An International Investigation of the Influence of Social Distance on Trust LetвЂ™s Get Personal: An International Examination of the Influence of Communication, Culture and Social Distance on Trust and Trustworthiness Nancy R. Buchan вЂ“ Univ. of Wisconsin Rachel T.A. Croson вЂ“ Univ. of Pennsylvania Eric J. Johnson вЂ“ Columbia University Goal of Research пЃ® How does trust manifest itself in different countries and different contexts? п‚Ё Four пЃ® countries China, Japan, Korea, United States п‚Ё Two cultural orientations п‚Ё Social Distance п‚Ё Personal vs. Impersonal Communication Perspective: Experimental Economics пЃ® The Trust Game (Berg, Dickhaut & McCabe 1995) п‚Ё Standard trust game п‚Ё Both parties receive $10 participation fee п‚Ё Amount sent is tripled, responder is a dictator with total п‚Ё One-shot, real money п‚Ё Deviations from equilibrium вЂ“ trust, trustworthiness The Influence of Social Distance пЃ® Need to incorporate social distance to explain individual decisions bearing social consequences (Akerlof 1997) п‚Ё Definition: Location in a social space (Akerlof 1997) пЃ® Degree of isolation from each other and experimenter (Hoffman, McCabe and Smith 1996) пЃ® Demographic similarity (Glaeser et al. 2000) пЃ® Social Distance and Culture пЃ® Minimal Group Paradigm (Tajfel et al. 1971) п‚Ё Arbitrary categorization into mutually exclusive groups п‚Ё Significant ingroup bias in the United States (Orbell, van de Kragt & Dawes 1988; Frey & Bohnet 1997) пЃ® The Influence of Culture on Group Definition and Formation п‚Ё Permanent / Natural groups vs. Temporary / Artificial (Triandis 1995) п‚Ё Likely resistance from collectively oriented participants (Mann, Radford, & Kanagawa 1985) Social Distance and Culture пЃ® Hypothesis (tested at the aggregate and individual levels): п‚Ё Countries (US) / Individuals with relatively more individualist orientation will display an ingroup bias. п‚Ё Countries (China, Japan, Korea) / Individuals with relatively more collectivist orientation will not display an ingroup bias. The Influence of Communication пЃ® Communication increases cooperation rates in SDs (Sally 1995) п‚Ё Why? п‚Ё Must пЃ® communication be strategy-relevant? Irrelevant communication (вЂњcheap talkвЂќ) п‚Ё Roth (1995): 2-min., face to face, social discussion пЃ® Led to same amounts of increase in cooperation as relevant communication п‚Ё Dawes, McTavish, & Shaklee (1977): 10-min., face to face, fact-related discussion пЃ® Led to same amount of cooperation as no communication Irrelevant Content пЃ® Compare 2 types of irrelevant communication п‚Ё Personal п‚Ё Impersonal пЃ® Hypothesis: п‚Ё 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 пЃ® 188 Ss вЂ“ sophomore or junior business or economics majors Ss: Nankai University вЂ“ China п‚Ё 50 Ss: Seoul National University вЂ“ Korea п‚Ё 44 Ss: University of Tokyo вЂ“ Japan п‚Ё 44 Ss: University of Pennsylvania - US п‚Ё 50 Experiment Procedures Communication пЃ® Participants randomly assigned to one of four color-coded groups (~ 12 members) п‚Ё Personal communication: Ss told to introduce themselves and to discuss, from a list provided, a question pertaining to their birthdays п‚Ё Impersonal communication: Ss told to answer from the given list, questions obtained from the world almanac Experiment Procedures вЂ“ Trust Game пЃ® пЃ® Participants randomly assigned to roles (sender / responder), and directed to separate rooms Received written instructions for game, experimental fee, and 2 colored index cards п‚Ё Numbered card represented color of Ss own discussion group, and ID number п‚Ё Blank card represented color of partnerвЂ™s discussion group Experimental Procedures Double-blind пЃ® Controls (Roth et al. 1999): пЃ® п‚Ё Ss pool equivalency п‚Ё Currency effects: 2,000 yen, 1,000 won, 10 yuan, $10 dollars п‚Ё Language effects п‚Ё Experimenter effects п‚Ё Comprehension Experimental Questionnaire Senders: Completed after sending money but before receiving any back пЃ® Responders: Completed before receiving money from senders пЃ® Hypothetical games (next study) пЃ® Measures of Cultural Orientation пЃ® Cultural orientation: Broad measures of culture reflecting two fundamental differences in culture (Hofstede 1980; Triandis et al. 1988) пЃ® Competition vs. Cooperation п‚Ё пЃ® вЂњTo what extent does it feel like you are competing or cooperating with the responder (sender) in this experiment?