an international investigation of the influence of social distance on trust
TRANSCRIPT
Let’s Get Personal: An International Examination of the Influence of Communication, Culture and Social Distance on Trust and Trustworthiness
Nancy R. Buchan – Univ. of WisconsinRachel T.A. Croson – Univ. of Pennsylvania
Eric J. Johnson – Columbia University
Goal of Research
How does trust manifest itself in different countries and different contexts?Four countries
China, Japan, Korea, United States
Two cultural orientationsSocial DistancePersonal vs. Impersonal Communication
Perspective: Experimental Economics The Trust Game (Berg, Dickhaut & McCabe 1995)
Standard trust game Both parties receive $10 participation fee Amount sent is tripled, responder is a dictator with total One-shot, real money
Deviations from equilibrium – trust, trustworthiness
The Influence of Social Distance
Need to incorporate social distance to explain individual decisions bearing social consequences (Akerlof 1997)
Definition: Location in a social space (Akerlof 1997)
Degree of isolation from each other and experimenter (Hoffman, McCabe and Smith 1996)
Demographic similarity (Glaeser et al. 2000)
Social Distance and Culture
Minimal Group Paradigm (Tajfel et al. 1971)
Arbitrary categorization into mutually exclusive groups Significant ingroup bias in the United States (Orbell, van
de Kragt & Dawes 1988; Frey & Bohnet 1997)
The Influence of Culture on Group Definition and Formation Permanent / Natural groups vs. Temporary / Artificial
(Triandis 1995) Likely resistance from collectively oriented
participants (Mann, Radford, & Kanagawa 1985)
Social Distance and Culture
Hypothesis (tested at the aggregate and individual levels): Countries (US) / Individuals with relatively
more individualist orientation will display an ingroup bias.
Countries (China, Japan, Korea) / Individuals with relatively more collectivist orientation will not display an ingroup bias.
The Influence of Communication
Communication increases cooperation rates in SDs (Sally 1995) Why? Must communication be strategy-relevant?
Irrelevant communication (“cheap talk”) Roth (1995): 2-min., face to face, social discussion
Led to same amounts of increase in cooperation as relevant communication
Dawes, McTavish, & Shaklee (1977): 10-min., face to face, fact-related discussion
Led to same amount of cooperation as no communication
Irrelevant Content
Compare 2 types of irrelevant communication Personal Impersonal
Hypothesis: Among in-group pairings, participants who engage in
personal communication will display higher levels of trust and trustworthiness than those who engage in impersonal communication.
Experiment Participants
188 Ss – sophomore or junior business or economics majors50 Ss: Nankai University – China50 Ss: Seoul National University – Korea44 Ss: University of Tokyo – Japan44 Ss: University of Pennsylvania - US
Experiment Procedures - Communication Participants randomly assigned to one of
four color-coded groups (~ 12 members)Personal communication: Ss told to introduce
themselves and to discuss, from a list provided, a question pertaining to their birthdays
Impersonal communication: Ss told to answer from the given list, questions obtained from the world almanac
Experiment Procedures – Trust Game Participants randomly assigned to roles
(sender / responder), and directed to separate rooms
Received written instructions for game, experimental fee, and 2 colored index cards Numbered card represented color of Ss own
discussion group, and ID number Blank card represented color of partner’s discussion
group
Experimental Procedures
Double-blind Controls (Roth et al. 1999):
Ss pool equivalencyCurrency effects: 2,000 yen, 1,000 won,
10 yuan, $10 dollarsLanguage effectsExperimenter effectsComprehension
Experimental Questionnaire Senders: Completed after sending money
but before receiving any back Responders: Completed before receiving
money from senders Hypothetical games (next study)
Measures of Cultural Orientation
Cultural orientation: Broad measures of culture reflecting two fundamental differences in culture (Hofstede 1980; Triandis et al. 1988)
Competition vs. Cooperation “To what extent does it feel like you are competing or
cooperating with the responder (sender) in this experiment?”
