Innovating at Cultural Crossroads: How Multicultural Social Networks Promote Idea Flow and Creativity

Multicultural Networks as Pipes and Prisms

A core feature of social networks is that they serve as conduits or “pipes” for the flow of valued resources between actors (Borgatti & Halgin, 2011; Levine & Cross, 2004; Podolny, 2001). Prior research has argued that diversity in networks is a source of nonredundant novel information, potentiating creative performance (Burt, 2004; Perry-Smith, 2006; Rodan & Galunic, 2004). However, new research on access to diverse informational resources revealed a more complex picture on how it affects creativity. Perry-Smith (2014) argued that the type of informational resources matter. Differentiating between information (i.e., data, ideas, and knowledge) and frames (i.e., perspectives and interpretations), Perry-Smith (2014) found that nonredundant frames facilitated creativity more effectively than nonredundant information because nonredundant frames encourage expansive thinking by providing different ways to look at a problem. Additionally, where these informational resources come from matters—content derived from weak ties facilitated creativity in general but strong ties appear to enhance creativity only when nonredundant frames are received. It should be noted that Perry-Smith (2014) differentiated between information and frames using experimental manipulations. The distinction between information and frames, however, can be harder to make when the flow of informational resources occurs naturally. For example, a network contact may provide a new piece of information or new idea that triggers a new frame or perspective on a problem. Nevertheless, these findings highlight the importance of considering the content and sources of informational resources in social networks when examining their effects on creativity.

What informational resources does a diverse network bring? Diversity in networks can come in different forms. For example, networks can be diverse in terms of gender composition, functional expertise, cultural backgrounds, and so forth. Current research assumes that diversity in network contacts’ background provides diverse informational resources, without specifying the content of these informational resources. Yet network research has long recognized that networks bring highly specific forms of resources (e.g., Borgatti & Halgin, 2011; Chua, Ingram, & Morris, 2008; Cross & Sproull, 2004). Specifically, different types of ties carry different forms of resources, ranging from social support, to task advice, to financial aid (Chua et al.). Some researchers have also investigated specific forms of networks by focusing on the flow of particular resources in network ties, for example, friendship networks (e.g., Gibbons, 2004; Gibbons & Olk, 2003) and advice networks (e.g., Gibbons; McGrath, Vance, & Grey, 2003). Thus, it behooves organizational scholars to be more specific about the kind of informational resources that diverse networks bring.

This study seeks to further specify the content of information that is received from culturally diverse networks. Culture as a form of diversity is a source of nonredundant information (Nouri, Erez, Rockstuhl, Ang, Leshem-Calif, & Rafaeli, 2013). When one’s network comprises contacts from different cultural backgrounds, one is more likely to receive informational resources that stem from different cultural contexts and traditions. For example, if an American has a network contact who is Indian, he or she is likely to receive information or ideas that pertain to Indian culture. Thus, social exchanges with individuals from different cultures in one’s network should give one access to a specific form of informational resources: culture-related novel ideas—ideas from other cultures that are unfamiliar to oneself.

However, is it necessarily the case that culture-related novel ideas are derived only from culturally different social contacts? Might individuals from one’s own culture also be a source of novel ideas about other cultures? Research on network diversity often implicitly assumes that novel ideas primarily arise from network contacts who are different from oneself (M. Baer et al., 2015; Fleming, Mingo, & Chen, 2007; Perry-Smith, 2006; Rodan & Galunic, 2004). Challenging this assumption, I propose that people with multicultural networks are likely to receive culture-related novel ideas via members of their networks, but the sources of such ideas are not necessarily limited to those who are culturally different. A culturally diverse network signals to observers that one is open-minded and highly receptive to foreign cultures, encouraging even same-culture others to share culture-related novel ideas with him or her. To give an example, an American who has culturally diverse business associates and coworkers gives the impression that he or she is worldly, cosmopolitan, and open to diverse foreign cultures; this impression, in turn, increases the likelihood that an American coworker recently returned from Cambodia would share insights gained from the trip with him or her.

Theoretical foundations for the argument that one’s social ties convey information about oneself can be traced back to different lines of psychological theories. For example, Heider’s (1958) balance theory argued that someone who is perceived to be the friend of a friend is also likely to be perceived positively because doing so achieves cognitive balance in the triad. Social psychologists have argued that people make known their social connections with positive sources because observers of these connections would evaluate them positively as well (Cialdini, Borden, Thorne, Walker, Freeman, & Sloan, 1976). Specifically, Cialdini’s (1989) research on basking in reflected glory shows that, to enhance their public image, people actively seek to display social connections to successful others. Additionally, research on social identity theory argued that individuals’ social connections shape how others perceive them, influencing interpersonal behaviors and interactions (Brewer & Gardner, 1996; Tajfel, Billig, Bundy, & Flament, 1971).

In social network research, scholars have similarly recognized that networks serve not only as conduits for resource flow but also as a signaling mechanism that conveys important information about the actor (Baum & Oliver, 1992; Kilduff & Krackhardt, 1994; Krackhardt & Kilduff, 1999; Podolny, 2001). Podolny argued that firms’ network configurations reflect informational cues about their status in the market, signaling the relative quality of their products. Kilduff and Krackhardt found that being perceived to have a prominent friend in an organization would boost one’s performance reputation. Recent experimental research found that the racial diversity of White students’ friendship networks was used by Black students to form expectations about how they would be perceived (Wout, Murphy, & Steele, 2010). Black students expect to be perceived more negatively by White students whose networks are culturally homogeneous (i.e., all White) and expect more interpersonal challenges in upcoming interactions with such students compared to White students with more culturally diverse social networks. More specific to the current thesis, research by Jang, Ramarajan, and Polzer (2010) found that people with culturally diverse online social networks are perceived to be more creative.