вЂќ Priority on self vs. group п‚Ё вЂњWhat 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?вЂќ Experimental Results пЃ® Dependent Variables: п‚Ё Trust: Amount sent by sender п‚Ё Trustworthiness: Proportion of total wealth returned by responder пЃ® Monetary amounts across countries standardized on scale from 0 to 1,000 units Amount Sent Independent V ariable E stim ate S t D ev t V alue Intercept 457.45 53.06 13.44 ** C om m unication 278.23 52.93 2.51 * G roups 221.61 53.53 -0.90 C hina 278.67 37.18 1.41 Japan 237.19 38.59 0.62 K orea 69.96 36.59 0.46 C om m *G roups 29.76 26.69 0.20 C om m *C hina -108.99 37.94 1.28 C om m *Japan -142.42 38.07 -0.63 C om m *K orea -3.51 36.55 0.78 G roups*C hina -328.16 38.00 -2.53 G roups*Japan -309.03 38.85 0.97 G roups*K orea -137.41 36.59 -1.00 -81.40 37.30 -0.30 41.20 34.25 -0.53 G ender E conom ics E ducaton A djusted R 2 # of O bserv ations **p< .01 * p< .05 ^ p< .10 0.06 92 ^ * Country x Social Distance Amount Sent (Country by Social Distance) 1000 900 Amount Sent 800 700 600 Ingroup Outgroup 500 400 300 200 100 0 China Japan Korea US Cultural Orientation Independent V ariable E stim ate S t D ev t V alue Intercept 9.1590 1.2200 7.57 ** C hina 1.9140 1.4600 2.05 * Japan 0.6740 1.5200 -0.33 K orea 0.3940 1.4500 -0.31 G roups 0.0060 1.2900 0.59 C om m unication 0.3520 1.2800 -0.25 P roposer 0.3510 1.2800 -0.16 2 A djusted R N um ber of O bserv ations **p< .01 * p< .05 0.12 184 Cultural Orientation x Social Distance Amount Sent (Cultural Orientation by Social Distance) 1000 900 Amount Sent 800 700 600 Ingroup Outgroup 500 400 300 200 100 0 Collectivist Individualist Cultural Orientation Cultural mediation Independent V ariable E stim ate S t D ev t V alue Intercept 460.24 48.23 14.76 ** C om m unication 108.43 47.65 2.28 ** G roups -65.16 47.47 -1.37 -316.65 48.08 -1.98 C hina 28.35 33.85 0.84 Japan 27.76 34.08 0.81 K orea 10.55 32.42 0.33 C ulture*C hina -38.31 33.74 -1.42 C ulture*Japan -20.28 33.91 -0.60 C ulture*K orea -32.31 34.22 -0.94 C om m *C ulture -9.25 24.22 -0.38 G roups*C ulture C ultural O rientation -73.94 34.03 -2.17 C om m *C hina 72.67 33.41 1.18 C om m *Japan -27.20 33.58 -0.81 C om m *K orea 11.89 33.94 0.35 C om m *G roups 3.00 24.19 0.12 G roups*C hina -29.00 24.55 -1.18 G roups*Japan -28.33 34.47 -0.82 G roups*K orea -51.54 32.28 -1.30 G ender -61.25 65.33 -0.60 54.20 33.70 -0.43 E conom ics E ducation 2 A djusted R N um ber of O bserv ations **p< .01 * p< .05 0.29 92 * * Communication вЂ“ Main effect Amount Sent (Communication by Social Distance) 1000 Amount Sent 900 800 700 . 600 500 400 Ingroup Outgroup 300 200 100 0 Personal Communication Impersonal Communication Proportion Returned Independent V ariable E stim ate S t D ev t V alue Intercept 0.1053 0.0369 10.60 ** C om m unication 0.1331 0.0318 1.99 * G roups 0.1512 0.0315 1.30 C hina 0.2140 0.0230 2.08 Japan 0.1155 0.0244 1.34 K orea 0.0587 0.0241 1.35 C om m *G roups 0.1345 0.0158 -2.13 C om m *C hina -0.0544 0.0231 -0.59 C om m *Japan -0.0030 0.0233 0.03 C om m *K orea -0.0162 0.0231 -0.18 G roups*C hina -0.1821 0.0224 -2.03 G roups*Japan -0.0638 0.0232 -0.69 G roups*K orea -0.0481 0.0223 0.54 G ender -0.0985 0.0200 2.46 0.0261 0.0240 -0.21 E conom ics E ducation 2 A djusted R N um ber of O bserv ations **p< .01 * p< .05 0.13 92 * * * ** Country x Social Distance Proportion Returned (Country by Social Distance) 