Does Travel Broaden the Mind? Breadth of Foreign Experiences Increases Generalized Trust

Breadth Versus Depth of Foreign Experiences

Consistent with Mark Twain’s assertion, we propose that foreign travel experiences may serve as one way to facilitate an increase in generalized trust. However, given the difficulty for intergroup contact to increase a generalized sense of trust, we suggest that it is important to differentiate between two distinct aspects of foreign experiences: the depth of experience (i.e., the length of time one spends abroad) and the breadth of experience (i.e., the number of countries one has visited). While depth has been shown to be associated with adapting to a local culture and subsequent enhanced creativity (Maddux & Galinsky, 2009), breadth captures the diversity of foreign experiences, an aspect of multicultural experiences that has received very little empirical attention thus far.

We propose that spending time in a foreign environment may be a critical experience that increases generalized trust, but especially when those foreign experiences involve broad experiences that afford the opportunity to engage with a variety of foreign individuals rather than individuals from just one specific group (i.e., deep experiences). Indeed, one possible reason for the inability of some intergroup contact experiences to increase a sense of generalized trust may be that most studies involved exposure to only one type of out-group. However, experiences that allow for contact with many different cultural or ethnic groups may increase the likelihood that one’s impressions derived from interactions with these different groups will be generalized and applied to other groups and people. These opportunities for contact with diverse groups are likely to occur during broad foreign experiences that take place across several countries. Deeper foreign experiences, on the other hand, such as extended time in a single country, may be less likely to lead to the diversity of exposure that is necessary to produce a generalized effect.

We propose that the breadth of foreign experiences may be particularly important for facilitating generalized trust because breadth provides the variety and diversity of experiences that are necessary to produce generalizations and learning. Indeed, Kelley’s classic analysis of variance attribution theory (Kelley, 1967) highlights the necessity of variance as being a critical factor when forming a generalized attribution of a target. For example, repeated information from dissimilar sources has been shown to be more valuable than information from similar sources in forming impressions (Himmelfarb, 1972). Evidence from the intergroup contact literature also supports the importance of breadth of experiences in leading to generalizability. For example, neighborhood ethnical diversity is associated with lower bias toward a number of different out-groups (Schmid, Hewstone, & Al Ramiah, 2013). Similarly, a diverse set of intergroup contact experiences (e.g., race, religion, nationality, culture, social class), rather than deep intergroup contact with just one category, predicts more favorable intergroup attitudes in general (Pettigrew, 1997). Further support for our hypotheses comes from studies showing that generalized trust is higher in social contexts that have greater social mobility (Macy & Sato, 2002; Yamagishi & Yamagishi, 1994), which provide more opportunities to interact with a diverse set of unfamiliar individuals compared to lower mobility contexts, where people mostly interact with known others. Overall, then, we predicted that broad rather than deep experiences within foreign environments would be more likely to produce generalized trust.

Overview

We conducted five studies to test the prediction that the breadth more than the depth of foreign experiences will increase generalized trust. In Study 1, we tested whether the breadth (i.e., the number of foreign countries one has traveled to) more than the depth (i.e. the length of time one has traveled abroad) of foreign experiences predicts behavior in the trust game (Berg, Dickhaut, & Mccabe, 1995). In Study 2 and Study 3, we established a causal relationship by directly manipulating a focus on broad or deep foreign experiences prior to a trust game (Study 2) and the generalized trust scale (Study 3; Yamagishi & Yamagishi, 1994). Study 4 used a longitudinal design and the generalized trust scale to assess people’s generalized trust before (Time 1) and after (Time 2) traveling abroad; this design allowed us to capture whether the number of countries participants traveled to during their trip predicted increases in generalized trust from Time 1 to Time 2. In Study 5, we directly tested our hypothesis that a diversity of experiences is critical by manipulating a difference or similarity focus to explore whether a difference focus increased generalized trust.

Study 1

Method

Results

The correlations between all the variables are presented in Table 1. A regression model that included only the breadth and the depth of foreign travel experiences revealed that breadth predicted the amount of money sent in the trust game, B = .17, standard error (SE) = .08, β = .16, p = .03, but the effect of depth was not significant, B = .00, SE = .00, β = −.10, p = .20 (see Model 1, Table 2). This effect of breadth held even when controlling for demographic and personality variables, B = .16, SE = .08, β = .16, p = .04; whereas the effect of depth was still not significant, B = −.00, SE = .00, β = −.08, p = .31 (see Model 2, Table 2). We also explored the interaction effect by adding the interaction term of breadth and depth and it was not significant, p = .80.

OPEN IN VIEWER

Table 1.

Correlation Table for All Variables, Study 1.