Drawing on this body of research, I argue that a multicultural network does not merely serve the straightforward function of providing conduits to receive new ideas from culturally different others (i.e., pipes) but also provides perceptual cues about individuals’ personal characteristics and propensities (i.e., prisms reflecting internal characteristics). A person’s network contacts may form impressions of his or her network cultural diversity via multiple pathways. One pathway is through observations—contacts observed that this person frequently socialized or collaborated with many culturally diverse others. Another pathway is through direct interaction with this person who reveals, in one way or another, his or her associations with culturally different others. Finally, contacts may personally know other culturally diverse contacts in this person’s network.

Drawing on social identity theory (Brewer & Gardner, 1996; Tajfel et al., 1971), I posit that this impression about one’s network cultural diversity, regardless of how it was formed, would shape how one is being perceived (in this case, as being cosmopolitan and open to foreign ideas) and encourage network contacts from the same cultural backgrounds as oneself to share culture-related ideas they have acquired elsewhere. Why would a same-culture network contact share culture-related novel ideas with a focal individual who is perceived to be cosmopolitan? Social network research has emphasized that ties need to be maintained after they are established (Borgatti & Halgin, 2011; Rivera, Soderstrom, & Uzzi, 2010; Uzzi & Dunlap, 2005). One way to maintain ties is to engage in repeated interactions that sustain the relationship (Chen, 2007; Rivera et al.). Thus, network contacts might be motivated to maintain or even strengthen their ties with focal individuals by sharing ideas that appeal to those individuals’ perceived interests (Byrne, 1961; Moreland & Zajonc, 1982; Wang & Noe, 2010). Specifically, knowing that individuals are cosmopolitan and interested in diverse cultures, network contacts would likely discuss with those individuals culture-related novel ideas that they have received from their own cross-cultural experiences. In these exchanges, novel ideas rather than old ideas about other cultures are more likely to be shared; this is because novel ideas can capture attention, generate excitement, and pique the interests of the other party (Olson & Evans, 1999; Van Swol & Ludutsky, 2007), helping to sustain the existing relationship, in part, by reducing relational boredom (Aron, Norman, Aron, McKenna, & Heyman, 2000; Canary, Stafford, Hause, & Wallace, 1993; Harasymchuk & Fehr, 2010).

Taken together, the above arguments suggest that, at a network level of analysis, the more culturally diverse one’s network is, the greater the total number of culture-related novel ideas one is likely to receive from the network. However, the sources of these ideas are not solely limited to culturally different social contacts in the network. Thus, at a dyadic level of analysis, I expect that these novel cultural ideas may come from either culturally different or identical social contacts.

Effects of Network Cultural Diversity on Creativity

The more one’s network includes individuals from different cultural backgrounds, the more one has access to and is stimulated by ideas and perspectives different from one’s own. The result should be greater creativity. However, it is unclear whether the positive effects of a culturally diverse network on creativity applies to all forms of tasks or only to specific types of tasks that require knowledge of multiple cultures. To date, research that investigates the effects of networks on creativity has been silent about the type of task. Many studies used general creative performance ratings by supervisors or observers (e.g., M. Baer, 2010; Perry-Smith, 2006; Rodan & Galunic, 2004; Zhou, Shin, Brass, & Zhang, 2009), leaving open to interpretation the type of tasks that evaluators had in mind in the rating process. Other studies used expert ratings on ideas generated by study participants in context-specific tasks (Burt, 2004) or secondary data, such as patents (Lee et al., 2007) and critics’ reviews of creative outputs (Uzzi & Spiro, 2005). For this second group of studies, the “tasks” are specific to the industry being studied (e.g., generating ideas for improving supply chain management or the successful staging of a musical), but it is unclear whether the creativity effects are generalizable to other tasks.

Recent research proposes that the positive effect of multicultural exposure on creativity applies only to tasks that draw on diverse cultural-knowledge systems (Cheng et al., 2008). In their study of biculturals (individuals with extensive experience in two cultures), Cheng and colleagues argued that the effect on creative performance of multicultural experiences was restricted to tasks belonging to specific cultural domains in which the individual has bicultural experience.

Creativity researchers have similarly argued that creative performance is enhanced by the possession of domain-specific skills and knowledge resources relevant to a given task (Amabile, 1983; J. Baer, 1993; Rietzschel et al., 2007). Specifically, Amabile proposed that familiarity with domain-specific facts, principles, performance “scripts,” and technical skills is a prerequisite for creativity in that given domain. Thus, receipt of domain-relevant informational resources from one’s network should directly aid individuals to become more creative at tasks in that domain. For any creative task that requires drawing on knowledge from diverse cultural sources, culture-related novel ideas are a form of domain-relevant knowledge as such ideas directly inform and stimulate how an individual goes about framing and solving the problem at hand. Specifically, ideas from diverse cultures prompt the individual to engage in more elaborate processing of culture-relevant information (i.e., deeply considering and integrating different cultural ideas and perspectives), directly benefiting the search for creative solutions (van Knippenberg, De Dreu, & Homan, 2004). For example, when creating a new fusion cuisine that combines ingredients from East and West, chefs’ knowledge about diverse cultures is highly relevant and can be directly applied to the task by informing, stimulating, and guiding their search for a new recipe. However, when the task at hand does not require diverse cultural knowledge (e.g., devising a new effective way to clean vegetables), diverse knowledge about other cultures that one draws from one’s social network would be less useful. Thus, one might expect the positive effect of a culturally diverse network on creativity to be highly domain specific. Because a multicultural network facilitates the flow of specific types of informational resources relating to other cultures, increasing one’s stockpile of cultural knowledge, its effect should be primarily applicable to task domains that require knowledge and ideas from multiple cultures.