0.50 Proportion Returned 0.45 0.40 0.35 0.30 Ingroup Outgroup 0.25 0.20 0.15 0.10 0.05 0.00 China Japan Korea US Cultural Orientation x Social Distance Proportion Returned (Cultural Orientation by Social Distance) Proportion Returned 0.50 0.40 0.30 Ingroup Outgroup 0.20 0.10 0.00 Collectivist Individualist Cultural Orientation Cultural Mediation Independent V ariable E stim ate S t D ev t V alue Intercept 0.1742 0.0903 1.93 ** C om m unication -0.0413 0.0204 -2.02 ** G roups 0.0244 0.0538 0.45 C ultural O rientation 0.0200 0.0085 2.35 C hina 0.0177 0.0693 0.26 Japan 0.0311 0.0595 0.52 K orea 0.0682 0.0566 1.20 C ulture*C hina -0.0045 0.0045 -1.00 C ulture*Japan -0.0219 0.0214 -1.03 C ulture*K orea 0.0004 0.0211 0.02 C om m *C ulture -0.0140 0.0206 -0.68 G roups*C ulture -0.0208 0.0524 -1.79 C om m *C hina -0.0240 0.0208 -1.16 C om m *Japan -0.0009 0.0200 -0.05 C om m *K orea 0.0003 0.0063 0.05 C om m *G roups -0.0023 0.0059 -0.39 G roups*C hina -0.0266 0.0145 -1.84 G roups*Japan 0.0034 0.0044 0.77 G roups*K orea -0.0001 0.0059 -0.01 G ender 0.0356 0.0185 1.93 E conom ics E duc aton 0.0351 0.0125 0.56 2 A djusted R N um ber of O bserv ations **p< .01 * p< .05 ^ p< .10 0.41 92 ** ^ ** Effect of Communication Proportion Returned (Communication by Social Distance) 0.50 Proportion Returned 0.45 0.40 0.35 0.30 Ingroup 0.25 Outgroup 0.20 0.15 0.10 0.05 0.00 Personal Communication Impersonal Communication Discussion пЃ® The interaction of cultural orientation with social distance п‚Ё Culturally 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) п‚Ё Implications for research methods used Discussion пЃ® пЃ® The content of non-strategy relevant communication influences the level of trust displayed 3 explanations for this result: п‚Ё Social пЃ® Decreasing social distance = higher trust п‚Ё Social пЃ® Distance (Hoffman, McCabe & Smith 1996) Identity (Brewer 1991; Tyler and Dawes 1993) Group identity increases expectations that members of ingroup will return money Discussion пЃ® Proposed explanation for communication effect п‚ЁPersonal communication prompts greater вЂњotherregardingnessвЂќ (Bohnet & Frey 1999) Examination of outgroup results reveals вЂњcarryoverвЂќ effect of communication пЃ® A representative heuristic at play? (Kahneman & Tversky пЃ® 1972) Conclusion вЂњCheap TalkвЂќ isnвЂ™t so cheap пЃ® Our results regarding trust and trustworthiness concur with Schelling (1968) who suggested, the more we know, the more likely we are to care. пЃ® The Boundaries of Trust: Own and OthersвЂ™ Actions in the US and China Nancy R. Buchan, University of Wisconsin Rachel T.A. Croson, University of Pennsylvania Exploring the Boundaries of Trust пЃ® Putting FukuyamaвЂ™s ideas to test (1995) levels vary by country вЂ“ US > China п‚Ё Location of trust boundary varies by country п‚Ё Trust пЃ® пЃ® пЃ® US boundary is wide вЂ“ including non-kin China boundary is close and tight вЂ“ limited to family members Trust: вЂњвЂ¦the expectation of cooperative behavior based on commonly shared norms.вЂќ Trust Games with Hypothetical Targets пЃ® Examining trust and trustworthiness within naturally occurring social networks пЃ® 7 targets varying in social distance (similar to Kollock 1998) п‚Ё п‚Ё п‚Ё п‚Ё п‚Ё п‚Ё п‚Ё Parent Sibling Cousin Student know well Student from another university Stranger from hometown Stranger from another country Intentions and Expectations пЃ® 2 types of measures collected: п‚Ё п‚Ё пЃ® пЃ® Intention Expectation п‚Ё Senders: вЂњImagine 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.вЂќ п‚Ё Responders: вЂњImagine 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.