Understanding the Vicious Circles of Segregation. A Geographic Perspective

R&D project 2019–2023

Project info in the Estonian Research Information System

A geographic research project aimed at gaining a better understanding of segregation between socioeconomic and ethnic groups. We go beyond residential neighborhoods by exploring segregation in other life domains as well, including work, schools, and leisure.

Segregation in those domains is (re)produced through individual level sorting processes and the contextual effects people get from those domains. We will use a unique combination of large-scale longitudinal microdata:

(a) linked Estonian population censuses (1989, 2000, 2011) and register data;
(b) smartphone data, complemented with GIS layers with activity sites and a post-tracking survey.

As an outcome of the project, we will

a) have a better understanding of the links between segregation in different life domains;
b) developed new methodological tools for smartphone-based spatial mobility and segregation research;
c) facilitate PhD studies;
d) contribute to the development of smart and inclusive urban planning.

---

Objectives

Growing levels of socioeconomic and ethnic segregation (SEG) can seriously harm social inclusion and ethnic integration, undermine the economic competitiveness of cities, and increase concerns about safety as well as the intergenerational transmission of (dis)advantage (OECD 2018). According to the Urban Agenda for the EU, spatial SEG is one of the key challenges facing European cities.

The key innovation of this project is the systematic study of SEG across all life domains—places of residence, schools, workplaces, and leisure activities—making it highly relevant to the social sciences at large, despite its primarily geographic focus.

We will use a unique combination of census, register, and smartphone data, which, for the first time, will enable a detailed study of mobility and SEG across multiple domains. As a result of this project, we will develop new methodologies and gain new knowledge on:

---

The Concept of ‘Vicious Circles of Segregation’ (VCS)

The starting point of our conceptual framework is the concept of vicious circles of segregation (VCS) (Van Ham and Tammaru 2016; Van Ham et al. 2018). According to this concept, SEG is correlated across different domains. From a geographic perspective, these domains encompass all activity sites within a given urban region: all residential neighborhoods form the residential domain, all workplaces form the work domain, all schools form the school domain, and all leisure sites form the leisure domain. SEG in each domain can be measured based on socioeconomic status and ethnicity. SEG in different domains arises from in situ changes (cf. Finney and Simpson 2009), coupled with:

(a) the micro-level sorting of individuals with particular characteristics into activity sites; and
(b) contextual effects on individual outcomes due to exposure to and interaction with others (e.g., neighbors, friends, colleagues, classmates) in these activity sites (Van Ham et al. 2018).

Contextual effects shape various factors, including individual choices, which, in turn, influence future sorting into activity sites. Policies targeting different domains, individual resources and preferences, and discrimination all affect the sorting processes into activity sites and, consequently, SEG (Hulchanski 2010; Kährik and Tammaru 2008; Leetmaa et al. 2015). This sorting process is further structured by time and space factors (Hägerstrand 1970; Silm and Ahas 2014; Van Ham et al. 2018).

Time

The VCS evolves throughout the life course and is partly intergenerational, leading to place stratification (Portes and Rumbaut 1996) across all domains (Van Ham et al. 2018). A child is born into a neighborhood in which that child’s parents can afford, and as time passes, the child will likely attend a local school and this may transmit residential SEG into school SEG (Bernelius and Vaattovaara 2016). Educational inequality, the sorting into schools, and school characteristics all have an effect on outcomes later in life.

This affects the sorting into workplaces and the incomes that people earn later in life (Lam et al 2017) which, in turn, shapes in which neighbourhoods individuals can afford to live (Hulchanski 2010). Socioeconomic status and ethnicity play crucial roles in this sorting process (Bolt and Van Kempen 2010). Furthermore, for ethnic minorities, improved socioeconomic status can be a important factor in breaking the VCS (Alba and Nee 2003).

Sorting and exposure mechanisms vary across individual life stages. For example, during the family formation stage, school quality often becomes a major consideration in housing choices (Owens et al. 2017).

Space

An individual’s visits to activity sites collectively form their activity space (Golledge and Stimson 1997). The sorting into activity sites is shaped by urban planning, such as the spatial distribution of housing, workplaces, schools, and leisure activity sites across the urban region (Van Ham and Tammaru 2016).

For most people, the place of residence serves as a central activity site, from where daily activities usually start and end (Silm et al. 2018). Proximity and connectivity (i.e., accessibility) influence daily trajectories, as activities closer to places of residence or other central sites tend to cost less time and money (cf. Hägerstrand 1970).

For elderly, ethnic minorities, and low-income people, the residential neighborhood is often the primary arena for daily interaction (Bolt and Van Kempen 2017). For those who are employed, the workplace location (Kamenik et al. 2015) and, for children, the school location (Bernelius and Vaattovaara 2016) also play crucial roles.

---

Methods

The project builds on a unique combination of datasets, including:

(a) the Infotechnological Mobility Observatory (IMO),
(b) smartphone tracking data, complemented by
(c) GIS layers on activity site characteristics, and
(d) a post-tracking survey.

The methodological framework includes an indices-based SEG approach, GIS-based analytical methods, longitudinal modeling, techniques for analyzing space-time behavior at the individual level, and social network analysis. The study’s case area is the Tallinn urban region.

Our research population consists of individuals residing in the urban region, though their activity spaces may extend beyond its boundaries, including cross-border commuting. All previously discussed domains will be analyzed—residence, work, school, and leisure. The focus is on socioeconomic and ethnic SEG. Socioeconomic status is measured using education, occupation, and (household) income, while ethnicity is assessed through country of birth, mother tongue, and self-defined ethnicity.

---

The project is based on the following three integrated work packages (WP):

WP 1: Vicious Circles of SEG: the macro-level perspective

This WP provides a macro-level analysis of SEG in different domains and builds a contextual database for studying individual level sorting and contextual effects in WP 2 and WP 3.

WP 2: Vicious Circles of SEG: a longitudinal micro-level analysis

This WP focuses on how the sorting of individuals into activity sites and exposure to others in those activity sites contributes to the (re)reproduction of SEG between different domains over the life course and across generations.

WP 3: Vicious Circles of SEG: a micro-level analysis of the activity space

The censuses and registers allow for the study of individual trajectories in different domains, but we still miss important pieces of information when it comes to understanding how SEG is (re)produced: (a) we do not always know whether members of different ethnic and socioeconomic groups are co-present, it means in the same activity sites at the same time; and (b) with whom they interact there. Guided by a time-geographic research perspective (Hägerstrand 1970; Silm et al 2018) and enriching it with new innovative smartphone tracking data, we aim to detect the co-presence of different ethnic and socioeconomic groups in different activity sites.