	Mean	SD	1	2	3	4	5	6	7	8	9	10	11
1. Money sent in trust game	4.23	3.33	—
2. Breadth	4.05	3.15	.11†	—
3. Depth	10.53	15.28	−.02	.48**	—
4. Gender (0 = male, 1 = female)	1.60	.50	−.14*	−.02	.12	—
5. Age	19.59	1.30	−.10	.12	.04	−.15**	—
6. Ethnicity group (0 = non-White, 1 = White)	.48	.50	.07	.01	−.05	−.11†	.15*	—
7. Extroversion	4.74	1.35	.07	.07	.03	.04	−.06	.07	—
8. Conscientiousness	5.31	1.16	−.13†	−.07	−.03	.08	.13	.02	−.10	—
9. Agreeableness	4.72	.99	.06	−.01	.00	.00	.04	−.02	.20**	−.07	—
10. Neuroticism	3.16	1.30	−.07	−.08	−.07	.16*	−.11†	−.01	−.90	−.10	−.07	—
11. Openness	5.35	1.08	.01	.14*	.05	−.00	.01	−.04	.38**	−.04	.14*	−.09	—

Note: SD = standard deviation. N = 237.

^† p < .10. ** p < .01. * p < .05.

OPEN IN VIEWER

Table 2.

Personality/Demographic and Foreign Travel Experiences Predictors of Money Sent in a Trust Game (N = 237), Study 1.

Independent Variables	Model 1	Model 2
Breadth	.17* (.08)	.16* (.08)
Depth	−.003 (.03)	−.00 (.00)
Gender (0 = male, 1 = female)		−.83† (.45)
Age		−.35* (.17)
Ethnicity group (0 = non-White, 1 = white)		−.83 (.45)
Extroversion		.05 (.09)
Conscientiousness		−.14 (.10)
Agreeableness		−.12 (.11)
Neuroticism		−.08 (.09)
Openness		−.06 (.11)
Constant	3.77** (.35)	12.64** (3.83)

Note. The table represents unstandardized regression coefficients, with standard errors in parentheses.

^† p < .10. ** p < .01. * p < .05.

We conducted several additional robustness checks. First, to test for the effect of outliers, we identified outliers using studentized deleted residual greater than 3; no outliers were identified. We also used Cook’s Distance as the outlier criterion, with the critical value at 0.01688 (4/N). Nine outliers were identified, but importantly, breadth still predicted the amount of money sent in the trust game after excluding these outliers, B = .18, SE = .08, β = .16, p = .03. Second, we log transformed the breadth and depth data to reduce skewness (adding 1 before the transformation to eliminate 0 values). The effect of breadth still held, B = 2.54, SE = 1.28, β = .23, p = .048. Third, we explored whether there was a nonlinear relationship between breadth and money sent; the quadratic term was not significant, p = .57, suggesting that a linear effect is a better representation for the relationship between breadth and the amount of money sent in the trust game.

Study 2

Method

Results

Participants who recalled a broad foreign travel experience (M = 6.21, SD = 3.43) sent significantly more money in the trust game than those who recalled a deep travel experience (M = 4.33, SD = 2.82), t(49) = 2.08, p = .04, d = .60, providing causal evidence for the role of breadth of foreign experiences in the development of greater generalized trust.

Study 3

Method

Results

Participants who recalled broad foreign travel experiences (M = 5.01, SD = .83) had significantly higher scores in the generalized trust survey than those who recalled a deep foreign travel experience (M = 4.68, SD = .87), t(115) = 2.10, p = .04, d = .39. Thus, Study 3 showed the robustness of the relationship by using a different manipulation and generalized trust measure.

We obtained the country-level trust scores from the World Values Survey (WVS)2 for the countries participants listed. Two participants were dropped off for this analysis because trust scores for the countries they listed (e.g. Bahamas, Nicaragua) were missing in the WVS. We tested whether the country-level trust scores moderated the effect of condition (breadth vs. depth) on generalized trust. We averaged the country-level trust ratings of the three countries in the breadth condition. We entered condition and country-level trust on the first step and added in their interaction on the second step to predict generalized trust. Breadth still predicted generalized trust, B = .32, SE = .15, β = .19, p = .04. Importantly, the interaction was not significant, B = −.01, SE = .01, β = −.32, p = .33.

We also tested whether participants recalled countries with different country-level trust depending on condition. The countries that participants recalled in the breadth and the depth conditions did not differ in country-level trust, t(113) = .86, p = .39. Thus, the result suggested that the effect is independent of country-level trust scores.