вЂќ Measures elicited during play of actual trust game Focusing on China and the US since they represent FukuyamaвЂ™s starkest predictions. Social Distance Hypothesis пЃ® 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 nonfamily members will be significantly greater than the drop in trust and trustworthiness among American participants. Social Norm Hypothesis пЃ® Ostrom (2003): вЂњOnce some members of a population acquire norms of behavior, the presence of these norms affect the expectations of other players,вЂќ (p.41). пЃ® 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. пЃ® Method of Analysis пЃ® пЃ® пЃ® пЃ® Mixed-model repeated ANOVA Amounts sent standardized scale 1-10 units Converted absolute amounts returned to proportions returned. Covariates: п‚Ё Exposure to game theory п‚Ё Gender вЂ“ replicating prior results (Croson and Buchan AER 1999, previous study) Comparison with Actual Game пЃ® пЃ® Actual game: participants paired with anonymous student from same university Hypothetical: Between вЂњStudent you know well,вЂќ and вЂњStudent from another universityвЂќ Amount P ro p o rtio n Sent R e tu rn e d A c tu a l 4 .9 5 22% S tu d e n t K n o w 7 .3 1 35% 4 .0 4 22% A c tu a l 6 .6 2 35% S tu d e n t K n o w 8 .1 2 35% 5 .4 1 30% US W e ll S tu d e n t A n o th e r U n iv e rs ity C h in a W e ll S tu d e n t A n o th e r U n iv e rs ity Amount Sent вЂ“ SenderвЂ™s Intentions As social distance increases, amounts sent decrease (F(7,308)=70.17, p<.01) пЃ® ChinaвЂ™s tighter boundary of trust? п‚Ё No social distance x target interaction F ig u re 2 : Am o u n t P ro p o s e rs W o u ld S e n d Ac ro s s C o u n trie s 10 9 8 7 A mount пЃ® 6 5 4 3 2 1 0 P a re n t S ib lin g C o u sin S tu d e n t Y o u K n e w W e ll S tu d e n t fro m A n o th e r U n ive rsity Responder C hina (n=24) U nited S tates (n=22) S tra n ge r fro m Hom e Town S tra n ge r fro m A n o th e r C o u n try Amount Sent вЂ“ Responder's Expectations пЃ® As social distance increases, expectations decrease (F(7,315)=36.6, F ig u re 3 : Am o u n t R e s p o n d e rs E x p e c t to R e c e ive Ac ro s s C o u n trie s 10 9 8 7 p<.01) пЃ® Significant country x recipient interaction (F(7,315)=10.64, p<.01) A mount 6 5 4 3 2 1 0 P a re n t S ib lin g C o u sin S tu d e n t Y o u K n e w W e ll S tu d e n t fro m A n o th e r U n ive rsity P ro p o s e r C hina (n= 25) U nited S tates (n= 22) S tra n ge r fro m Hom e Town S tra n ge r fro m A n o th e r C o u n try Proportion Returned вЂ“ ResponderвЂ™s Intentions пЃ® As social distance increases, amounts sent F ig u re 4: P ro p o rtio n R esp o n d ers W o u ld R etu rn Acro ss C o u n tries 1.00 0.90 0.80 decrease (F(7,315)=20.06, пЃ® ChinaвЂ™s tighter boundary of trust? п‚Ё Again, no social distance x target interaction 0.60 A m ount p<.01) 0.70 0.50 0.40 0.30 0.20 0.10 0.00 P arent S ibling C ousin S tudent Y ou K new W ell S tudent from A nother U niversity P ro p o ser C hina (n=25) U nited S tates (n=22) S tranger from S tranger from H om e T ow n A nother C ountry Proportion Returned вЂ“ SenderвЂ™s Expectations Significant country x recipient interaction 1 .0 0 (F(7,308)=8.72, p<.01) 0 .7 0 F ig u re 5: P ro p o rtio n P ro p o sers W o u ld E xp ect T o H ave R etu rn ed Acro ss C o u n tries 0 .9 0 0 .8 0 Am ount пЃ® 0 .6 0 0 .5 0 0 .4 0 0 .3 0 0 .2 0 0 .1 0 0 .0 0 P a re n t S ib lin g C o u sin S tu d e n t You Knew W e ll S tu d e n t fro m A n o th e r U n ive rsity R esp o n d er C hina (n=24) U nited S tates (n=22) S tra n g e r fro m Hom e Town S tra n g e r fro m A n o th e r C o u n try 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. пЃ® Analysis: Repeated measures anovas using intentions and expectations as the DV, role as the IV (for both sending and returning) пЃ® US Comparison Returning 10.00 1.00 9.00 0.90 8.00 0.80 7.00 0.70 Amount Amount Sending 6.00 5.00 4.00 3.00 0.60 0.50 0.40 0.30 2.00 0.20 1.00 0.10 0.00 Parent Sibling Cousin Student You Knew Well Student from Another University Stranger from Home Tow n Stranger from Another Country 0.00 Parent Sibling Cousin Student You Knew Well Recipient Student from Another University Stranger from Home Tow n Stranger