The Endogenous Nature of Social Preferences

1
The Endogenous Nature of Social Preferences

• John Smith
• Rutgers University-Camden

2
Introduction

• Can social preferences be measured in a
  nonintrusive manner?
• Do measurement techniques affect the quantity
  which they are designed to measure?
• If so, what can we say about the relationship?

3
Introduction

• Experiment
• measure social preferences
• observe play in dictator game
• vary the order
• If measurement has no effect on preferences
• We expect no difference between the two
  treatments

4
Main Results

• Standard Dictator Game
• Split 10
• Increase amount kept
• equivalent reduction in amount sent
• We find that prosocial people
• Act even more procially when
• Preferences measured first
• Selfish people
• unaffected by the order of the measurement

5
Main Results

• Modified dictator game
• Increase amount kept by 0.50
• Reduces amount sent by 1.50
• Decision also involves surplus
• Prosocial and selfish people are unaffected by
  timing of measurement
• Endogeneity related to decisions involving surplus

6
Related Literature SVO

• Social Value Orientation (SVO)
• Van Lange et. al. 1997
• 9 dictator allocation decisions
• with 3 choices each
• Prosocial
• Positive value others material payoffs
• Individualistic
• Do not consider others material payoffs
• Competitive
• Negatively value others material payoffs
• If 6 responses of particular type
• Classified as such

7
Related Lit Other Measurement Methods

• Ring measure of SVO
• 24 pair-wise dictator allocation decisions
• GARP
• Andreoni and Miller (2000)
• Dictator allocation decisions
• Charness and Rabin (2002)
• Chen and Li (2009)
• Dictator allocation decision and simple games
• Our measure is transparent and easy to interpret
• Unknown precisely how these relate

8
Related Lit Endogenous preferences

• Carpenter (2005)
• Guth et. al. (2008)
• Canegallo et. al. (2008)

9
Study 1 Procedure

• 96 student subjects
• Attendance for course credit
• Pay 25 of each class amount earned
• Standard 10 dictator game

10
Study 1 Procedure

• SVO
• A
• You 480 points
• Other 80 points
• B
• You 540 points
• Other 280 points
• C
• You 480 points
• Other 480 points

• 1 Point0.002103
• SVO amount earned
• Ranges from 0.94 to 1.06
• SVO amount sent
• Ranges from 0.19 to 0.94

Competitive
Individualistic
Prosocial
11
Study 1 Procedure

• Half SVO measurement then dictator game
• SVO First
• Half dictator game then SVO measurement
• SVO Last
• Completely no strategic considerations

Triadic
12
Study 1 Results
13
Study 1 Results
14
Study 1 Results
p-value 0.074
Only 9
p-value 0.054
15
Study 1 Discussion

• Prosocial subjects are affected by timing of
  measure
• Individualists are unaffected by timing of
  measure
• Why?

16
Study 1 Discussion

• No choice of surplus in standard dictator game
• Perhaps if both have decisions involving surplus
• Prosocials will not be affected by timing
• Prosocials and Individualists converge
• Maybe too easy to be mean in dictator game
• Perhaps if being selfish is more expensive
• Selfish subjects will be affected by timing
• Prosocials and Individualists diverge

17
Study 2 Procedure

• 90 subjects
• Exactly as Study 1
• Except that used modified dictator game
• Relative allocation price is 1 to 3
• Most selfish allocation
• Keep 10, send 0
• Least selfish allocation
• Keep 0, send 30
• Dictator decision involves surplus
• Being selfish is relatively expensive

18
Study 2 Results
ProSocial and ProSelf different
t3.24 plt0.01
19
Study 2 Results
Not significant
20
Study 1 Results
p-value 0.72
Timing not significant
for either types
p-value 0.30
21
Discussion-Study 1

• Standard dictator game and measure SVO
• Keep additional 1
• Reduce amount sent by 1
• Prosocial subjects
• are affected by the timing
• Individualists
• not affected by the timing

22
Discussion-Study 2

• Nonstandard dictator game and measure SVO
• Keep additional 0.50
• reduce amount sent by 1.50
• Neither prosocials nor individualists
• are affected by the timing

23
Discussion

• Seems that creation of surplus is key to
  endogenous preferences of prosocials
• When dictator game involves choice of surplus
• Timing does not matter
• Make being selfish more expensive
• Does not induce individualists to be affected by
  timing

24
Future Work

• Timing affect individualists?
• Or are preferences less malleable?
• Significance that SVO only measures preferences
• when not behind other subject

25
Future Work

• Timing matter in
• Ring Measure of SVO
• GARP?
• Charness-Rabin, Chen-Li experiments?
• Conjecture that it will
• measure used here is extremely transparent
• Implications in ultimatum or trust game?
• In the presence of strategic considerations

26
Triadic Design

• You allocate between
• You and Other1
• Other1 allocates between
• Other1 and Other 2
• Other2 allocated between
• Other 2 and You

Study 1 Data
27
Study 1 Results
28
Study 2 Results