Warfare Experience during Ontogeny Increases

Egalitarian and Parochial Motivations

Michal Bauer
Charles University in Prague and CERGE-EI

Alessandra Cassar
University of San Francisco

Julie Chytilov
Charles University in Prague

Joseph Henrich
University of British Columbia

University of Maastricht
February 16th, 2012

Motivation
Other-regarding preferences
are common and facilitate cooperation in large groups (Henrich et
al. 01, Camerer 03, Fehr and Gachter 02, Fehr and Fischbacher
03)
exogenous? Shaped by a society and its history?

Legacies of war
Development in reverse (Collier 2003)
Negative effects on physical and human capital, at least in the short term (e.g., Deininger
2003, Verwimp 2005, Do and Iyer 2009)
Break down of social norms, deterioration of institutions and re-occurrence of conflict
(Collier et al. 2003; Collier and Hoeffler 2004)

Positive effects on social capital and institution-building?

More voting in Uganda (Blattman 09), participation in local collective action in Sierra
Leone (Bellows and Miguel 08), more sharing with neighbors (Voors et al. fothcoming)
State-formation and nation-building in Europe after WWII (Tilly and Ardant 75)

Motivation
Evolutionary approach: Inter-group conflict as a source of
human prosociality (Bowles 2008, Choi and Bowles 07)
Parochialism: stronger group identity to harness out-group hostility
Suppress within-group competition: inequality aversion, altruism
Promote within-group cooperation

Research question: Does experience of warfare

intensify other-regarding preferences and
increase parochialism (in-group bias)?

This project
Two post-conflict societies
Republic of Georgia: short-term effects of war (6 moths)
Sierra Leone: long-term effects of war (10 years)

Other-regarding types and parochialism (ingroup bias)

A set of binary dictator games
Varying identity of anonymous partner

Experience of warfare during childhood and adolescence vs.

adulthood
Georgia: sample of young children
Sierra Leone: sample of adults (some were children or adolescents during
the war)

Literature
Lab experiments
Intergroup competition intensifies group-based altruism and punishment
in the lab (Bornstein 03, Bornstein and Ben-Yossef 94)
Thoughts of death induce a greater adherence to social norms (terror
management theory), Greenberg et al. 86, Pysczynski et al. 04

Field experiments in post-conflict societies

More rewarding of cooperation and punishing non-cooperators during
Isreal-Hezbollah conflict (Fessler and Gneezy 11)
Experience of violence in Burundi increases sharing with neighbors
(Voors et al forthcoming)

Childhood and adolescence as a formative period

Social preferences are acquired and internalized during childhood
(Eisenberg and Mussen 99, Fehr et al. 08, Almas et al. 2010)
Differences in socio-economic background affect formation of skills
(Heckman 06, Cunha et al. 07) and other-regarding preferences (Bauer
et al 11)

Republic of
Georgia

A small Caucasus country

Since 90s transition from central planning to market
War in August 2008
A quick war with Russia over South Ossetia, a breakaway region
Largely based on aerial, artillery and tank fire strikes; no interpersonal
fighting
Substantial human and material losses among civilians, more than
100 ths. IDPs (HRW 2009)

Data collection
Data collection 6 months after war (Jan-Feb 2009)
564 children across the age 4-11yrs

Experimental design
4 binary choices in dictator games
Sharing game: (1,1) vs. (2,0)
Choosing (1,1): inequality aversion, generosity
Choosing (2,0): selfishness, spitefulness
Envy game: (1,1) vs. (2,3)
Choosing (1,1): inequality aversion, spitefulness
Choosing (2,3): selfishness, generosity, efficiency concerns
Costless sharing game: (1,1) vs. (1,0)
Costless envy game: (1,1) vs. (1,2)
Sharing game

Envy game

(2,0)

(1,1)

(2,3)

Selfish

Generous

(1,1)

Spiteful

Inequality averse

Experimental design
Measuring parochialism group identity
Ingroup partner: an anonymous child from the same class
Outgroup partner: a child from an unknown class

Payments
tokens to be exchanged for sweets, toys, etc.

Procedure
One-by-one explanation, cross-check questions, randomization of order
and spatial allocation of (1,1) option

Survey data
War experiences and IDP status
Observable characteristics (age, gender, siblings)

Choice situation: Georgia

Motivation to reveal preferences

Summary statistics: Georgia

All

7-11 years old

Mean
(SD)

3-6 years old

Mean
(SD)

Mean

(SD)

Panel A: Warfare exposure

Non-affected

0.319

(0.466)

0.313

(0.464)

0.33

(0.471)

Affected & Non-IDP

0.442

(0.497)

0.488

(0.501)

0.351

(0.479)

Affected & IDP

0.239

(0.427)

0.199

(0.400)

0.319

(0.467)

Panel B: Child's characteristics

Female

0.484

(0.500)

0.493

(0.501)