Other research, however, suggests that the effects of multicultural exposure on creativity are generalizable to all kinds of creative endeavors. Tadmor and Tetlock (2006) argue that multicultural experiences help individuals recognize and integrate viewpoints grounded in diverse cultures and, consequently, enhance their ability to tolerate and even reconcile conflicting ideas and perspectives. Leung and colleagues (2008) propose that exposure to foreign cultures may also challenge or expand an individual’s routine knowledge structure by providing a new perspective toward problem solving, spurring divergent thinking. These claims are consistent with Perry-Smith’s (2014) argument that novel frames derived from social networks can enhance creative problem solving. Indeed, some cross-cultural psychologists have found evidence that multicultural exposure improves performance on a wide variety of tasks, including negotiation, idea generation, remote association of ideas, and even tasks such as drawing depictions of aliens (Leung & Chiu, 2008, 2010; Maddux & Galinsky, 2009). Taken together, these studies suggest that if multicultural networks indeed increase creativity, it might well be generalizable to a range of creative tasks.

Drawing on these two lines of argument and associated empirical evidence, I investigate two competing hypotheses. One is that individuals with more culturally diverse networks will be more creative but primarily at tasks that draw on multiple cultural-knowledge systems. The alternative hypothesis is that individuals with more culturally diverse networks will be creative at a broad range of tasks that may not require cultural knowledge because such networks provide novel frames and perspectives to look at problems, enhancing creative thinking in general.

I conducted two studies to test the above hypotheses. In Study 1, a laboratory experiment, I test Hypotheses 1 and 2 to shed light on the content and sources of informational resources that are derived from culturally diverse networks. In addition, I test the competing Hypotheses 3a and 3b about the domain-specific effects of multicultural networks on creativity. Specifically, I investigate the impact of network’s cultural diversity on tasks that require (a) knowledge about multiple cultures from around the world, (b) knowledge about a given local culture, and (c) little knowledge about any particular culture. Study 2 replicates the effects found in Study 1 for Hypotheses 3a and 3b in a field setting to increase external validity.

Study 1

Analyses and Results

I analyzed the data at both the participant and dyadic (participant-contact) levels. The participant-level dependent variables are (a) the degree of creativity exhibited in the task of generating ideas for an advertisement and (b) the total number of culture-related novel ideas received from network contacts over the preceding 3 months. The dyadic-level dependent variable is whether a participant received a culture-related novel idea from a given contact within the preceding 3 months.

For the participant-level (i.e., network-level) analysis of creative outcomes, I used ordinary least squares regression. I used contrast coding to capture the three types of task. Specifically, I used two indicators: (a) global versus imagination (coded 1, 0, and −1 for the global, local, and imagination conditions, respectively) to represent the comparison between the global task and the imagination task, and (b) global versus others (coded 2, −1, and −1 for the global, local, and imagination conditions, respectively) to represent the comparison between the global task and the other two tasks. The latter indicator is of key interest as it allows me to test Hypotheses 3a and 3b.

I also tested whether the degree of cultural diversity in networks influences the number of culture-related novel ideas that participants receive from their network contacts. Because this dependent variable is a count measure, I used negative binomial regression for this analysis. For the dyadic-level analyses, the dependent variable is binary, with a value of 1 if a culture-related novel idea is received from a given contact and 0 otherwise. Because the data contain hierarchically nested variables (up to 15 dyadic relationships are nested with a given participant), nonindependence of observations is a methodological concern (Hausman, Hall, & Griliches, 1984; Klein, Dansereau, & Hall, 1994). In response, I used STATA’s probit regression with robust cluster on the participant to control for the influence of a given participant on multiple observations. This approach allows for estimating dyadic-level effects within egocentric networks (Hoffman, Griffin, & Gavin, 2000) and effects for key participant-level variables, particularly the degree of cultural diversity in the network.

Table 1 presents the descriptive statistics for and correlations among variables. I first tested whether the cultural diversity of participants’ networks predicts the type of novel ideas they receive from contacts in their networks. Table 2 presents the results. As I predicted, cultural diversity in participants’ networks has a positive effect on the total number of culture-related novel ideas received (b = 1.27, p < .05). But network cultural diversity had no effect on receipt of other types of novel ideas; results for a sample selection of these idea categories (e.g., leisure and general knowledge) appear in Table 2.

