Workshop Programme


Beyond Schelling and Axelrod: Computational Models of Ethnocentrism and Diversity
The Business School & Student Hub Building, All Saints Campus, Manchester Metropolitan University, Oxford Road, Manchester M15 6BH, UK.
details
maps

Day 1: Wednesday, June 7th 2017 (Room 5.02)

12:00 Arrival / registration / lunch (lunch provided in room)
12:45
Welcome and Introduction
Bruce Edmonds & Laurence Lesard-Phillips & David Hales

13:00 The mediating effect of deprivation in the association between ethnic density and racism: An ABM Approach.
Frencis Bras
abstract

13:30 Understanding the impact of residential segregation on the emergence of polarized attitudes towards ethnic minorities.
Thomas Feliciani
abstract

14:00
An Agent-based Modeling Approach to Predicting Effects of Open Enrollment School Choice Policies on Racial Integration in District Schools
Matt Kasman
abstract
14:30
Disruptive Norms - Assessing the impact of ethnic minority immigration on non-immigrant voter turnout using a complex model.
Thomas Loughran
abstract
15:00
Coffee break

15:30
Immigration, social networks and the emergence of ethnic segmentation in low-skilled labour markets.
Huw Vasey
abstract
16:00
Hammond and Axelrod’s model is not useful for studying ethnocentrism.
Fredrik Jansson
abstract
16:30
Panel 1: What do our models really tell us and how do we “sell” them?
Fredrik Jansson, Laurence Lesard-Philips, Edmund Chattoe-Brown, Bruce Edmonds.
Chair: David Hales
details
17:30
Workshop ends for the day




19:00
Social Dinner in Manchester: Umani at 147-153 Oxford Rd, Manchester M1 7EE.
Bruce Edmonds will provide further information on meeting-up etc. on the day.
map




Day 2: Thursday, June 8th 2017 (Room 5.05)

08:30
Arrival / coffee

09:00
Schelling models of Immigration: Parameters and Sensitivities.
Linda Urselmans
abstract
09:30
An Agent-Based model of interaction between immigrants and a host population: self-organized and regulated adaptation.
Laurence Lessard-Philips
abstract
10:00
Community diversity and inter-ethnic marriage: an agent-based approach to a complex social phenomenon.
Ruth Meyer
abstract
10:30
Coffee break

11:00
An Agent-Based Modelling Study of Persistent Segregation in Metropolitan Cape Town.
Cobus Van Rooyen
abstract
11:20
Endogenous Segregation Dynamics and Housing Market Interactions: An ABM approach.
Benjamin Bonakdar
abstract
11:40
What Actually Stops Us Going “Beyond” Schelling and Axelrod: Three Challenges.
Edmund Chattoe-Brown
abstract
12:00
Panel 2: The “new nationalisms” and ethical dimensions.
Martin Neumann, Benjamin Bonakdar, David Hales.
Chair: Laurence Lessard-Philips
details
13:00
Workshop ends


Talk Details



The mediating effect of deprivation in the association between ethnic density and racism: An ABM Approach.
Frensis Bras. The University of Manchester, UK.

Short abstract:  It has been theorised that ethnic minorities experience fewer racist events in areas with a higher concentration of ethnic minorities. At the same time, research has found that those who commit racially motivated assaults are more likely to live in deprived communities with higher unemployment rates, where prejudice against ethnic minorities is prevalent in the wider community (Ray et. al, 2004; Sibbitt, 1997). Yet, there is a strong correlation between ethnic concentration and area deprivation. This indicates that deprivation might play a complex and important role in the observed relationship between the concentration of ethnic minorities, prejudice towards ethnic minority groups, and the occurrence of racist events. This paper sheds light on potential mechanisms underlying these dynamic relationships through the implementation of an Agent-Based model. Through simulating various neighbourhoods with different levels of deprivation and concentration of ethnic minorities, the effect of deprivation and ethnic concentration was explored. The model indicates that ethnic minorities living in areas with a higher concentration of ethnic groups might experience fewer accounts of racism despite higher levels of deprivation. The model further indicates that ethnic minorities living outside of highly concentrated areas are more likely to become victims of racism.