0.465

(0.500)

Age (years)

7.724

(2.068)

8.912

(1.345)

5.34

(0.834)

Has a sibling

0.862

(0.345)

0.87

(0.337)

0.847

(0.361)

Has a brother

0.558

(0.497)

0.585

(0.493)

0.5

(0.501)

Has a sister

0.541

(0.499)

0.531

(0.500)

0.563

(0.497)

Height (cm)
Lived in urban area before war

130.3
0.554

(13.24)
(0.498)

137.1
0.546

(9.42)
(0.499)

116.7
0.571

(8.41)
(0.496)

Lived in South Ossetia before

war

0.22

(0.415)

0.199

(0.400)

0.262

(0.441)

Warfare and choices

(7-11yr-old children)

Warfare and choices

(7-11yr-old children)

Warfare and choices (7-11yr-old children)

Dependent variable
Sample

Affected&Non-IDP
Affected&IDP
In-group
Affected&Non-IDP*In-group
Affected&IDP*In-group
Age
Female
Has a brother
Has a sister
Observations

Egalitarian choice in the

sharing game: (1,1) vs. (2,0)
All
(1)
-0.183*
(0.055)
-0.190*
(0.091)
0.0005
(0.996)
0.148
(0.244)

In-group
(2)
-0.048
(0.583)
0.130
(0.206)

Out-group
(3)
-0.186**
(0.048)
-0.197*
(0.069)

Egalitarian choice in the

envy game: (1,1) vs. (2,3)
All
(4)
-0.082
(0.362)
-0.126
(0.242)
-0.179*
(0.063)
0.288**
(0.023)

0.317**

0.499***

(0.032)

(0.0009)

0.055***
(0.009)
-0.042
(0.459)
0.011
(0.860)
-0.034
(0.557)
341

0.069**
(0.016)
-0.030
(0.693)
-0.050
(0.544)
-0.046
(0.567)
184

0.041
(0.192)
-0.056
(0.493)
0.071
(0.400)
-0.028
(0.739)
157

-0.041**
(0.042)
-0.046
(0.391)
-0.090
(0.118)
-0.115**
(0.039)
341

In-group
(5)
0.198**
(0.023)
0.382***
(0.000)

Out-group
(6)
-0.086
(0.334)
-0.133
(0.208)

-0.030
(0.280)
-0.010
(0.889)
-0.099
(0.224)
-0.124
(0.121)
184

-0.055*
(0.057)
-0.091
(0.247)
-0.085
(0.292)
-0.102
(0.195)
157

Probit, marginal effects (dF/dx) reported ; p-values in parentheses, *** denotes significance at 1% level, ** at 5% level and * at 10% level.

Other-regarding types in in-group treatment

(7-11yr-old children)

Other-regarding types in in-group treatment

(7-11yr-old children)

Dependent variable
Sample
Affected&Non-IDP
Affected&IDP
Age
Female
Has a brother
Has a sister
Observations

Selfish

Egalitarian

Generous

7-11 year-old children in Georgia

(1)
(2)
(3)
-0.0356
0.118
-0.144*
(0.609)
(0.106)
(0.082)
-0.223***
0.298***
-0.133
(0.006)
(0.002)
(0.162)
-0.034
0.006
0.064**
(0.136)
(0.798)
(0.021)
0.015
-0.022
-0.003
(0.813)
(0.707)
(0.970)
0.063
-0.085
0.036
(0.347)
(0.148)
(0.654)
0.085
-0.084
0.035
(0.202)
(0.146)
(0.653)
184
184
184

Spiteful

(4)
0.092
(0.181)
0.121
(0.185)
-0.034
(0.124)
0.016
(0.787)
-0.011
(0.863)
-0.040
(0.535)
184

Probit, marginal effects (dF/dx) reported ; p-values in parentheses, *** denotes significance at 1% level, ** at 5%
level and * at 10% level.

Other-regarding types in out-group treatment

(7-11yr-old children)

Other-regarding types in out-group treatment

(7-11yr-old children)

Dependent variable
Sample
Affected&Non-IDP
Affected&IDP
Age
Female
Has a brother
Has a sister

Observations

Selfish
Egalitarian Generous
Spiteful
7-11 year-old children in Georgia
(1)
(2)
(3)
(4)
0.302***
0.00657
-0.189**
-0.0927
(0.00216)
(0.913)
(0.0247)
(0.204)
0.273**
-0.100
-0.0865
-0.0333
(0.0243)
(0.163)
(0.359)
(0.695)
-0.00496
-0.0139
0.0572**
-0.0363
(0.870)
(0.496)
(0.0440)
(0.128)
0.127
-0.0204
-0.0362
-0.0726
(0.122)
(0.695)
(0.620)
(0.264)
-0.0665
-0.0701
0.135*
-0.0129
(0.426)
(0.192)
(0.0749)
(0.848)
0.0506
-0.0794
0.0528
-0.0191
(0.537)
(0.130)
(0.480)
(0.775)
157

157

157

157

Probit, marginal effects (dF/dx) reported ; p-values in parentheses, *** denotes significance at 1%
level, ** at 5% level and * at 10% level.