Priority on self vs. group “What is more important to you in this experiment, maximizing
the amount of money that you and the (sender) responder will gain, or maximizing the amount of money you alone will gain?”
Experimental Results
Dependent Variables:Trust: Amount sent by senderTrustworthiness: Proportion of total wealth
returned by responder
Monetary amounts across countries standardized on scale from 0 to 1,000 units
Amount Sent Independent Variable Estimate St Dev t Value Intercept 457.45 53.06 13.44 ** Communication 278.23 52.93 2.51 * Groups 221.61 53.53 -0.90 China 278.67 37.18 1.41 ^ Japan 237.19 38.59 0.62 Korea 69.96 36.59 0.46 Comm*Groups 29.76 26.69 0.20 Comm*China -108.99 37.94 1.28 Comm*Japan -142.42 38.07 -0.63 Comm*Korea -3.51 36.55 0.78 Groups*China -328.16 38.00 -2.53 * Groups*Japan -309.03 38.85 0.97 Groups*Korea -137.41 36.59 -1.00 Gender -81.40 37.30 -0.30 Economics Educaton 41.20 34.25 -0.53
Adjusted R2 0.06 # of Observations 92 **p<.01 * p<.05 ^ p<.10
Country x Social Distance
Amount Sent (Country by Social Distance)
0
100
200
300
400
500
600
700
800
900
1000
China Japan Korea US
Am
ou
nt
Se
nt
Ingroup
Outgroup
Cultural Orientation
Independent Variable Estimate St Dev t Value Intercept 9.1590 1.2200 7.57 ** China 1.9140 1.4600 2.05 * Japan 0.6740 1.5200 -0.33 Korea 0.3940 1.4500 -0.31 Groups 0.0060 1.2900 0.59 Communication 0.3520 1.2800 -0.25 Proposer 0.3510 1.2800 -0.16
Adjusted R2 0.12 Number of Observations 184 **p<.01 * p<.05
Cultural Orientation x Social Distance
Amount Sent (Cultural Orientation by Social Distance)
0
100
200
300
400
500
600
700
800
900
1000
Collectivist Individualist
Cultural Orientation
Am
ou
nt
Se
nt
Ingroup
Outgroup
Cultural mediation Independent Variable Estimate St Dev t Value Intercept 460.24 48.23 14.76 ** Communication 108.43 47.65 2.28 ** Groups -65.16 47.47 -1.37 Cultural Orientation -316.65 48.08 -1.98 * China 28.35 33.85 0.84 Japan 27.76 34.08 0.81 Korea 10.55 32.42 0.33 Culture*China -38.31 33.74 -1.42 Culture*Japan -20.28 33.91 -0.60 Culture*Korea -32.31 34.22 -0.94 Comm*Culture -9.25 24.22 -0.38 Groups*Culture -73.94 34.03 -2.17 * Comm*China 72.67 33.41 1.18 Comm*Japan -27.20 33.58 -0.81 Comm*Korea 11.89 33.94 0.35 Comm*Groups 3.00 24.19 0.12 Groups*China -29.00 24.55 -1.18 Groups*Japan -28.33 34.47 -0.82 Groups*Korea -51.54 32.28 -1.30 Gender -61.25 65.33 -0.60 Economics Education 54.20 33.70 -0.43
Adjusted R2 0.29 Number of Observations 92 **p<.01 * p<.05
Communication – Main effect
Amount Sent (Communication by Social Distance)
0
100
200
300
400
500
600
700
800
900
1000
PersonalCommunication
ImpersonalCommunication
Am
ou
nt
Se
nt
Ingroup
Outgroup
.
