Psychological factors surrounding disagreement in multicultural design team meetings

Abstract

Our research contrasts two theories of creativity in multicultural teams. The dual-process model focuses on the degree of diversity, whereas cross-cultural psychology focuses on specific cultural compositions. In individualistic cultures, team members express more conflicts and benefit from it, compared to in collectivistic cultures that emphasise harmony. The relative representation of members from these cultures may affect team dynamics, conflict and creativity. We coded over 3100 speaker turns in the 11th Design Thinking Research Symposium data-set for the presence of disagreements and examined the effects of conflict phase and team diversity on creativity, promotion and prevention approaches using Linguistic Inquiry Word Count measures. We found that micro-conflicts increased insight words in the moment of the conflict. Individuals in more diverse team meetings of Scandinavians and South-East Asians expressed fewer conflicts than teams dominated by Scandinavians and were less likely to focus on potential gains when experiencing micro-conflicts. Regardless of conflict, the more culturally diverse teams were more likely to use insight and promotion words overall. There were no effects for prevention. These findings extend extant theory to different types of heterogeneous teams in a real-world design setting. This study is novel in combining theory on team cultural diversity with a micro-process method.

Keywords:

• Teams
• culture
• diversity
• conflict
• promotion
• creativity

Acknowledgement

The authors are grateful to Shauna Sweet for statistical advice.

Notes

1. Kappas from 0.40 to 0.59 are considered moderate, 0.60–0.79 substantial and 0.80 and above outstanding (Landis and Koch 1977). In addition, when rating categories are used relatively frequently or infrequently, intercoder reliability may be lower (Smith 2000).

2. The final model for assent words, which included effects for team size, team composition, conflict and the interaction between conflict and team composition, was significant, χ² (11, N = 475) = 87.16, p < 0.001. The results for during conflict versus baseline were B = −0.88 (−1.19, −0.58), SE = 0.15, Wald χ² = 32.93, Exp(B) = 0.41 (0.31, 0.56), p < 0.001; for delayed conflict versus baseline contrast, B = −0.32 (−0.63, −0.02), SE = 0.15, Wald χ² = 4.43, Exp(B) = 0.72 (0.54, 0.98), p = 0.035. Multicultural teams had 47% fewer assent words, (p < 0.001), seven person teams had 54% more assent words compared to four person teams, p = 0.002) and multicultural teams had 85% more assent words just after conflict (postcon1) compared to Western-only teams (p = 0.011).

3. For differentiation words, the results for during conflict versus baseline were B = 0.32 (0.17, 0.48), SE = 0.08, Wald χ² = 16.09, Exp(B) = 1.38 (1.18, 1.62), p < 0.001; for postcon1 vs. baseline, B = 0.22 (0.02, 0.41), SE = 0.10, Wald χ² = 4.75, Exp(B) = 1.24 (1.02, 1.51), p = 0.03. Only conflict and team size were included in the final model of differentiation words, χ² (6, N = 475) = 27.00, p < 0.001.

4. The final model for negation words, which included main effects for team composition, conflict and the interaction between the two, was significant, χ² (9, N = 475) = 38.56, p < 0.001 (although team composition itself was not significant, p > 0.80). In addition to the effects for during conflict versus baseline, B = 0.61 (0.27, 0.95), SE = 0.17, Wald χ² = 12.54, Exp(B) = 1.84 (1.31, 2.58), p < 0.001, during the post1 segment, multicultural teams were twice as likely to use negation words compared to Scandinavian-only team meetings during the same phase, B = 0.69 (0.02, 1.37), SE = 0.34, Wald χ² = 4.04, Exp(B) = 2.00 (1.02, 3.93), p = 0.044.

5. Correlations tested using Spearman rho due to the lack of normality of the variables (n = 475).

6. Specifically, the natural log of the total words per segment.

7. For number of conflict turns overall, B = −0.66 (−1.54, 0.23), SE = 0.45, Wald χ² = 2.11, Exp(B) = 0.52 (0.21, 1.26), p = .146; insight words, B = 0.10 (−0.04, 0.23), SE = 0.07, Wald χ² = 2.01, Exp(B) = 1.10 (0.96, 1.26), p = 0.156; or risk words, B = −0.14 (−0.64, 0.36), SE = 0.26, Wald χ² = 0.31, Exp(B) = 0.87 (0.53, 1.43), p = 0.577.