Small children: No effects

Sample: 3-6 yr-old children

War exposure exogenous?

Regional differences in social norms
Sub-sample analysis: urban vs. rural areas
Controlling for region fixed effects

Selective targeting
Unlikely given the bombing and speed of the conflict
Children are less likely to singled out
Sub-sample analysis: sample restricted to those from locations with
little pre-war interaction with separatist groups

Controlling for region fixed effects

(7-11yr-old children)

Dependent variable

Egalitarian choice in the

Sharing game: (1,1) vs. (2,0)

Sample

Affected&Non-IDP
Affected&IDP
In-group
Affected&Non-IDP*In-group
Affected&IDP*In-group
Age
15 Region fixed effects
Observations

Egalitarian choice in the

Envy game: (1,1) vs. (2,3)

7-11 year-old children in Georgia

All

In-group

(1)
-0.136
(0.142)
-0.148
(0.286)
0.033
(0.730)
0.119
(0.329)
0.339**
(0.022)
0.075***
(0.001)
Yes
377

(2)
-0.0202
(0.818)
0.145
(0.328)

Outgroup
(3)
-0.131
(0.168)
-0.099
(0.531)

0.086***
(0.007)
Yes
197

0.063*
(0.067)
Yes
172

All

In-group

Out-group

(4)
-0.115
(0.203)
-0.117
(0.385)
-0.213**
(0.026)
0.316**
(0.010)
0.493***
(0.001)
-0.0726***
(0.001)
Yes
373

(5)
0.195**
(0.025)
0.461***
(0.003)

(6)
-0.140
(0.147)
-0.198
(0.207)

-0.052*
(0.087)
Yes
202

-0.097***
(0.004)
Yes
171

Probit, marginal effects (dF/dx) reported; p-values in parentheses, *** denotes significance at 1% level, ** at 5% level and * at 10% level.

Civil war in Sierra Leone

West African country: 5.3 mil., extremely poor
Long and extremely bloody civil war (1991-2002)
Bad governance, diamonds and outside support of rebel groups
Atrocities committed by Revolutionary United Front (RUF) but also other
groups including Sierra Leone Army
Victims were mainly civilians (50,000 killed, half of the population displaced)

Welldocumented by representative surveys

Conibere et al 04, Humphries and Weinstein 06, Bellows and Miguel 08
Ethnic or religious affiliation did not play a major role in targeting civilians ->
villages [not individuals] were a major target
More experience of violence, more local collective action (BM)

Sierra Leone

Sierra Leone
Sample
Northern part of SL, 20 different villages (selected based on previous data of
BM on exposure to violence)
587 adult subjects aged 18-84
7-20 years during the civil war
>20 during the civil war

Data collected in July 2010 -> long-term effects of war

Survey data
Measuring exposure to violence
Were any members of your household killed during the conflict? (42%)
Were any members injured or maimed during the conflict? (35%)

Wealth, ethnicity, religion, gender, education, family size, etc.

Group membership and participation in local collective action

Sample

Our sample

IRCBP

Mean

(SD)

Mean

Warfare exposure
HH member injured
HH member killed
Least affected
Mid-affected
Most affected

0,350
0,415
0.449
0.334
0.217

(0.477)
(0.493)
(0.498)
(0.472)
(0.412)

0,35
0,44

Individual characteristics
Female
Age (years)
Attended school
Number of siblings
Religion - Christian
Religion - Muslim
Religion - other
Ethnicity - Temne
Ethnicity - Limba
Ethnicity - other

0.705
40.715
0.261
7.474
0.362
0.631
0.007
0.635
0.286
0.078

(0.456)
(13.688)
(0.439)
(4.784)
(0.481)
(0.483)
(0.082)
(0.482)
(0.452)
(0.269)

0,49
42,1
0,31

Real life behavior

Attended community meeting in past year
Member of a traditional society
Member of a religious group
Voted in recent election
Participated in road brushing in past year
Knows name of Local Councilor
Knows name of Paramount Chief
Local participation index

0,960
0,381
0,512
0,947
0,602
0,534
0,831
4,795

(0.196)
(0.486)
(0.500)
(0.224)
(0.490)
(0.499)
(0.375)
(1.293)

0,57
0,94
0,37
0,42
0,77

Experimental design

Children in Georgia
Option A
Option B
Own Partner's
Own Partner's
payoff
payoff
payoff payoff
Costly games
Sharing game
Envy game
Costless games
Sharing game
Envy game