from Another Country Recipient Amount Proposers Would Send Amount Responders Expect to Receive Proportion Proposers Would Expect to Have Returned to Them Proportion Responders Would Return US Comparison пЃ® пЃ® пЃ® пЃ® No significant effect of role for either sending or returning вЂ“ thus the levels of lines are the same. Significant effect of counterpart for both sending and returning вЂ“ thus the slopes of the lines are significant and negative. Interactions are not significant вЂ“ the slopes of the lines are the same as each other. Strong relationship between intended behaviors and expectations in US. China - Comparison Returning 10.00 1.00 9.00 0.90 8.00 0.80 7.00 0.70 6.00 0.60 Amount Amount Sending 5.00 4.00 0.50 0.40 3.00 0.30 2.00 0.20 1.00 0.10 0.00 0.00 Parent Sibling Cousin Student You Knew Well Student from Another University Stranger from Home Town Stranger from Another Country Recipient Amount Proposers Would Send Amount Responders Expect to Receive Parent Sibling Cousin Student You Knew Well Student from Another University Stranger from Home Town Stranger from Another Country Recipient Proportion Proposers Would Expect to Have Returned to Them Proportion Responders Would Return China Comparison Significant interactions for both sending and returning вЂ“ the slopes of the lines are the different from each other. пЃ® Intended actions are more responsive to social distance than expectations. пЃ® Weak relationship between intended actions and expectations in China. пЃ® Discussion пЃ® пЃ® Lack of support for FukuyamaвЂ™s tighter boundary of trust in China Strong support for influence of social distance in both countries among natural groups п‚Ё пЃ® Counterpoint to findings in previous studies with experimental groups Stresses the importance of (embeddedness in) personal networks of relationships in building trust (Granovetter 1985) п‚Ё Trust in China can occur outside kin (despite вЂњbamboo networkвЂќ) пЃ® 3 categories of interpersonal relationships based on expressive, instrumental, or mixed ties (Bond and Hwang 1995) The connection between own and otherвЂ™s actions US: RespondersвЂ™ expectations about proposersвЂ™ reactions to social distance are consistent with proposersвЂ™ actual reactions, and vice versa. пЃ® China: Expectations are systematically less consistent. пЃ® п‚Ё Actors are sensitive to social distance, but neither expects the other side to be. Role of Norms пЃ® Apriori expectations: п‚Ё In collectivist cultures strict norms governing behavior in different contexts (Bond & Hwang 1986); adherence to norm demanded (Bierbrauer et al. 1994). п‚Ё In individualist cultures, variety of norms is tolerated (Bierbrauer et al. 1994). пЃ® Test of variances yields no distinct pattern Comparison of IndividualвЂ™s Own Actions with His Expectations пЃ® Correlation analysis: What a proposer says he will send to each counterpart vs. what he expects to receive from each. Parent Sibling C ousin Student You Knew W ell Student from Another U niversity Stranger from H om e T own Stranger from Another C ountry U nited States 0.336 0.709 0.815 0.764 0.775 0.808 0.833 C hina 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. пЃ® US: r = .794, p<.01 пЃ® China: r = .384, p<.09 пЃ® Possible Explanations пЃ® Projection (Dawes, McTavish & Shaklee 1977) п‚Ё Americans are more likely than Chinese to project own behavior onto others in situations involving possibility of cooperation. пЃ® Avoidance of вЂњRenqinвЂќ (King 1991) п‚Ё If you do not wish to be indebted to others, you do not expect favors from them. Implications of micalibration пЃ® Central premise in trust research: trust is based on positive expectations of reciprocity (Rousseau et al. 1998) п‚Ё Perhaps there are motivations for trust other than expectations of reciprocity. п‚Ё Or, possibly the strength of expectations as a motivation for cooperative behavior varies across countries.