Proportion Returned Independent Variable Estimate St Dev t Value Intercept 0.1053 0.0369 10.60 ** Communication 0.1331 0.0318 1.99 * Groups 0.1512 0.0315 1.30 China 0.2140 0.0230 2.08 * Japan 0.1155 0.0244 1.34 Korea 0.0587 0.0241 1.35 Comm*Groups 0.1345 0.0158 -2.13 * Comm*China -0.0544 0.0231 -0.59 Comm*Japan -0.0030 0.0233 0.03 Comm*Korea -0.0162 0.0231 -0.18 Groups*China -0.1821 0.0224 -2.03 * Groups*Japan -0.0638 0.0232 -0.69 Groups*Korea -0.0481 0.0223 0.54 Gender -0.0985 0.0200 2.46 ** Economics Education 0.0261 0.0240 -0.21
Adjusted R2 0.13 Number of Observations 92 **p<.01 * p<.05
Country x Social Distance
Proportion Returned(Country by Social Distance)
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
China Japan Korea US
Pro
po
rtio
n R
etu
rne
d
Ingroup
Outgroup
Cultural Orientation x Social Distance
Proportion Returned(Cultural Orientation by Social Distance)
0.00
0.10
0.20
0.30
0.40
0.50
Collectivist Individualist
Cultural Orientation
Pro
po
rtio
n R
etu
rne
d
Ingroup
Outgroup
Cultural Mediation Independent Variable Estimate St Dev t Value Intercept 0.1742 0.0903 1.93 ** Communication -0.0413 0.0204 -2.02 ** Groups 0.0244 0.0538 0.45 Cultural Orientation 0.0200 0.0085 2.35 ** China 0.0177 0.0693 0.26 Japan 0.0311 0.0595 0.52 Korea 0.0682 0.0566 1.20 Culture*China -0.0045 0.0045 -1.00 Culture*Japan -0.0219 0.0214 -1.03 Culture*Korea 0.0004 0.0211 0.02 Comm*Culture -0.0140 0.0206 -0.68 Groups*Culture -0.0208 0.0524 -1.79 ^ Comm*China -0.0240 0.0208 -1.16 Comm*Japan -0.0009 0.0200 -0.05 Comm*Korea 0.0003 0.0063 0.05 Comm*Groups -0.0023 0.0059 -0.39 Groups*China -0.0266 0.0145 -1.84 Groups*Japan 0.0034 0.0044 0.77 Groups*Korea -0.0001 0.0059 -0.01 Gender 0.0356 0.0185 1.93 ** Economics Educaton 0.0351 0.0125 0.56
Adjusted R2 0.41 Number of Observations 92 **p<.01 * p<.05 ^ p<.10
Effect of Communication
Proportion Returned (Communication by Social Distance)
0.00
0.05
0.10
0.15
0.20
0.25
0.30
0.35
0.40
0.45
0.50
PersonalCommunication
ImpersonalCommunication
Pro
po
rtio
n R
etu
rne
d
Ingroup
Outgroup
Discussion
The interaction of cultural orientation with social distanceCulturally influenced definitions and methods
of group formation may moderate extent to which ingroup bias in trust is demonstrated (see also Buchan, Croson & Dawes, AJS, 2002)
Implications for research methods used
Discussion
The content of non-strategy relevant communication influences the level of trust displayed
3 explanations for this result: Social Distance (Hoffman, McCabe & Smith 1996)
Decreasing social distance = higher trust
Social Identity (Brewer 1991; Tyler and Dawes 1993) Group identity increases expectations that members of
ingroup will return money
Discussion
Proposed explanation for communication effectPersonal communication prompts
greater “otherregardingness” (Bohnet & Frey 1999)
Examination of outgroup results reveals “carryover” effect of communication
A representative heuristic at play? (Kahneman & Tversky 1972)
Conclusion
“Cheap Talk” isn’t so cheap Our results regarding trust and
trustworthiness concur with Schelling (1968) who suggested, the more we know, the more likely we are to care.
The Boundaries of Trust: Own and Others’ Actions in the US and China
Nancy R. Buchan, University of Wisconsin
Rachel T.A. Croson, University of Pennsylvania
Exploring the Boundaries of Trust
Putting Fukuyama’s ideas to test (1995) Trust levels vary by country – US > China Location of trust boundary varies by country
US boundary is wide – including non-kin China boundary is close and tight – limited to family
members
Trust: “…the expectation of cooperative behavior based on commonly shared norms.”