Adults in Sierra Leone

Option A
Option B
Own Partner's
Own Partner's
payoff
payoff
payoff
payoff

1
1

1
1

2
2

0
3

10
10

10
10

15
13

5
16

1
1

1
1

1
1

0
2

10
10

10
10

10
10

05
15

Choice situation: Sierra Leone

Experimental design
Measuring parochialism/group identity
In-group partner: an anonymous person from the same village
Out-group partner: an anonymous person from a different distant village

Payments
Monetary rewards (10 tokens=$1.25=daily wage)

Warfare during childhood and

adolescence
(Sharing game, comparison of Georgia and Sierra Leone)

Sierra Leone: 7-20 yr-old during the war

Percentage of egalitarian choices

1.0

Sharing game
(10,10) vs. (15,5)
Ingroup
Outgroup

0.8
0.6
0.4
0.2
0.0
Least affected
(N=76)

Mid affected
(N=48)
War exposure

Most affected
(N=39)

Warfare during childhood and

adolescence
(Sharing game, comparison of Georgia and Sierra Leone)

Sierra Leone: 7-20 yr-old during the war

Percentage of egalitarian choices

1.0

Sharing game
(10,10) vs. (15,5)
Ingroup
Outgroup

0.8
0.6
0.4
0.2
0.0
Least affected
(N=76)

Mid affected
(N=48)
War exposure

Most affected
(N=39)

Georgia: 7-11 yr-old children

Warfare during childhood and

adolescence
(Envy game, comparison of Georgia and Sierra Leone)
Sierra Leone: 7-20 yr-old during the war

Percentage of egalitarian choices

1.0

Envy game
(10,10) vs. (13,16)
Ingroup

Outgroup

0.8
0.6
0.4
0.2
0.0
Least affected Mid affected Most affected
War exposure

Warfare during childhood and

adolescence
(Envy game, comparison of Georgia and Sierra Leone)
Sierra Leone: 7-20 yr-old during the war

Percentage of egalitarian choices

1.0

Envy game
(10,10) vs. (13,16)
Ingroup

Outgroup

0.8
0.6
0.4
0.2
0.0
Least affected Mid affected Most affected
War exposure

Georgia: 7-11 yr-old children

Warfare and Choices (Sierra Leone, younger subjects)

Dependent variable
Sample
Mid affected
Most affected
In-group
Mid affected*In-group
Most affected*In-group
Age
Female
Number of brothers
Number of sisters
Attended school
Muslim religion
Temne ethnicity
Observations

Egalitarian choice in the

Egalitarian choice in the
Sharing game: (10,10) vs. (15,5)
Envy game: (10,10) vs. (13,16)
Adults who were 7-20 yr-old during the civil war In Sierra Leone
All
In-group Out-group
All
In-group Out-group
(1)
(2)
(3)
(4)
(5)
(6)
0.0486
0.286**
0.103
-0.0783
0.253**
-0.144
(0.732)
(0.0317)
(0.479)
(0.529)
(0.0447)
(0.267)
0.0454
0.264*
0.102
-0.203
0.313**
-0.227**
(0.764)
(0.0692)
(0.511)
(0.109)
(0.0257) (0.0495)
0.00545
-0.210*
(0.965)
(0.0706)
0.205
0.358*
(0.293)
(0.0645)
0.205
0.520**
(0.322)
(0.0121)
-0.0111
-0.0219
0.000574 -0.000432 0.00818
-0.0146
(0.309)
(0.172)
(0.972)
(0.966)
(0.574)
(0.309)
0.0341
-0.0990
0.153
-0.105
-0.196
-0.0867
(0.756)
(0.520)
(0.332)
(0.319)
(0.166)
(0.592)
-0.0163
0.0251
-0.0638**
-0.00503 -0.00743 0.00120
(0.345)
(0.304)
(0.0361)
(0.748)
(0.747)
(0.958)
-0.0190 -0.0457*
-0.0188
-0.00482 3.96e-05 -0.00284
(0.332)
(0.0860)
(0.572)
(0.793)
(0.999)
(0.927)
0.0743
0.0101
0.150
-0.00481
-0.0276
0.106
(0.457)
(0.939)
(0.359)
(0.957)
(0.800)
(0.479)
0.0403
-0.0318
0.152
0.0680
-0.139
0.397***
(0.709)
(0.828)
(0.344)
(0.488)
(0.297) (0.00458)
-0.0170
0.110
-0.176
0.0177
0.0753
-0.251
(0.874)
(0.441)
(0.295)
(0.855)
(0.559)
(0.111)
157

82

75

157

82

75

Probit, marginal effects (dF/dx) reported ; p-values in parentheses, *** denotes significance at 1% level, ** at 5% level and * at 10% level.