Study 4

Method

Results

The correlations of all the variables are presented in Table 3. We first ran a regression model including only the breadth and the depth of foreign travel experiences as predictors of increases in generalized trust from Time 1 to Time 2 (Time 1 generalized trust subtracted from Time 2 generalized trust). Breadth predicted increases in generalized trust, B = .14, SE = .06, β = .18, p = .02, but the effect of depth was not significant, B = −.005, SE = .005, β = −.088, p = .27 (see Model 1, Table 4). Next, we conducted a second regression model controlling for demographic and personality variables. Again, breadth predicted increases in generalized trust, B = .14, SE = .06, β = .18, p = .02, whereas depth did not, B = −.006, SE = .005, β = −.10, p = .19 (see Model 2, Table 4). We also examined our predictions by using generalized trust at Time 2 as the dependent variable and generalized trust at Time 1 as a control variable. Again, breadth predicted generalized trust at Time 2 after controlling for generalized trust at Time 1, personality, and demographic variables, B = .12, SE = .06, β = .14, p = .04, but depth did not, p = .25. We also explored the interaction effect by adding the interaction term of breadth and depth and it was not significant, p = .76.

OPEN IN VIEWER

Table 3.

Correlation Table for All Variables, Study 4.

	Mean	SD	1	2	3	4	5	6	7	8	9	10	11	12	13
Generalized trust change	.22	.79	—
Breadth	1.24	1.01	.13*	—
Depth (Days)	10.82	12.77	.02	.60**	—
Generalized trust (T1)	5.38	.80	−.48**	−.05	−.02	—
Generalized trust (T2)	5.60	.81	.50**	.07	.00	.52**	—
Gender (0 = male, 1 = female)	1.68	.47	−.08	−.10	−.01	.05	−.03	—
Age	28.37	5.79	−.08	.10	.09	.13*	.05	.03	—
SES	5.30	1.32	−.17**	.04	−.02	.21**	.04	.00	.02	—
Extroversion	4.20	1.24	.00	.00	−.09	.12†	.12†	.12†	.08	.09	—
Conscientious	5.15	1.05	−.09	.05	.04	.16*	.07	−.09	.24**	.03	.04	—
Agreeableness	5.32	.96	−.19**	−.09	−.08	.28**	.09	.07	.11†	.06	.02	.02	—
Neuroticism	3.26	1.08	.18**	−.01	.11†	−.22**	−.05	.14*	−.21**	−.16*	−.29**	−.38**	−.39**	—
Openness	5.01	1.10	−.09	−.05	−.06	.14*	.04	−.10	−.06	.07	.48**	.13*	.08	−.37**	—

Note. SES = socioeconomic status; SD = standard deviation. N = 245.

** p < .01. * p < .05. ^† p < .10.

OPEN IN VIEWER

Table 4.

Personality/Demographic and Foreign Travel Experiences Predictors of Increases in Generalized Trust from Time 1 to Time 2 (N = 245), Study 4.

Independent Variable	Model 1	Model 2
Breadth	.14* (0.06)	.14* (0.06)
Depth	−.005 (0.005)	−.006 (0.005)
Gender (0 = male,1 = female)		−.15 (0.11)
Age		−.008 (0.009)
SES		−.09* (0.04)
Extroversion		.06 (0.05)
Conscientiousness		−.004 (0.05)
Agreeableness		−.10† (0.06)
Neuroticism		.08 (0.06)
Openness		−.06 (0.05)
Constant	.22** (.05)	.47* (.19)

Note. SES = socioeconomic status. The table represents unstandardized regression coefficients, with standard errors in parentheses.

^† p < .10. ** p < .01. * p < .05.

We conducted several robustness checks. First, we tested whether there were outliers driving the result. We used studentized deleted residual greater than 3, which identified three outliers, and Cook’s Distance greater than the critical value at 0.0163 (4/N), which identified 17 outliers. Breadth still predicted increases in generalized trust after eliminating the studentized deleted residual outliers, B = .18, SE = .06, β = .26, p < .01, and the Cook’s Distance outliers, B = .19, SE = .06, β = .30, p < .01. Second, we log transformed the breadth and depth data to reduce skewness (adding 1 before the transformation to eliminate 0 values); breadth still marginally predicted increases in generalized trust, B = .84, SE = .52, β = .20, p = .10, and the effect became significant if we eliminated the outliers in the analysis, B = 1.17, SE = .47, β = .31, p = .01, for outliers identified by studentized deleted residual; B = .98, SE = .45, β = .29, p = .03 for outliers identified by Cook’s Distance. Third, we explored whether there was a nonlinear relationship between breadth and increases in generalized trust. The quadratic term for breadth was not significant, p = .12, suggesting that a linear effect is a better representation for the relationship between breadth and increases in generalized trust.