Trust Games with Hypothetical Targets Examining trust and trustworthiness within
naturally occurring social networks
7 targets varying in social distance (similar to Kollock 1998) Parent Sibling Cousin Student know well Student from another university Stranger from hometown Stranger from another country
Intentions and Expectations
2 types of measures collected: Intention Expectation
Senders: “Imagine you were doing this experiment with the following responders. Please state how much you would send the responder, and how much you would expect to receive back from the responder.”
Responders: “Imagine you were doing this experiment with the following senders. Please state how much you would expect to receive from the sender, and how much you would return to them.”
Measures elicited during play of actual trust game Focusing on China and the US since they represent Fukuyama’s
starkest predictions.
Social Distance Hypothesis
Although decreases in trust and trustworthiness are expected as social distance increases for both American and Chinese participants, the drop in levels of trust and trustworthiness among Chinese participants between family and non-family members will be significantly greater than the drop in trust and trustworthiness among American participants.
Social Norm Hypothesis
Ostrom (2003): “Once some members of a population acquire norms of behavior, the presence of these norms affect the expectations of other players,” (p.41).
Expectations of the trusting actions of others will decrease as social distance increases in both the US and in China.
Expectations of the trustworthy actions of others will decrease as social distance increases in both the US and in China.
Method of Analysis
Mixed-model repeated ANOVA Amounts sent standardized scale 1-10 units Converted absolute amounts returned to
proportions returned. Covariates:
Exposure to game theory Gender – replicating prior results (Croson and Buchan
AER 1999, previous study)
Comparison with Actual Game
Actual game: participants paired with anonymous student from same university
Hypothetical: Between “Student you know well,” and “Student from another university”
Amount Sent
Proportion Returned
USActual 4.95 22%Student Know Well
7.31 35%
Student Another University
4.04 22%
ChinaActual 6.62 35%Student Know Well
8.12 35%
Student Another University
5.41 30%
Amount Sent – Sender’s Intentions
As social distance increases, amounts sent decrease (F(7,308)=70.17,
p<.01) China’s tighter boundary
of trust? No social distance x target
interaction
Figure 2: Amount Proposers Would Send Across Countries
0
1
2
3
4
5
6
7
8
9
10
Parent Sibling Cousin Student YouKnew Well
Student fromAnother
University
Stranger fromHome Town
Stranger fromAnother Country
Responder
Am
ou
nt
China (n=24) United States (n=22)
Amount Sent – Responder's Expectations As social distance
increases, expectations decrease (F(7,315)=36.6, p<.01)
Significant country x recipient interaction (F(7,315)=10.64, p<.01)
Figure 3: Amount Responders Expect to Receive Across Countries
0
1
2
3
4
5
6
7
8
9
10
Parent Sibling Cousin Student YouKnew Well
Student fromAnother
University
Stranger fromHome Town
Stranger fromAnotherCountry
Proposer
Am
ou
nt
China (n=25) United States (n=22)
Proportion Returned – Responder’s Intentions As social distance
increases, amounts sent
decrease (F(7,315)=20.06, p<.01)
China’s tighter boundary of trust? Again, no social distance x
target interaction
Figure 4: Proportion Responders Would Return Across Countries
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Parent Sibling Cousin Student YouKnew Well
Student fromAnother
University
Stranger fromHome Town
Stranger fromAnotherCountry
Proposer
Am
ount
China (n=25) United States (n=22)
Proportion Returned – Sender’s Expectations Significant country x
recipient interaction (F(7,308)=8.72, p<.01)
Figure 5: Proportion Proposers Would Expect To Have Returned Across Countries
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Parent Sibling Cousin StudentYou Knew
Well
Studentfrom
AnotherUniversity
Strangerfrom
HomeTown
Strangerfrom
AnotherCountry
Responder
Am
ount
China (n=24) United States (n=22)
What do these interactions mean?
Directly compare the actions senders or responders report they would take with the expectations about those actions from responders or senders.