Other-regarding types in in-group treatment

(Sierra Leone, younger subjects)

Spiteful

Relative frequency

1.00
0.80
0.60

0.108
0.270

Egalitarian
0.120
0.320

0.054
0.280

0.40
0.20

Generous

0.568

Selfish

0.190
0.333

0.238

0.280

0.238

Mid affected

Most affected

0.00
Least affected

War exposure

Other-regarding types in in-group treatment

(Sierra Leone, younger subjects)

Dependent variable
Sample
Mid affected
Most affected
Age
Female
Number of brothers
Number of sisters
Attended school
Muslim religion
Temne ethnicity

Observations

Selfish
Egalitarian Generous Spiteful
Adults who were 7-20 yr-old during the civil
war In Sierra Leone
(1)
(2)
(3)
(4)
-0.261**
0.271***
0.00521
-0.0172
(0.0362) (0.00977)
(0.966)
(0.813)
-0.337***
0.186*
0.0673
0.0686
(0.00861) (0.0789)
(0.619)
(0.391)
0.0381**
0.0174
-0.0423*** -0.00980
(0.0124)
(0.117)
(0.00655) (0.232)
0.0626
-0.212*
0.130
0.00595
(0.673)
(0.0919)
(0.351)
(0.941)
0.00425
0.0201
0.00825
-0.0219
(0.861)
(0.108)
(0.690)
(0.122)
0.0290
-0.0132
-0.0258
0.0141
(0.283)
(0.365)
(0.315)
(0.352)
0.103
0.0838
-0.0734
-0.122**
(0.428)
(0.187)
(0.548)
(0.0369)
-0.0273
-0.192
0.191
0.0567
(0.846)
(0.124)
(0.122)
(0.367)
0.0847
0.185*
-0.189
-0.174**
(0.524)
(0.0807)
(0.151)
(0.0334)
82

82

82

82

Probit, marginal effects (dF/dx) reported ; p-values in parentheses, *** denotes significance at 1%
level, ** at 5% level and * at 10% level.

Warfare experience during adulthood

Sierra Leone: >20 yr-old during the war
Envy game
(10,1) vs. (13,16)

1.0
Percentage of egalitarian choices

Percentage of egalitarian choices

1.0

Sharing game
(10,10) vs. (15,5)
Ingroup
Outgroup

0.8
0.6
0.4
0.2

Ingroup

Outgroup

0.8
0.6
0.4
0.2
0.0

0.0
Least affected

Mid affected
War exposure

Most affected

Least affected

Mid affected
War exposure

Most affected

Warfare during adulthood

Dependent variable:
Sub-sample:

Most affected
Mid affected
Age
Female
Number of siblings
Observations

Selfish

Generous

Inequality averse

Spiteful

Ingroup Outgroup

Ingroup Outgroup

Ingroup Outgroup

Ingroup Outgroup

(1)

(2)

(3)

(4)

(5)

(6)

(7)

(8)

-0.167*

-0.041

0.147

0.095

0.020

0.009

0.018

-0.050

(0.091)

(0.088)

(0.096)

(0.082)

(0.075)

(0.064)

(0.068)

(0.044)

-0.104

-0.082

0.084

0.028

-0.036

0.045

0.031

0.006

(0.074)

(0.082)

(0.069)

(0.074)

(0.055)

(0.061)

(0.052)

(0.047)

-0.000

-0.001

-0.001

-0.001

0.002

0.001

0.000

0.001

(0.003)

(0.003)

(0.002)

(0.003)

(0.002)

(0.002)

(0.002)

(0.002)

0.006

-0.052

-0.002

-0.027

0.030

0.045

-0.042

0.012

(0.073)

(0.076)

(0.064)

(0.067)

(0.053)

(0.050)

(0.050)

(0.043)

0.014*

-0.000

-0.013*

0.002

-0.004

-0.000

0.002

-0.003

(0.008)

(0.007)

(0.007)

(0.006)

(0.006)

(0.005)

(0.005)

(0.004)

213

206

213

206

213

206

213

206

Probit, marginal effects (dF/dx) reported; standard errors in parentheses, *** denotes significance at 1% level, ** at 5% level
and * at 10% level.

Robustness checks
Endogeneity of war exposure?
Selective targeting
No evidence of selective targeting (Bellows and Miguel 08)
Sub-sample analysis: stronger results for a younger cohort that is less
likely to be singled out
Regional differences
Controlling for village fixed effects

External validity
Survey information about group membership and participation in local
collective action -> index
Attended community meeting, membership in religious group,
traditional society, voting during last local elections, road-brushing,
knowing local councilor, knowing paramount chief

Behavior outside of the lab

Dependent variable:

Local participation index (0-7)

Sample:
Most affected
Mid affected

(1)
0.408**
(0.204)
0.175
(0.165)

Generous
Inequality averse
Spiteful
Observations
R-squared

288
0.131

In-group treatment
(2)

0.127
(0.180)
0.578***
(0.211)
0.227
(0.243)
288
0.142

(3)
0.357*
(0.205)
0.144
(0.165)
0.079
(0.182)
0.533**
(0.212)
0.192
(0.243)
288
0.151

Out-group treatment
(4)
(5)
(6)
0.472***
0.484***
(0.177)
(0.179)
0.383**
0.392**
(0.163)
(0.164)
-0.130
-0.175
(0.177) (0.176)
-0.167
-0.173
(0.205) (0.205)
-0.122
-0.091
(0.233) (0.231)
276
277
276
0.178
0.154
0.182

OLS; standard errors in parentheses, *** denotes significance at 1% level, ** at 5% level and * at
10% level. We control for observable characteristics.

