: Field-experiment: 190
subjects from the Dappermarket-area
3. Design
21/11/2011 Results of a solidarity game experiment 9
21/11/2011 Results of a solidarity game experiment
4. Results of study 1
Number of credits that are voluntarily shared by the winners in round 1(maximum of 20):
Condition 1: 1.87 (N=16)Condition 2: 4.62 (N=24)Condition 3: 4.05 (N=20)Condition 4: 7.28 (N=18)
Conclusion: solidarity is possible
21/11/2011 Results of a solidarity game experiment
4. Results of study 1
One shot (condition 1 and 2): 3.53Four shots (condition 3 and 4): 5.48p = 0.07
Conclusion: when people are in a long-term, stable relationship with one another, they are more inclined to solidarity
21/11/2011 Results of a solidarity game experiment
4. Results of study 1
Performance (condition 1 and 3): 3.08 Random (condition 2 and 4): 5.76P= 0.02
Conclusion: When people feel / know that they do not deserve their wealth, they are more inclined to solidarity (or: when they do feel they deserve their wealth, they are less inclined to share).
21/11/2011 Results of a solidarity game experiment
4. Results of study 1
Solidarity in:
Round 1: 5.48 (Mean of both four shot conditions)Round 2: 3.58 (Mean of both four shot conditions)Round 3: 3.63 (Mean of both four shot conditions)Round 4: 0.74 (Mean of both four shot conditions)
- Strategic considerations?- Dissapointed by lack of solidarity in previous
rounds?
21/11/2011 Results of a solidarity game experiment
4. Results of study 1
Direct reciprocity: r = +0.326 (n=114), p < 0.001
(Gifts of the winners of round 1 with direct returns in round 2)
Indirect reciprocity: r=+0.423 (n=76, p<0.001)(Cumulative gifts to others with cumulative gifts
from others)
Conclusion: solidarity seems to be returned (to some extent).
• No proof of inequality-aversion• Correlation between gifts from
others in round 4 and cumulative earnings in round 1, 2 en 3 (controlled for being a winner or a loser in round 4): rp = + 0.048 (n.s.)
4. Results of study 1
21/11/2011 Results of a solidarity game experiment 15
21/11/2011 Results of a solidarity game experiment
5. Phase 2
• In phase 2 of the experimental study, people from the Dappermarkt-area in Amsterdam will be used (rather than students)– This is the same neighborhood that is
used in Laurens’ qualitative study
21/11/2011 Results of a solidarity game experiment
5. Phase 2
• More importantly, social distance is introduced in the design.– The participants know crucial aspects of
the identity of the other people in their group:• Sex• Age• Marital status• Etnicity / cultural background• Number of years of living in the
Dappermarkt-area
• We see substantially more solidarity than in study 1: 8.94 versus 4.53 (aggregated mean)
• P<0.001
5. Results from study 2
21/11/2011 Results of a solidarity game experiment 18
• We see less differences between the different conditions
• No greater solidarity in random conditions
• No greater solidarity in four shots conditions
• No proof of inequality-aversion• However, strong proof of
reciprocity
5. Results from study 2
21/11/2011 Results of a solidarity game experiment 19
Social distance within the group• We see most solidarity when there are
precisely two cultural groups represented in a group of four players, either in a 2:2 or a 3:1 ratio (Herfindahl-indices of 0.5 / 0.48).
• Cultural composition of the group seems to matter: F(4, 85) = 2.22 (p=0.073)
• Age compostion does not seem to matter
5. Results from study 2
21/11/2011 Results of a solidarity game experiment 20
Characteristics of the contributors• Solidarity is positvely correlated with age
of contributor (r= +0.175, p<0.10, n=90).• Indication that autochthones share
somewhat more than allochthones (9.21 versus 8.17, n.s.)
• We further need to analyze the effect of social distance between specific contributors and specific beneficiaries.
5. Results from study 2
21/11/2011 Results of a solidarity game experiment 21
• Too pessimistic view of mankind, based on experimental research with student subjects?
• Potential for solidarity rather fixed or context-dependent and variable?
• Is the role of social distance (in terms of age) underestimated or overestimated?
6. Implications and questions
21/11/2011 Results of a solidarity game experiment 22