The Emergence of Inequality in Social Groups
Leader: Prof. Dr. Claudia Wagner
Scientific unit: Computational Social Science (CSS)
One of the most challenging problems facing our society today is the growing prevalence of socioeconomic inequality. In social organizations ranging from small communities to entire nations, many undesirable social outcomes can result from lopsided distributions of income and wealth. In society at large, imbalanced distributions of wealth may negatively impact happiness [Alesina et al. 2004; Luttmer et al. 2005] and health [Wilkinson and Pickett 2010], and increase violence [Fajnzylber et al. 2002], or lead to political instability [Alesina and Perotti 1996]. In communities and smaller organizations such as firms, higher income inequality decreases individual productivity and job satisfaction, inhibits collaboration, and results in lower organizational performance [Pfeffer and Langton 1993; Bloom 1999]. The significance of growing inequality in society is of concern not only to academics, but also governance organizations [OECD 2011], and the general public through popular literature [Piketty and Goldhammer 2014] and the Occupy movement [Adam 2011; Jacobs 2011; Thompson 2011].
Previous research on inequality may be categorized in two main strands. On the one hand, sociological work on social stratification and mobility has examined how structural and psychological factors, such as parents’ socioeconomic standing, individual and peer status aspirations, and innate ability, translate to individual outcomes such as social class and income [Kerckhoff 1995; DiPrete and Eirich 2006]. These analyses generally disregard interactions between individuals and rarely investigate the behavioral mechanisms that produce the observed correlations. Part of the reason for this is that observational studies have significant limitations in establishing the causal effects of emergent outcomes.
Behavioral experiments can overcome this inability to establish causality. Thus, the other major strand of inequality research stems from behavioral experiments, primarily in social psychology. These lab experiments have investigated how social differentiation by status, power, or even arbitrary group membership makes individuals accept and propagate inequalities in small groups [Baron and Pfeffer 1994; Thye 2000; Cook and Rice 2003; Correll and Ridgeway 2003; Ridgeway 2006]. While these studies allow for interactions between individuals, they are largely limited in size and often confined to individual-level effects, rather than emergent macro-level properties.
To unify these literatures, then, the project partners require controlled experiments at a larger scale than before. However, such experiments are inaccessible using the traditional method of brick-and-mortar laboratory experiments. Recent progress in computational social science overcomes these limitations through “virtual lab” experiments [Mason and Suri 2012; Mao et al. 2012], which use software systems and large numbers of online participants, and allow experimental research at significantly greater scale and complexity. These experiments point toward the potential of studying large groups of participants interacting over long periods of time [Mao et al. 2016a] and the ability to create ever more realistic “virtual communities” in the lab.
Large-scale experiments with online participants have investigated the emergence of cooperation [Rand et al. 2011; Suri and Watts 2011; Wang et al. 2012; Mao et al. 2016a], social contagion [Tsvetkova and Macy 2014, 2015], social learning [Mason and Watts 2012], and collective intelligence [Mao et al. 2016b]. A handful of recent studies have also studied inequality. Salganik et al.  investigated unpredictability and inequality in an artificial cultural market, Shah et al.  studied the consequences of poverty and scarcity on individual behavior, van de Rijt et al.  tested for “rich-get-richer effects” in online communities, and Nishi et al. [2015b] investigated if visibility of wealth affects cooperation and aggravates inequality.
Building upon this work, the project partners propose to use large-scale controlled experiments with web and mobile-based participants to causally investigate how group-level inequality emerges and sustains through social interactions.
Runtime01.02.2017 - 28.2.2022
Dr. Andrew Mao, Microsoft Research, Computational Social Science, New York, USA
- AssistProf. Dr. Milena Tsvetkova, Department of Methodology London School of Economics and Political Science, London School of Economics and Political Science, London, UK