Concluding remarks
Individual exposure to war-related violence reduces selfishness
and increases in-group inequality aversion
More enduring if war is experienced during young age
Robust across completely different settings (Caucasus vs. West Africa)

Consistent with evolutionary theories of human cooperation

emphasizing intergroup conflict
Intensification of in-group-out-group gap (not overall pro-sociality, Voors
et al forthcoming, Fessler and Gneezy 11)

Concluding remarks
Understanding post-conflict recovery
positive effects on local cooperation in post-conflict societies, conditional on
war being ended (BM 08, Voors et al 2011)
Stronger group identity -> may also explain why wars are difficult to end as
well as repeated cycles of conflicts
Open question:
Inter-group conflict as one type of personal insecurity? Environmental
disasters?

Preference formation during lifecycle

childhood and adolescence as a particularly sensitive period (consistent with
Heckman 06, Cunha et al 09)

Thank you

Dependent variable: War index

probit (ME), p-values in parentheses
Sierra Leone
War index

Dependent variable
All

Female
Age
Attended school
Number of brothers
Number of sisters
Muslim religion
Temne ethnicity
Constant
Village fixed effects
Observations
R-squared

Younger
(1)
(2)
-0.0906
-0.169
(0.245)
(0.378)
-0.00206 0.00202
(0.398)
(0.907)
0.100
0.275*
(0.236)
(0.0693)
-0.0111
0.0135
(0.380)
(0.611)
0.0310** 0.0204
(0.0263)
(0.511)
0.0847
0.193
(0.351)
(0.270)
-0.00195 0.00142
(0.983)
(0.993)
0.761*** 0.512
(3.20e-06) (0.321)
no
no
577
158
0.020
0.068

Older
(3)
-0.0690
(0.423)
-0.00430
(0.182)
-0.0130
(0.900)
-0.0225
(0.106)
0.0354**
(0.0270)
0.0393
(0.712)
0.00350
(0.974)
0.934***
(1.09e-05)
no
419
0.018

All
(4)
-0.166**
(0.0373)
-0.00183
(0.466)
0.0976
(0.256)
-0.0123
(0.310)
0.0347**
(0.0168)
0.126
(0.222)
0.0957
(0.483)
0.309
(0.219)
yes
577
0.105

Younger
(5)
-0.336*
(0.0620)
-0.0130
(0.461)
0.201
(0.237)
-0.00512
(0.862)
0.0506
(0.127)
-0.0215
(0.927)
-0.306
(0.211)
0.976*
(0.0806)
yes
158
0.257

Older
(6)
-0.116
(0.199)
-0.00310
(0.353)
-0.0292
(0.781)
-0.0238*
(0.0830)
0.0384**
(0.0261)
0.150
(0.209)
0.241
(0.112)
0.269
(0.377)
yes
419
0.103

Dependent variable: Egal. choice in the Sharing game [(10,10) vs. (15,5)]
probit (ME), p-values in parentheses
Sierra Leone, VFE
Dependent variable

Ingroup
War index==1
War index==2
Ingroup*War index==1
Ingroup*War index==2
Female
Age
Attended school
Number of brothers
Number of sisters
Muslim religion
Temne ethnicity
Village fixed effects
Observations

Egalitarian choice in the sharing game: 10/10 vs. 15/5

Teenagers during war
Adult during war
All
Ingroup
Outgroup All
Ingroup
Outgroup
(1)
(2)
(3)
(4)
(5)
(6)
-0.0535
0.0326
(0.700)
(0.679)
-0.0599
0.319
0.258
0.0456
0.101
0.0175
(0.726)
(0.117)
(0.231)
(0.606)
(0.274)
(0.854)
-0.0546
0.227
0.142
0.0751
0.256**
0.0438
(0.755)
(0.282)
(0.547)
(0.429)
(0.0375)
(0.655)
0.315
-0.00858
(0.168)
(0.942)
0.264
0.0750
(0.268)
(0.596)
0.0334
-0.0397
-0.447
0.00321
0.0626
-0.0191
(0.794)
(0.844)
(0.197)
(0.957)
(0.490)
(0.836)
-0.0140
-0.0577** -0.00428 0.000906 0.00364
-0.00113
(0.268)
(0.0174)
(0.809)
(0.690)
(0.275)
(0.752)
0.0520
-0.0696
-0.0113
-0.00373 0.229**
-0.159*
(0.647)
(0.676)
(0.956)
(0.957)
(0.0468)
(0.0978)
-0.0330
0.0789*
-0.159*** -0.000175 -0.0158
0.0146
(0.123)
(0.0636)
(0.000274) (0.987)
(0.393)
(0.300)
-0.0152
-0.0907** -0.0824
-0.00665 -0.0227
0.00990
(0.494)
(0.0156)
(0.146)
(0.562)
(0.212)
(0.537)
-0.0308
-0.143
0.251
0.0981
0.310**
0.0171
(0.850)
(0.634)
(0.326)
(0.228)
(0.0138)
(0.891)
-0.175
0.300
-1.000*** -0.116
-0.0112
-0.0959
(0.386)
(0.342)
(0)
(0.318)
(0.963)
(0.549)
yes
yes
yes
yes
yes
yes
149
68
69
414
202
199