OPEN IN VIEWER

Table 1

Study 1—Descriptive Statistics and Correlations (Dyadic Level)

Variable	1	2	3	4	5	6	7	8	9	10	11	12	13	14	15	16	17	18	M	SD	Min	Max
1. Creativity a	1.00																		2.96	1.19	1.08	6.03
2. Network cultural diversity	−.03	1.00																	0.39	0.23	0	0.83
3. Contact provides culture-related novel idea	−.01	.07*	1.00																0.05	0.22	0	1
4. Network size (number of contacts)	−.12*	.01	−.02	1.00															11.62	3.13	6	15
5. Network density	.04	−.09*	−.01	−.11*	1.00														0.53	0.27	0	1
6. Contact’s embeddedness	.03	−.07*	−.00	−.09*	.86*	1.00													0.53	0.32	0	1
7. Number of countries lived in (6 months or more)	−.06*	.05*	.01	.08*	−.01	−.01	1.00												0.15	0.41	0	2
8. Participant’s work experience	−.04	−.15*	−.04	.11*	−.07*	−.06*	−.01	1.00											3.17	3.22	0	16
9. Contact is of different culture	.01	.41*	.05*	.00	−.05*	−.07*	.01	−.09*	1.00										0.40	0.49	0	1
10. Contact is of different gender	.02	−.01	.00	.01	−.07*	−.06*	−.00	.06*	−.02	1.00									0.35	0.48	0	1
11. Contact is source of economic resources	.08*	−.08*	.01	−.05*	.06*	.11*	−.05*	.05*	−.11*	.08*	1.00								0.16	0.37	0	1
12. Contact is source of career guidance	.04	−.00	.03	−.06*	−.05*	−.05*	−.03	−.03	.00	.03	.36*	1.00							0.36	0.48	0	1
13. Contact is source of task advice	.07*	−.01	.02	−.02	−.08*	−.03	−.04	−.06*	−.04*	−.05*	.21*	.38*	1.00						0.53	0.50	0	1
14. Contact is source of friendship	−.01	−.06*	.00	.08*	.13*	.17*	.06*	−.01	−.06*	−.07*	−.17*	−.27*	−.00	1.00					0.86	0.34	0	1
15. Tie strength	−.02	.04	−.02	.05*	−.23*	−.33*	.01	.15*	.07*	.05*	−.10*	.04*	−.17*	−.32*	1.00				1.86	0.76	1	4
16. Participant’s gender (male)	.05*	.05*	−.01	−.11*	.16*	.13*	.04	−.07*	.06*	−.12	.05*	−.03	−.02	.02	−.04*	1.00			0.48	0.50	0	1
17. Participant’s age	−.10*	−.12*	−.05*	.08*	−.14*	−.12*	.07*	.59*	.00	.06*	.05*	.06*	−.00	−.09*	.21*	−.09*	1.00		21.27	2.73	18	30
18. Time spent abroad (months)	−.07*	.23*	.07*	−.01	.05*	.05*	.30*	.00	.12*	.01	−.04*	−.02	−.04*	.01	−.04	.04	.07*	1.00	19.38	46.46	0	306

a

Creativity scores for all three tasks have similar standard deviations, suggesting similar degree of sensitivity.

*

p < .05.

OPEN IN VIEWER

Table 2

Study 1—Number of Each Type of Novel Idea Received (Participant Level)

Type of idea	Dependent variable: Culture related	Sample of other idea categories
		Leisure related	Business opportunities	General knowledge	Academic advice	Work/career advice	Philosophy of life
Key predictor
Network cultural diversity	1.27 (0.53)*	−0.32 (0.32)	0.70 (0.67)	−0.23 (0.54)	−0.62 (0.54)	−0.81 (0.46)†	−1.02 (0.64)
Control variable
Network size	0.07 (0.04)†	0.10 (0.02)**	0.16 (0.05)**	0.00 (0.04)	0.10 (0.04)**	0.04 (0.03)	0.11 (0.05)*
Network density	−0.01 (0.43)	0.73 (0.28)**	0.74 (0.55)	−0.20 (0.45)	−0.25 (0.47)	−0.64 (0.38)†	0.15 (0.57)
Gender (male)	−0.19 (0.23)	0.16 (0.15)	0.13 (0.29)	−0.21 (0.24)	−0.03 (0.25)	0.05 (0.21)	−0.57 (0.30)*
Number of countries lived in (6 months or more)	0.03 (0.28)	0.00 (0.20)	0.03 (0.37)	−0.26 (0.31)	0.00 (0.33)	−0.13 (0.27)	0.41 (0.38)
Time spent abroad (months)	0.00 (0.00)†	0.00 (0.00)	−0.01 (0.00)	0.00 (0.00)	0.00 (0.00)	0.01 (0.00)*	−0.01 (0.01)
Participant’s age	−0.08 (0.06)	−0.06 (0.04)†	0.13 (0.07)*	0.10 (0.05)†	0.01 (0.05)	0.05 (0.05)	−0.05 (0.07)
Work experience	−0.01 (0.05)	0.02 (0.03)	−0.12 (0.06)*	−0.08 (0.05)†	0.00 (0.05)	−0.03 (0.04)	−0.02 (0.06)
Intercept	−0.13 (1.31)	0.50 (0.80)	−5.20 (1.62)	−2.15 (1.26)	−1.54 (1.29)	−0.91 (1.09)	0.25 (1.61)
Likelihood ratio χ2	19.11*	30.42**	15.95*	6.05	10.09	13.15	18.06*
R2	.05	.04	.03	.01	.02	.02	.03
Number of participants	209	209	209	209	209	209	209

Note: The table presents unstandardized coefficients with standard errors in parentheses.

†

p < .10.

*

p < .05.

**

p < .01.