Analysis: Repeated measures anovas using intentions and expectations as the DV, role as the IV (for both sending and returning)
US Comparison
Sending
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
Parent Sibling Cousin Student YouKnew Well
Student fromAnother
University
Stranger fromHome Tow n
Stranger fromAnother Country
Recipient
Am
ount
Amount Proposers Would Send Amount Responders Expect to Receive
Returning
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Parent Sibling Cousin Student YouKnew Well
Student fromAnother University
Stranger fromHome Tow n
Stranger fromAnother Country
RecipientA
mou
nt
Proportion Proposers Would Expect to Have Returned to Them Proportion Responders Would Return
US Comparison
No significant effect of role for either sending or returning – thus the levels of lines are the same.
Significant effect of counterpart for both sending and returning – thus the slopes of the lines are significant and negative.
Interactions are not significant – the slopes of the lines are the same as each other.
Strong relationship between intended behaviors and expectations in US.
China - Comparison
Sending
0.00
1.00
2.00
3.00
4.00
5.00
6.00
7.00
8.00
9.00
10.00
Parent Sibling Cousin Student YouKnew Well
Student fromAnother
University
Stranger fromHome Town
Stranger fromAnother Country
Recipient
Am
ount
Amount Proposers Would Send Amount Responders Expect to Receive
Returning
0.00
0.10
0.20
0.30
0.40
0.50
0.60
0.70
0.80
0.90
1.00
Parent Sibling Cousin Student YouKnew Well
Student fromAnother
University
Stranger fromHome Town
Stranger fromAnother Country
Recipient
Am
ount
Proportion Proposers Would Expect to Have Returned to Them Proportion Responders Would Return
China Comparison
Significant interactions for both sending and returning – the slopes of the lines are the different from each other.
Intended actions are more responsive to social distance than expectations.
Weak relationship between intended actions and expectations in China.
Discussion
Lack of support for Fukuyama’s tighter boundary of trust in China
Strong support for influence of social distance in both countries among natural groups Counterpoint to findings in previous studies with experimental
groups Stresses the importance of (embeddedness in) personal
networks of relationships in building trust (Granovetter 1985) Trust in China can occur outside kin (despite “bamboo network”)
3 categories of interpersonal relationships based on expressive, instrumental, or mixed ties (Bond and Hwang 1995)
The connection between own and other’s actions US: Responders’ expectations about
proposers’ reactions to social distance are consistent with proposers’ actual reactions, and vice versa.
China: Expectations are systematically less consistent.Actors are sensitive to social distance, but
neither expects the other side to be.
Role of Norms
Apriori expectations: In collectivist cultures strict norms governing
behavior in different contexts (Bond & Hwang 1986); adherence to norm demanded (Bierbrauer et al. 1994).
In individualist cultures, variety of norms is tolerated (Bierbrauer et al. 1994).
Test of variances yields no distinct pattern
Comparison of Individual’s Own Actions with His Expectations Correlation analysis: What a proposer says he
will send to each counterpart vs. what he expects to receive from each.
Parent Sibling Cousin
Student You Knew
Well
Student from Another
University
Stranger from Home
Town
Stranger from Another Country
United States 0.336 0.709 0.815 0.764 0.775 0.808 0.833 China 0.175 0.153 0.425 0.465 0.731 0.641 0.394
Data from Actual Game
Correlation analysis: What proposers sent vs. what they expected to receive back.
US: r = .794, p<.01 China: r = .384, p<.09
Possible Explanations
Projection (Dawes, McTavish & Shaklee 1977)Americans are more likely than Chinese to
project own behavior onto others in situations involving possibility of cooperation.
Avoidance of “Renqin” (King 1991) If you do not wish to be indebted to others,
you do not expect favors from them.
Implications of micalibration
Central premise in trust research: trust is based on positive expectations of reciprocity (Rousseau et al. 1998)Perhaps there are motivations for trust other
than expectations of reciprocity.Or, possibly the strength of expectations as a
motivation for cooperative behavior varies across countries.