Dependent variable: Egal. choice in the Envy game [(10,10) vs. (15,5)]
probit (ME), p-values in parentheses
Sierra Leone, VFE
Dependent variable

Ingroup
War index==1
War index==2
Ingroup*War index==1
Ingroup*War index==2
Female
Age
Attended school
Number of brothers
Number of sisters
Muslim religion
Temne ethnicity
Village fixed effects
Observations

Egalitarian choice in the envy game: 10/10 vs. 13/16

Teenagers during war
Adult during war
All
Ingroup
Outgroup All
Ingroup
Outgroup
(1)
(2)
(3)
(4)
(5)
(6)
-0.218*
0.0845
(0.0735)
(0.215)
-0.0908
0.176
-0.153
0.0576
-0.0564
0.0249
(0.487)
(0.260)
(0.324)
(0.456)
(0.456)
(0.751)
-0.255** 0.217
-0.378*** -0.0276
-0.0549
-0.0806
(0.0340)
(0.269)
(0.00633) (0.732)
(0.580)
(0.298)
0.339
-0.0875
(0.107)
(0.366)
0.631***
0.0258
(0.00688)
(0.835)
-0.163
-0.0175
-0.153
0.0820
0.0525
0.142*
(0.164)
(0.920)
(0.567)
(0.100)
(0.479)
(0.0590)
0.00192
-0.0119
0.00255
0.00210
0.00261
0.00337
(0.852)
(0.546)
(0.882)
(0.289)
(0.334)
(0.269)
0.00123
-0.0927
0.110
0.140**
0.148
0.104
(0.990)
(0.512)
(0.585)
(0.0287)
(0.180)
(0.221)
-0.00656 0.0198
-0.0118
0.00671
0.0156
0.0107
(0.719)
(0.603)
(0.681)
(0.442)
(0.309)
(0.369)
-0.00356 -0.0265
0.0360
-0.00143 -0.00793 -0.00587
(0.851)
(0.483)
(0.285)
(0.880)
(0.620)
(0.669)
0.103
-0.162
0.283
0.158**
0.123
0.166
(0.445)
(0.460)
(0.153)
(0.0194)
(0.236)
(0.148)
0.298
0.484*
0.752**
-0.0522
-0.179
0.0663
(0.111)
(0.0594)
(0.0301)
(0.600)
(0.389)
(0.590)
yes
yes
yes
yes
yes
yes
152
71
59
415
210
191

Warfare and survey measures of

group participations
Table A4: War and real life behavior

Dependent
variable
PANEL A
Sample
Most affected
Mid affected
Observations
PANEL B
Sample
Most affected
Mid affected
Observations
PANEL C
Sample
Most affected
Mid affected
Observations

Attended Traditional Religious

community society
group
meeting member member
(1)
(2)
(3)

Voted in
last year
(4)

Voted in
pres. el.
(5)

Knows
Knows
local
paramount
Voted in Roadbrush
councillor
chief
local el.
ing
(6)
(7)
(8)
(9)