Understanding the impact of residential segregation on the emergence of polarized attitudes towards ethnic minorities.
Thomas Feliciani1, Andreas Flache1, Jochem Tolsma2, Michael Mas1
1. Dept. of Sociology, University of Groningen, Groningen, The Netherlands;
2. Dept. of Sociology, Radboud University, Nijmegen, The Netherlands.

Short abstract: Agent-based models of social influence help us understand how attitudes can polarize in a population divided into two demographic groups. Here we focus on two prominent models of opinion polarization, the negative influence model and the persuasive argument model. Previous work has investigated how the two models make conflicting predictions about the direction and the strength of the effect that the spatial segregation of the groups has on the emergence of attitude polarization. Following this line of research, we test the robustness of previous findings, and investigate which segregation patterns are more likely to contribute to the emergence of polarized attitudes according to the two models. We do so by imposing a spatial setting that is much more realistic than assumed by previous studies, that of real cities, and assuming a multi-faceted definition of individual attributes. Individual attributes comprise some important predictors of attitudes towards ethnic minorities: ethnicity, age and household income. Population density and spatial distribution of demographic attributes are going to be calibrated on very fine-grained census data on Dutch cities, which provide a rich variety of segregation patterns.

Links: Extended abstract



An Agent-based Modeling Approach to Predicting Effects of Open Enrollment School Choice Policies on Racial Integration in District Schools
Matt Kasman. The Brookings Institution, Washington, DC, USA.

Short abstract: Open enrollment school choice policies allow (and sometimes require) families to make initial school selections rather than assigning students to schools by default. Many have argued that open enrollment policies have the potential to increase racial integration within district schools by creating opportunities for families in overwhelmingly impoverished, minority neighborhoods to access schools other than highly segregated neighborhood schools. Unfortunately, in practice the impact of open enrollment policies on diversity has been underwhelming, with many schools in large urban school districts that have implemented these policies still serving a substantial majority of students from a single racial background. In order to explain how the details of these policies influence their outcomes and might be altered to improve racial integration in public schools, I construct agent-based models of school enrollment in a large, urban school district using open enrollment. Using these models, I find that increasing participation in the school choice process would have the largest positive impact on racial integration in the district and increasing the priority given to residential proximity during student assignment would decrease integration.

Links: Extended abstract



Disruptive Norms - Assessing the impact of ethnic minority immigration on non-immigrant voter turnout using a complex model.
Thomas Loughran1, Edward Fieldhouse1, Laurence Lesard-Phillips2, Lee Bentley3
1. The University of Manchester, UK;
2. University of Birmingham, UK;
3. University of Liverpool, UK.

Short abstract: There is a substantial, but contradictory, sociological and political science literature related to the contextual effects of increased ethnic diversity on socio-political behaviour.  Much empirical work in this area has concentrated on providing testable measures that can contribute to the academic debate between contact theory and conflict theory regarding the impact of aggregate area level ethnic diversity on individual level attitudes and behaviours.   The paper will present findings from a number of simulations utilising the voter model of social processes connected to turnout.  The findings show that increased levels of immigration lead to an increased level of turnout among the non-immigrant majority population over time but that this effect is mitigated by the level of civic duty among the immigrant population.

Links: Extended abstract



Immigration, social networks and the emergence of ethnic segmentation in low-skilled labour markets.
Huw Vasey1, Ruth Meyer2
1. School of Arts, Languages and Cultures, University of Manchester, Manchester, UK
2. Centre for Policy Modelling, Manchester Metropolitan University, UK.

Short abstract: It has been widely reported that post-World War II immigration to more developed countries has gone hand-in-hand with the development of ethnically segmented labour markets, particularly in low-skill roles where entry requirements are minimal (Bauder, 2006; Piore, 1979; Sassen, 1996). Whilst numerous theories have been forwarded as to why such situations occur, it has remained difficult to empirically test such conceptualisations because of the numerous interacting processes which produce segmentation in the labour market. In this paper, we investigate the processes of ethnic segmentation in low-skilled labour markets, where referral hiring is the norm, with particular reference to the role of ethnically homogenous social networks and forms of ‘conservative’ discrimination. We employ an agent-based modelling approach, adapting key elements from Waldinger & Lichter’s (2003) widely cited networked explanation of ethnic labour market segmentation in late twentieth-century Los Angeles.Results indicate that ethnically homogenous social networks have the effect of increasing the level of ethnic segmentation within a referral-based labour market, but that these networks also help immigrant populations grow and protect them from the negative impacts of employer discrimination. We conclude that empirically-informed ABMs allow us to provide important insights into the manner and extent in which changes in social conditions may effect group behaviour. Such impacts are not random or occasional, but are tightly related to the non-linear and emergent nature of a multifaceted social system.