To pinpoint the sources of culture-related novel ideas, I further analyzed the data at the dyadic level, controlling for a host of participant/contact variables. Table 3 shows that the cultural diversity of participants’ networks significantly predicts the likelihood that a culture-related novel idea is received from a network contact (b = 0.60, p < .05). Importantly, the receipt of culture-related novel ideas is not predicted by whether the other person is culturally different (b = 0.11, p > .10). Nor does network cultural diversity have any effect on the likelihood of receiving novel ideas of other types. For example, higher network cultural diversity does not predict receipt of a novel idea regarding business opportunities (b = 0.07, p > .10). Results for a sample selection of these idea categories appear in Table 3. Overall, these results indicate that individuals whose networks are highly culturally diverse are more likely to receive culture-related novel ideas but not other types of ideas; culture-related novel ideas do not necessarily come from culturally different network contacts. Thus, there is support for Hypotheses 1 and 2.

OPEN IN VIEWER

Table 3

Study 1—Probit Maximum Likelihood Estimation on Exposure to Novel Ideas From Each Contact (Dyadic Level)

Type of idea	Key dependent variable: Culture related	Sample of other idea categories
		Leisure related	Business opportunities	General knowledge	Academic advice	Work/career advice	Philosophy of life
Key predictors
Network cultural diversity	0.60 (0.27)*	−0.13 (0.23)	0.07 (0.29)	−0.08 (0.25)	−0.38 (0.31)	−0.49 (0.26)†	−0.34 (0.32)
Contact is of different culture	0.11 (0.10)	−0.17 (0.08)*	0.15 (0.13)	0.02 (0.12)	0.05 (0.10)	0.06 (0.12)	−0.08 (0.09)
Control variables
Contact’s embeddedness	−0.01 (0.17)	0.42 (0.14)**	0.18 (0.19)	−0.23 (0.19)	−0.05 (0.18)	−0.34 (0.17)*	0.20 (0.26)
Number of countries lived in (6 months or more)	−0.02 (0.15)	−0.05 (0.11)	0.08 (0.12)	−0.04 (0.19)	−0.02 (0.12)	−0.03 (0.12)	0.14 (0.20)
Time spent abroad (months)	0.00 (0.00)	0.00 (0.00)	−0.00 (0.00)	0.00 (0.00)	0.00 (0.00)	0.00 (0.00)**	−0.00 (0.00)
Network size	−0.01 (0.02)	0.00 (0.02)	0.02 (0.02)	−0.05 (0.02)**	0.01 (0.02)	−0.03 (0.02)	0.01 (0.02)
Intercept	−1.76 (0.42)**	−0.98 (0.27)**	−1.67 (0.42)**	−0.30 (0.32)	−1.95 (0.36)**	−0.68 (0.33)*	−1.21 (0.44)**
Wald χ2	18.65	55.16	25.80	25.49	24.97	58.55	31.68
Number of observations	2,233	2,233	2,233	2,233	2,233	2,233	2,233

Note: The table presents unstandardized coefficients with standard errors in parentheses. Control variables for the content of network ties, tie strength, and gender are not presented due to space constraints.

†

p < .10.

*

p < .05.

**

p < .01.

Table 4 reports the effects of the degree of network cultural diversity on creativity. Model 1 presents the effects of network cultural diversity and the key control variables. Model 2 adds the predictors for different types of tasks. In neither model did network cultural diversity exhibit any significant main effect on creativity. Model 3 adds the Network Cultural Diversity × Global Versus Others interaction term to test whether network cultural diversity predicts creativity only for a task that requires a global perspective. This interaction was significant in the expected direction (b = 0.45, p < .05). Specifically, network cultural diversity positively predicts the creativity of proposed ideas for the global condition (b = 1.12, p < .05) but not for the local (b = −0.51, p = .40) or imagination (b = −0.21, p = .75) conditions. This pattern of interaction is depicted in Figure 1. This finding supports the hypothesis that a multicultural network promotes creativity only on tasks that require knowledge and ideas from multiple cultures (Hypothesis 3a).

OPEN IN VIEWER

Table 4

Study 1—Regression Results for Participant Creativity

	Model 1	Model 2	Model 3
Key predictors
Network cultural diversity	−0.10 (0.38)	0.19 (0.33)	0.18 (0.33)
Global versus imagination		−1.06 (0.19)**	−1.04 (0.19)**
Global versus others (imagination and local)		0.28 (0.11)*	0.09 (0.15)
Network Cultural Diversity × Global Versus Others			0.45 (0.23)*
Control variables
Gender (participant is male)	0.10 (0.17)	0.02 (0.15)	−0.01 (0.15)
Number of countries lived in (6 months or more)	−0.06 (0.23)	−0.15 (0.20)	−0.13 (0.20)
Time spent abroad (months)	0.00 (0.00)	−0.00 (0.00)	−0.00 (0.00)
Network size	−0.04 (0.03)	−0.04 (0.02)†	−0.04 (0.02)†
Network density	0.21 (0.32)	0.07 (0.28)	0.11 (0.28)
Participant’s age	−0.03 (0.04)	−0.01 (0.03)	0.00 (0.03)
Work experience	−0.00 (0.03)	−0.01 (0.03)	−0.02 (0.03)
Intercept	3.91 (0.88)**	3.52 (0.78)**	3.52 (0.77)**
R 2	.03	.25	.27
Number of observations	205	205	205

Note: The table presents unstandardized coefficients with standard errors in parentheses.

†

p < .10.

*

p < .05.

**

p < .01.