Adults, all
0.020
-0.017
(0.013)
(0.031)
-0.015
-0.058**
(0.015)
(0.028)
577
576

0.141***
(0.052)
0.083*
(0.048)
576

-0.001
(0.056)
0.045
(0.048)
577

-0.059
(0.043)
-0.042
(0.035)
577

0.013
(0.018)
0.011
(0.017)
565

0.220***
(0.055)
0.154***
(0.048)
577

0.122**
(0.055)
0.081*
(0.049)
577

0.013
(0.014)
-0.023
(0.016)
577

0.029
(0.020)
0.009
(0.022)
289

0.204**
(0.082)
0.199***
(0.065)
295

0.088
(0.079)
0.077
(0.066)
295

Adults, ingroup treatment

0.036
(0.034)
-0.065** -0.065**
-0.063*
(0.028)
(0.028)
(0.036)
237
237
294

0.140*
(0.075)
-0.022
(0.067)
294

-0.022
(0.082)
0.063
(0.066)
295

-0.067
(0.066)
-0.028
(0.047)
295

-0.007
(0.044)
0.008
(0.039)
198

0.219***
(0.075)
0.126*
(0.071)
282

0.154**
(0.077)
0.087
(0.072)
282

Adults, outgroup treatment

0.011
0.026
-0.057
(0.026)
(0.028)
(0.047)
-0.048
(0.040)
146
146
282

0.165**
(0.072)
0.194***
(0.067)
282

0.020
(0.076)
0.033
(0.071)
282

-0.057
(0.058)
-0.060
(0.053)
282

Marginal effects (dF/dx) reported ; standard errors in parentheses, *** denotes significance at 1% level, ** at 5% level and * at

Sierra Leone: 5-20 yr-old during the war

Percentage of egalitarian choices

1.0

Costless sharing game

(10,10) vs. (10,5)
Ingroup
Outgroup

0.8
0.6
0.4
0.2
0.0
Least affected

Mid affected

Most affected

War exposure

1.0
Percentage of egalitarian choices

Georgia: 7-11 yr-old children

Costless envy game

(10,10) vs. (10,15)
Ingroup

Outgroup

0.8
0.6
0.4
0.2
0.0
Least affected Mid affected Most affected
War exposure

More detailed classification of otherregarding types

Observed
frequency

Choice in
Costly sharing Costless sharing Costly envy
game
game
game

Aheadness averse
Behindness averse
Inequality averse
Generous
Spiteful
Selfish

(1,1) vs.(2,0)

(1,1) vs.(1,0)

(1,1) vs.(2,3)

(1,1)
any
(1,1)
(1,1)
(2,0)
(2,0)

(1,1)
any
(1,1)
(1,1)
(1,0)
any

any
(1,1)
(1,1)
(2,3)
(1,1)
(2,3)

Costless
envy game
(1,1)
vs.(1,2)
any
(1,1)
(1,1)
(1,2)
(1,1)
any

in %

0,36
0,32
0,11
0,10
0,06
0,33

Other-regarding types: Georgia

(more detailed classification of types)

1.0
0.8

B
Aheadness averse
Ingroup

Outgroup

0.6
0.4

Prevalence of the type

Prevalence of the type

1.0
0.8

Behindness averse
Ingroup

Outgroup

0.6
0.4
0.2

0.2
0.0
0.0
Non-Affected Affected &
Non-IDP
War exposure

Affected &
IDP

Non-Affected Affected &

Non-IDP
War exposure

Affected &
IDP

Other-regarding types: Georgia

(more detailed classification of types)

1.0
0.8

1.0

D
Egalitarian
Ingroup

Outgroup

0.6
0.4

Prevalence of the type

Prevalence of the type

0.8

Generous
Ingroup

Outgroup

0.6
0.4
0.2

0.2
0.0
0.0
Non-Affected Affected & Affected &
Non-IDP
IDP
War exposure

Non-Affected Affected &

Non-IDP
War exposure

Affected &
IDP

Other-regarding types: Georgia

(more detailed classification of types)

1.0
0.8

1.0

F
Spiteful
Ingroup

Outgroup

0.6
0.4
0.2
0.0

Prevalence of the type

Prevalence of the type

0.8

Selfish
Ingroup

Outgroup

0.6
0.4
0.2
0.0

Non-Affected Affected &

Non-IDP
War exposure

Affected &
IDP

Non-Affected Affected &

Non-IDP
War exposure

Affected &
IDP

Other-regarding types: Sierra Leone

(more detailed classification of types)

1.0

1.0

B
Aheadness averse

0.8

Ingroup

Outgroup

0.6
0.4
0.2

Prevalence of the type

Prevalence of the type

0.8

Behindness averse
Ingroup

Outgroup

0.6
0.4
0.2
0.0

0.0
Least
affected

Mid affected
War exposure

Most
affected

Least
affected

Mid affected
War exposure

Most
affected

Other-regarding types: Sierra Leone

(more detailed classification of types)

1.0

1.0
Egalitarian
Ingroup

Prevalence of the type

Prevalence of the type

0.8

0.8

Outgroup

0.6
0.4

Generous
Ingroup

Outgroup

0.6
0.4
0.2

0.2

0.0

0.0
Least
affected

Mid affected
War exposure

Most
affected

Least
affected

Mid affected
War exposure

Most
affected

Other-regarding types: Sierra Leone

(more detailed classification of types)

1.0

1.0

F
Spiteful

0.8

Ingroup

Outgroup

0.6
0.4

Prevalence of the type

Prevalence of the type

0.8

0.0

0.0

War exposure

Most
affected

Outgroup

0.4
0.2

Mid affected

Ingroup

0.6

0.2

Least
affected

Selfish

Least
affected

Mid affected
War exposure

Most
affected