Hammond and Axelrod’s model is not useful for studying ethnocentrism.
Fredrik Jansson, Centre for the Study of Cultural Evolution, Stockholm University, Sweden.

Short abstract: Hammond and Axelrod’s ‘ethnocentrism’ model was initially published as a model on the armpit effect and inclusive fitness among simple organisms. The same model was later reframed as a model on ethnocentrism among humans, and inspired a branch of research on group discrimination. The question, though, is whether such a reframing is warranted. In short, I will argue that the resulting ‘ethnocentrism’ is driven by the fact that agents interact mostly with their clones, and group markers work as fairly reliable proxies for identifying nonclones. The applicability to what we know about ethnocentrism should be practically nil, and the model is sensitive to relaxing any of these assumptions. The conclusion is that there is little we can learn about group discrimination from the model in its current form, and attempts at generalisations are likely to fail.

Links: Extended abstract



Schelling models of Immigration: Parameters and Sensitivities.
Linda Urselmans. Dept. of Government, University of Essex, UK.

Short abstract: The Schelling model of racial segregation has made a vital contribution to our understanding of how ethnic segregation can occur even if residents of a neighbourhood are not particularly segregationists. Since its introduction in 1971 the model has become widely used in the field of immigration research. Building on previous iterations of the Schelling model I explore the effect of physical migrants represented as new agents entering the grid. The aim of this adaptation is to model immigration as an external shock to the existing system, rather than treat it as an initial state. The findings indicate that short-term effects of the rate and size migration weigh heavily on agent happiness and segregation behaviour, but that in the long run, population density is the most crucial determinant in model outcomes.

Links: Extended abstract



An Agent-Based model of interaction between immigrants and a host population: self-organized and regulated adaptation.
Carlos Lemos1, Ross Gore2, Laurence Lessard-Phillips3, LeRon Shults1
1. Dept. of Religion, Philosophy and History, Faculty of Humanities and Education, University of Agder, Kristiansand, Norway;
2. Virginia Modeling, Analysis and Simulation Center, Old Dominion University, Norfolk, VA, USA;
3. Institute for Research into Superdiversity, School of Social Policy, University of Birmingham, Birmingham, UK.

Short abstract: We present an ABM of “abstract” type for describing the interactions between a host population and a minority of immigrants, combining ideas from existing Agent-based models of segregation, dissemination of culture and ethnocentrism. The model also takes into account the framework elaborated by Berry (1997) for describing the role played by culture, contact and participation on processes of acculturation. The model entities are agents of a single type, a network between agents, and a “government” modeled as a “proto-agent,” i.e. as a procedure where the user can implement “policies” that affect agents. Agents interact with their network neighbors with a probability that depends on their education, cultural and identity “distance”. During the simulation the model produces a dynamic network of links between the agents. The “government” can then act in several ways, such as varying the probability of immigrants to acquire “security” (e.g. jobs) or lower the cultural barrier. The evolving network is analyzed using network segregation indices and algorithms for detection of social circles in ego networks. These patterns can be discussed under the light of sociological theories, for both “self-organized” and “regulated” interactions.

Links: Extended abstract



Community diversity and inter-ethnic marriage: an agent-based approach to a complex social phenomenon.
Huw Vasey1, Laurence Lessard-Phillips2, Ruth Meyer3
1. School of Arts, Languages and Cultures, University of Manchester, Manchester, UK;
2. Institute for Research into Superdiversity, School of Social Policy,University of Birmingham, Birmingham, UK;
3. Centre for Policy Modelling, Manchester Metropolitan University, UK.