We also explored whether the means of the country-level trust scores moderated the relationship between breadth and increases in generalized trust. Given that the data of many visited countries was missing in the WVS, we had 209 data points available in this analysis. The interaction term was not significant, B = .001, SE = .003, β = .07, p = .77, demonstrating that the country-level trust scores did not influence the effects.

Study 5

Method

Results

Participants who focused on the differences among broad foreign travel experiences (M = 5.23, SD = .75) had significantly higher generalized trust than those who focused on the similarities (M = 4.75, SD = .95), t(61) = 2.21, p = .03, d = .56.

Having participants focus on the diversity of their foreign experiences increased their generalized trust. The experiment provides supports for our proposed mechanism—the diversity of experiences provided by broad foreign experiences plays a key role in increasing generalized trust because diversity is essential for the generalization process.

Discussion

Across five studies, regardless of the types of research method, the cultural samples and operationalizations of generalized trust, we consistently found a robust relationship between the breadth of foreign travel experiences and increases in generalized trust. Our longitudinal study and experiments provide causal evidence that broad foreign experiences led to greater generalized trust. Our final experiment offered direct evidence for the idea that focusing on the diversity of one’s foreign experiences increases generalized trust.

Despite the importance of generalized trust in interpersonal interactions, most research on generalized trust comes from economics, sociology, and political science, which focus on macro-environmental factors, such as income inequality (Neville, 2012), wealth (Delhey & Newton, 2005), and corruption (Rothstein & Uslaner, 2005). Our research offers an individual and developmental perspective by showing the impact of foreign experiences on generalized trust.

A critical contribution of the current article is that it makes a novel distinction between the breadth and depth of foreign experiences. Globalization has given birth to a host of research on the psychological effects of foreign experiences, most of which has only investigated the effect of deep foreign experiences (e.g., Maddux & Galinsky, 2009). Less research has addressed the distinctive role of breadth of foreign experiences, either within the multicultural experiences literature or within the intergroup contact literature, and the differential psychological benefits provided by broader versus deeper experiences.

The distinction between breadth and depth is critical in practice, because the previous focus on the depth of intergroup contact (i.e., repeated interactions with the same individuals or across many individuals but from the same group) has led to policy prescriptions that often emphasize the depth of interactions, such as intergroup roommates and friendships (Pettigrew & Tropp, 2006). However, these deeper experiences within a single out-group, while helpful for future interactions toward that particular out-group, may also result in limited generalizability toward other groups and individuals. Our research suggests the importance of broad diverse experiences in creating a high trust environment across a potentially larger number of groups and individuals.

Across our studies, we never found significant main effects for depth or a significant interaction between depth and breadth. One open question is what role depth of foreign experiences might play in the development of generalized trust. For example, we found that breadth and depth were highly correlated, suggesting that both experiences are likely to be important. Indeed, depth of experience likely provides the time and opportunity for intergroup contact that then allows for breadth to drive the generalization process. Thus, it may be that some threshold of time spent in different countries is initially important, with further increases in breadth subsequently becoming the critical factor once a threshold of depth is reached. Indeed, we would expect that foreign experiences that involve very brief visits with minimal contact with others, even if across many countries, are unlikely to result in the positive gains in generalized trust that we have found here. Nevertheless, it does seem that the typical broad foreign experience, at least as experienced by participants in our samples, does involve enough contact to affect the generalization process. Future research should do more to explore the role of depth in the development of generalized trust, investigating whether a threshold level of depth is required and what that threshold might be.

One remaining question is to what extent broad foreign experiences overlap with the concept of residential mobility (e.g., Oishi, 2010). It is important to note that high residential mobility does not necessarily imply broad experiences (i.e., several repeated moves within one city or state would be considered residentially mobile but would not imply breadth of experiences). Future research could explore whether broad domestic experiences within a single country lead to higher generalized trust.

Future research should also continue to explore how people learn from experiences and develop generalized expectations toward others. Negative attitudes toward target groups persist because negative expectations create avoidance and reduced contacts (Fazio, Eiser, & Shook, 2004). This is especially problematic when these expectations are false but are never subject to disconfirmation because of avoidance (Fazio et al., 2004). Broad foreign experiences may serve to disconfirm negative expectations regarding the general trustworthiness of others. Although it is certainly the case that not all foreign travel experiences will be positive, individuals who travel broadly are more likely than those without such experiences to have at least some negative expectations disconfirmed.

The current research provides support for study abroad programs and expatriate assignments in organizations, but with a twist—seeing more of the world may be as or more important than spending a longer period of time seeing less of it.