OPEN IN VIEWER

Figure 1

Study 1—Interaction Between Network Cultural Diversity and Task Type

Finally, I conducted further tests to see whether the number of culture-related novel ideas that participants received from their networks mediates the effect of network cultural diversity on creative performance. Although network cultural diversity has a positive effect on both the number of culture-related novel ideas received (b = 0.57, p < .05) and creativity in the global-condition task, which requires drawing on other cultural-knowledge resources (b = 1.12, p < .05), the mediation effect was not significant. Specifically, when I added the number of culture-related novel ideas that each participant received into a model that regressed creativity on network cultural diversity in the global-condition task, the effect of network cultural diversity on creativity remained positive and significant (b = 1.24, p < .05). I discuss this finding in the Discussion section.

Study 2

Study 1 tests the hypotheses that individuals with multicultural networks are more likely to receive culture-related novel ideas and that these ideas are not necessarily conveyed by network contacts who are culturally different (Hypotheses 1 and 2). Study 1 also provides the first evidence that the effect of a network’s cultural diversity on creativity is domain specific (supporting Hypothesis 3a but not Hypothesis 3b). Study 2 seeks to replicate the effects of multicultural networks on creative performance demonstrated in Study 1 (i.e., Hypotheses 3a and 3b) by using a different sample, namely, a pool of working professionals tasked to tackle a major challenge in their profession. A strength of Study 2 is its external validity, as the participants were drawn from a field context involving a real-world problem that participants cared deeply about.

Analyses and Results

I analyzed the data at the participant (ego) level by using ordinary least squares regression. To test the effects of network cultural diversity on different types of creative tasks, I used the same form of contrast coding as used in Study 1 to capture the three types of tasks.

Table 5 presents the descriptive statistics for and correlations among the key variables. Table 6 presents results of the effects of a network’s degree of cultural diversity on creativity. Model 1 shows the effects of network cultural diversity and the key control variables. Model 2 adds the predictors for the different types of tasks—global versus imagination and global versus others. Network cultural diversity did not show any significant main effect on creativity in either model. Model 3 adds the Network Cultural Diversity × Global Versus Others interaction term to test whether network cultural diversity predicts creativity only for a task that requires a global perspective. This interaction was significant in the expected direction (b = 1.45, p < .05). Specifically, network cultural diversity positively predicts the novelty of proposed ideas for the global condition (b = 4.00, p < .05) but not for the local (b = 0.49, p = .76) or imagination (b = −0.30, p = .86) conditions. Figure 2 shows the graphical depiction of this interaction effect. Taken together, these results suggest that cultural diversity in networks increases creativity only for tasks that require drawing on knowledge from multiple cultures. Cultural diversity in networks has no effect on tasks that require knowledge about only one culture or do not involve any cultural knowledge whatsoever. In sum, Hypothesis 3a, but not Hypothesis 3b, is supported.

OPEN IN VIEWER

Table 5

Study 2—Descriptive Statistics and Correlations

Variable	1	2	3	4	5	6	7	8	9	M	SD	Min	Max
1. Creativity a	1.00									3.40	1.48	1	6.22
2. Network cultural diversity	.14	1.00								0.20	0.21	0	0.65
3. Network size (number of ties)	.03	.13	1.00							8.71	4.83	3	15
4. Number of countries lived in (6 months or more)	−.13	.07	.19	1.00						0.91	1.78	0	10
5. Time spent abroad (in months)	.18	.00	.02	.46*	1.00					118	233	0	1,228
6. Network density	.04	−.08	−.19	.01	.03	1.00				0.44	0.34	0	1
7. Participant’s work experience	−.22	.02	−.08	.20	.09	.00	1.00			28.40	14.38	0	62
8. Gender (participant is male)	.20*	−.07	−.16	−.05	.15	−.15	.26*	1.00		0.48	0.50	0	1
9. Participant’s age	−.25*	−.03	−.06	.26*	.07	.02	.93*	.24*	1.00	51.56	14.93	19	80

*

p < .05.

a

Creativity scores for all three tasks have similar standard deviations, suggesting similar degree of sensitivity.

OPEN IN VIEWER

Table 6

Study 2—Regression Results for Participant Creativity

	Model 1	Model 2	Model 3
Key predictors
Network cultural diversity	1.08 (0.82)	1.04 (0.81)	1.39 (0.79)†
Global versus imagination		0.40 (0.41)	0.46 (0.40)
Global versus others (imagination and local)		0.05 (0.25)	0.01 (0.24)
Network Cultural Diversity × Global Versus Others			1.45 (0.63)*
Control variables
Network size (number of ties)	0.02 (0.04)	0.03 (0.04)	0.06 (0.04)
Time spent abroad (in months)	0.00 (0.00)	0.00 (0.00)†	0.00 (0.00)†
Number of countries lived in (6 months or more)	−0.19 (0.12)	−0.20 (0.12)	−0.20 (0.12)†
Network density	0.43 (0.54)	0.26 (0.54)	0.38 (0.53)
Participant’s work experience	−0.01 (0.04)	−0.02 (0.04)	−0.03 (0.03)
Gender (participant is male)	0.20 (0.38)	0.24 (0.38)	0.48 (0.38)
Participant’s age	−0.01 (0.04)	0.01 (0.04)	0.01 (0.04)
Intercept	3.85 (1.03)**	3.23 (1.06)**	2.76 (1.04)**
R 2	.16	.22	.28
Number of observations	72	72	72

Note: The table presents unstandardized coefficients with standard errors in parentheses.