Short abstract: Inter-ethnic marriage is both a cause and a consequence of immigrant integration. It is, however, unclear how individual preferences and opportunities for contact may combine to produce the spectrum of rates of inter-ethnic marriage we observe in the UK and elsewhere. In this presentation we investigate the processes of partnering in diverse communities, focusing on individual preferences, opportunities for contact, and group size to uncover how these may lead to differing rates of inter-ethnic marriage. We employ an agent-based modelling approach, utilising quantitative and qualitative sources from across the social sciences, in order to develop a complex model of emergent processes of differentiation and change in the marriage patterns of ethnic groups in a variety of different spatial settings. Results indicate that, in line with existing evidence, diversity (especially in areas with low ethnic homogeneity) fosters higher rates of inter-ethnic marriage. However, this is strongly mediated by group size, network types, and the extent of search ranges. We conclude that agent-based models allow us to provide important insights into the manner and extent in which changes in certain social conditions may affect group behaviour. Such impacts are not random trends, but are tightly related to the non-linear and emergent nature of a multifaceted social system. By extension, therefore, researchers need to remain willing and able to incorporate an analysis of the complexities of how inter-ethnic marriage rates interact with levels of diversity and social amalgamation, and how these may vary across space and place. In this presentation we shall also discuss on-going extensions to the original model, including modelling migration within small areas, and adapting partner preferences based on social scientific evidence.



An Agent-Based Modelling Study of Persistent Segregation in Metropolitan Cape Town.
Cobus Van Rooyen. Dept. of Geography, Birkbeck, University of London, UK.

Short abstract: The South African city was shaped more by a turbulent political past than by the inherent dynamics of urban growth.  The legacy of apartheid was ingrained in the fibre of urban life and more than twenty years after democracy, segregation is persisting in the major cities and emphasizes the substantial influence social and political factors had and still have on the urban development of the South African city. The present research project aims to extend research on the spatial distribution and socio-economic exclusion of the segregated urban areas of South Africa and improve the understanding of the foundations of segregation in the urban environment in the country. The aim is to construct an agent-based model, which will provide for the ability of producing alternative ‘what-if’ scenarios to study the impact of complex dynamical mechanisms on the persistence of racial and socio-economic exclusion in the study area. This will serve as theoretical foundation on which the specification and development of modelling methodology is based for the research study.

Links: Extended abstract



Endogenous Segregation Dynamics and Housing Market Interactions: An ABM approach.
Benjamin Bonakdar. Institute for Macroeconomics Ruhr-University Bochum, Germany.

Short abstract: In contrast to previous research, I hypothesize that residential segregation patterns do not only result from an individual’s perception of different ethnicities in a particular neighborhood, but is rather influenced by socioeconomic factors. The underlying assumption here is that the interpretation of Schelling’s statement “being close to people of your own kind” can be extended to the social status of an individual, which is part of the comparison from oneself to the society and the respective peer group. Accordingly, agents are endowed differently, which leads to the emergence of a system with a higher degree of heterogeneity. In order to analyze these dynamics, I implement an agent-based model with several features, where the decision criterion of moving is connected to housing affordability. The endogenous segregation dynamics get determined by an endogenous tolerance function, a multidimensional dissimilarity index and a happiness function, which serves as determinant for the actual moving decision. Since agents are bound to their individual housing budget, they can only move, if a suitable spot was found. The ability or disability of moving elsewhere might lead to further segregation outcomes and thus, to other incentives for segregation behavior.

Links: Extended abstract



What Actually Stops Us Going “Beyond” Schelling and Axelrod: Three Challenges.
Edmund Chattoe-Brown. Dept. of Sociology, University of Leicester, University Road, Leicester, UK.

Short abstract: Agent-Based Modelling finds itself in the unusual position of having a methodology that almost everyone seems to agree on, with examples showing that it works (which, incidentally almost nobody cites) but, at the same time, almost nobody follows. This presentation shows how, in three different ways, this oddity harms the possibility of really going “beyond” Schelling and Axelrod. The first challenge involves “element selection”. How do we decide (and much more importantly how do we justify) the selection of some elements (and, either explicitly or implicitly, the rejection of others) in a particular ABM? The second challenge is dealing with “heaps of models”. Because there currently seems to be no principled basis for element selection, models simply emphasise the interests or disciplinary backgrounds of their designers or evolve until they deliver the “right kind” of answer. The third challenge is that of “research design” and validation. What would it mean to say that we had supported (or failed to support) a particular variant of the Schelling (or Axelrod) models with evidence? I will discuss these challenges and suggest possible solutions.

Links: Extended abstract