†

p < .10.

*

p < .05.

**

p < .01.

OPEN IN VIEWER

Figure 2

Study 2—Interaction Between Network Cultural Diversity and Task Type

Theoretical Implications

This research makes several theoretical contributions. First, the finding that the effect of multicultural networks on creativity applies to some types of tasks but not others contributes to creativity research. Although creativity scholars have long documented the effects on creativity of various individual characteristics (e.g., personality) and contextual factors (e.g., environmental stimuli and social network; see Zhou & Shalley, 2008, for a review), this literature has generally not paid much attention to the contexts in which creativity is required (Zhou & Su, 2010). Findings from this research revisit an often neglected line of inquiry for creativity researchers—under what task contexts do known drivers of creativity exert their effects? The current findings provide evidence that the effects of known drivers of creativity (in this case, network diversity) may not be exerted across all contexts.

Second, the present findings contribute to network research by showing that the effects of certain network properties, such as the degree of diversity, might well be domain specific. Multicultural networks promote the flow of culture-related novel ideas more than other types of ideas, resulting in benefits only for creative tasks that draw on multiple cultural-knowledge systems. Existing network research seldom differentiates types of outcomes or explicitly examines specific domains in which network variables exert an impact (M. Baer et al., 2015). Although the flow of diverse ideas and information is a commonly theorized mechanism underlying the effects of various network properties and creative performance (Burt, 2004; Perry-Smith, 2006; Rodan & Galunic, 2004), the content and sources of these ideas were also seldom examined. The present research looks into these issues. A key finding is that culturally diverse networks appear to promote the flow of culture-related novel ideas both directly (from different-culture network contacts sharing ideas about their own or other cultures) and indirectly (from same-culture network contacts sharing ideas about foreign cultures they have encountered). The finding that culture-related novel ideas might come from same-culture social network contacts is of theoretical importance because it challenges the implicit assumption that novel ideas arise primarily from network contacts who are different from oneself. Additionally, effects of network diversity are likely to be contingent on the types of task. Taken together, these findings compel network research to build more nuanced theories by more explicitly specifying the conditions under which proposed network effects are expected to occur as well as the type of information and ideas that flow in the networks.

Third, the findings in the present research also speak to diversity research in general. Extant research on organizational and team diversity has found mixed results (Hoever, van Kippenberf, van Ginkel, & Barkema, 2012; Williams & O’Reilly, 1998). A body of research has found positive effects of team cultural diversity on creativity (McLeod, Lobel, & Cox, 1996; O’Reilly, Williams, & Barsade, 1998). However, some research on functional diversity found no significant effects on organizational innovation (Sethi, Smith, & Park, 2001). To date, research linking diversity and creativity seldom distinguishes the different types of creative tasks when investigating diversity’s effect on creative performance. One possible explanation that the present research offers is that diversity’s effect on creativity and innovation is highly contingent on the type of diversity and whether the diverse informational resources it confers on the innovators are relevant to the tasks at hand.

Fourth, the present research contributes to a growing body of work that links multicultural experiences and creativity (Leung & Chiu, 2008; Leung et al., 2008; Maddux & Galinsky, 2009). Prior measures of multicultural experiences, such as living abroad, tend to conflate many different types of cultural exposures (e.g., making foreign friends, eating foreign food, exposure to foreign cultural icons, and experiencing local customs). Theories about the specific forms that cultural diversity takes—whether in the form of team composition, network profile, or experiences—and about the associated mechanisms and boundary conditions that underlie its effects on creativity are also still developing. It is therefore unclear exactly what aspects of multicultural experiences accounted for the observed increase in creativity in prior studies. This research identifies a multicultural network as a specific channel from which one could gain cultural knowledge that matters for creativity.

The finding that multicultural networks did not engender creativity in different types of tasks appears to run counter to an earlier finding that living abroad has broad implications for creativity (Maddux & Galinsky, 2009). One explanation is that living abroad and having a multicultural network are qualitatively different multicultural experiences. Living abroad provides an intense firsthand experience of immersion in and adaptation to a foreign culture. Such an experience may well have a deep influence, permanently altering one’s cognitive habits and promoting creative thinking ability in general. By contrast, a multicultural network represents relatively low-intensity, vicarious, but persistent and simultaneous, exposure to a range of cultures. An individual can engage or disengage network members at will, but they remain available to consult when the need arises. Ideas and perspectives gleaned from culturally diverse network members might be experienced vicariously through conversations with one’s social contacts and, thus, not as deeply imprinted compared to those that one experienced firsthand. Overall, my findings, when taken together with extant findings, suggest that the effects of multicultural experiences on creativity likely depend on the specific type of exposure. Thus, researchers should be specific about the type of multicultural exposure when investigating how multicultural experiences shape creativity. Future research could also explore the specific conditions under which a generalized effect of cultural diversity on creativity would materialize.

Practical Implications

This research has practical implications for managers and employees in the era of globalization. Creating a multicultural workplace is often touted as a strategy to foster organizational creativity. This advice, though intuitively appealing, should be selectively prescribed. Managers can most realistically expect enhanced creative performance from a multicultural workforce on tasks that call for combining ideas and knowledge from different cultural perspectives. On other types of task, the effects on creativity of having a culturally diverse workforce would probably be more uncertain and indirect; managers might be better served in these situations to seek other drivers of creativity to improve creative performance.

Understanding how multicultural networks foster the flow of culture-related novel ideas can nevertheless provide managers some additional tools for increasing organizational creativity. For instance, managers could create more opportunities (such as organizing an “international day” event) for employees of different cultures to exchange ideas. Managers could also explore how to better enable the flow of ideas across cultural lines. A recent study found trust to be a key facilitator of intercultural idea exchange (Chua et al., 2012). Hence, creating an organizational climate that fosters psychological safety and trust would be especially helpful.

Beyond learning from culturally different others, the present research also suggests that one can learn culture-related novel ideas from one’s own fellow citizens if one has a multicultural network. The finding from Hypothesis 2 highlights to individuals that as they strive to develop creative ideas by drawing on culturally diverse ideas from their networks, they should not ignore social contacts from their own cultures as potential sources of novel ideas about other cultures. As people’s exposure to multiple cultures increases as a result of globalization, social contacts from individuals’ own culture could very well provide them with novel ideas about other cultures that could translate into creative performance.

Limitations and Future Research

As with all research, the studies reported here have limitations. In hindsight, the measurement of novel-idea flow, although supportive of testing certain hypotheses (Hypotheses 1 and 2), was inadequate to capture the flow of ideas between individuals fully enough to test how creative performance on particular tasks comes about. My method relied on participants to select the ideas they wished to report and was thus subject to recall biases and social-desirability biases. For example, participants might have reported not the most novel ideas but those that impressed them most strongly. Or a controversial or provocative idea might have gone unreported. These limitations might partially explain why the number of cultural-related novel ideas did not significantly mediate the effect of network cultural diversity on creativity. Nevertheless, my methodology represents a starting point for research to examine the content and sources of ideas that flow in networks. Future research could build on my methodology to improve idea-reporting measurement.

The precise mechanisms underlying the relationship between culturally diverse networks and creativity deserve further investigation. Although extant research provides evidence that people with multicultural networks are perceived to be more creative (Jang et al., 2010), it would be useful to explicitly measure and demonstrate that how focal individuals are perceived predicts whether others are more likely to share novel ideas with them. For example, future research could ask network contacts to report on their perceptions of whether a focal individual is indeed cosmopolitan and open to new foreign ideas.

Researchers should also investigate other accounts that seem plausible. Specifically, there are alternative accounts as to why individuals with culturally diverse networks might receive culture-related novel ideas from same-culture social contacts. In my theory development, I argued that same-culture network contacts share culture-related novel ideas with a focal individual so as to maintain existing ties by appealing to the perceived interests of this individual. One alternative explanation could be that the focal individual, because of his or her exposure to diverse cultures in the network, might become adept at extracting culture-related novel ideas from network contacts—regardless of their cultural background. Another explanation could be that homophily plays a role. That is, when focal individuals have a culturally diverse social network, they might be drawn to same-culture network contacts who also have such networks. The confluence of common interests in diverse cultures would promote an exchange of culture-related novel ideas. These explanations are not mutually exclusive with my central thesis. In all likelihood, an individual’s multicultural network potentially promotes a two way process of receiving as well as extracting culture-related novel ideas from a range of network contacts regardless of their cultural backgrounds.

Relatedly, a research direction worth exploring is whether having a multicultural network has any effect on individuals’ intercultural competence. If multicultural networks indeed increase intercultural competence, what role does an individual’s intercultural competence play in improving creativity? Recent research on cultural intelligence and creativity suggests that cultural metacognition—reflective thinking about one’s own cultural knowledge and assumptions—might be a crucial determinant of whether intercultural interaction promotes creativity (Chua et al., 2012). One hypothesis is that individuals with high cultural metacognition are better positioned than others to reap the benefits of multicultural networks because they are better able to track and monitor opportunities for cultural cross-fertilization. For example, individuals with high cultural metacognition might be more sensitive and receptive toward culture-related ideas, and these individuals might actively solicit such ideas from their networks. Future research could quiz research participants on whether they actively seek out certain ideas or information from their social networks. Alternatively, it is plausible that individuals with high intercultural competence are more likely to develop multicultural networks and share culture-related ideas with their network contacts, prompting them to reciprocate by sharing other culture-related ideas from their own experiences.

In the present research, I did not measure openness to experience as a personality antecedent to creativity, but prior research has found some links (e.g., Leung & Chiu, 2008). Individual differences are not a major concern in this paper because I used an experimental approach based on randomization of participants into different experimental conditions. Participants in each experimental condition are therefore likely to have a range of openness to experience tendencies. However, future research can explore whether there is potential interaction between openness to experience, types of tasks, and network cultural diversity on creativity. Perhaps individuals high in openness to experience might be better able to harness the power of a culturally diverse network when the task at hand calls for drawing on knowledge from diverse cultures. Individuals with greater openness to experience might also be more likely to develop and maintain culturally diverse networks; thus, it would be useful to control for this variable in future research.

Another potential limitation to the present research is that the samples in both studies are drawn from the United States. Yet using a U.S.-centric sample should not have any significant implications on my central thesis. My focus is on individuals’ networks and the extent to which they are culturally diverse. The United States is a highly culturally diverse country, and there are significant variations in the degree of cultural diversity in the participants’ networks. Moreover, in all my analyses, I have controlled for participants’ overseas experiences, work experiences, age, and gender. In Study 1, I also controlled for the individual contact’s gender and culture for the dyadic level of analyses. Nevertheless, future work could draw on a more international sample to see whether the results hold up elsewhere.