Migration as a network, while not a novel idea, has recently gained significant attention for analyzing both international and internal migration patterns (Danchev & Porter, 2021; Pitoski et al., 2021a). However, despite its potential, network analysis has often been applied to migration with insufficient attention to interpretability(Pitoski et al., 2021a), focusing on estimating network metrics and tools, such as community detection algorithms, even when they are conceptually inadequate. Furthermore, the lack of comprehensive migration flow data has impeded effective network analysis. When flow data is available, assigning weights to ties remains a challenge, with proposed solutions often inadequate—for instance, simply taking the number of migrants moving from one place to the other (Abramski et al., 2020; Chen et al., 2021; Fagiolo & Mastrorillo, 2013; Pitoski et al., 2021c, 2021b; Zhang et al., 2020), or binarizing flows based on arbitrary thresholds (Carvalho & Charles-Edwards, 2020; Peres et al., 2016). Rarely has weight attribution accounted for origin and destination population sizes, and never for overall migration levels or all three factors simultaneously.

Adequately measuring and interpreting migration as a network is crucial, as the conceptual framework of migration networks offers exciting opportunities for advancing migration research. By merging classical migration theories with long-theorized systemic approaches (Mabogunje, 1970), deeper insights into migration phenomena can emerge, particularly in the context of internal migration, where the structure of the migration system is less shaped by political factors. The relatively better availability of flow data for internal migration also paves the way for robust longitudinal analysis of migration networks. In recent years, systemic analyses exploring changes in the structure and interconnectivity of internal migration have gained traction (DeWaard et al., 2020; Huang & Butts, 2023), particularly in response to declining migration intensity in some Western countries—trends that have proven difficult to explain, especially with bilateral-level analyses. Moreover, major societal shifts (e.g., the COVID-19 pandemic and the rise of remote work) may have long-term implications for the (re)configuration of migration systems. However, realizing these analytical promises depends on adequately measuring and transposing migration systems into networks.

This paper proposes a novel method for rescaling migration networks that neutralizes the effects of population size and yearly migration fluctuations while ensuring that edge weights accurately reflect the attractiveness between geographical areas. This approach enhances the overall interpretability of the network and enables the use of network tools that are otherwise difficult to apply when edges are improperly weighted. Notably, it facilitates the identification of repulsive ties, which are critical for a more comprehensive understanding of migration systems and improves the validity of community detection. It also enables the calculation of network metrics that were previously inapplicable to migration systems (e.g., density) or had limited informative value (e.g., certain centrality measures), providing more direct, intuitive, and meaningful interpretations. As an example, we will apply this method to Canadian migration data to demonstrate its potential for network-based migration analysis.

Urban studies have shown that youth living in disadvantaged and remote areas are less likely to have relationships to people from different social backgrounds and milieus, sometimes called bridging social capital. This so-called neighbourhood effect may restrict their life opportunities. However, existing evidence is largely based on the area where young people (the ‘egos’) live and often ignores the personal (egocentric) network context and where the individuals connected to the young people (their ‘alters’) live, possibly leading to misattribution errors. This paper aims to bridge this gap by examining how the spatial patterns of young adults’ personal networks relate to social capital. The data come from the 2020-23 Swiss Federal Survey of Adolescents (www.chx.ch) on personal networks of young adults. It includes the almost complete national cohort of young Swiss men (N approx. = 60,000) aged between 18 and 21 years, and a sample of about 15,000 Swiss women aged 19. Based on data from this large-scale representative national youth survey, we first develop a typology of four distinct spatial patterns based on how alters are clustered in and scattered across places defined as Swiss (or transnational) employment areas and foreign locations. We then analyse how these patterns are related to area deprivation where young people reside and social capital, measured in terms of the size, composition (homophily) and structure of emotional support and conflict ties between young people and the members of their personal networks.

Spatial isolation approaches to racial disparities have long assumed that lack of access to well-resourced communities is a critical mechanism through which racial segregation limits the wellbeing and life chances of Black youth. Recent research indicates that Black youth spend a substantial amount of time outside of segregated Black spaces due both to residence and everyday routines. Although Black youth may benefit from access to organizational and amenity resources in advantaged neighborhoods, they may also experience elevated exposure to the risk of scrutiny, microaggression, and discrimination. To date, few quantitative studies have addressed the emotional experience of Black youth as they navigate socioeconomically advantaged urban areas, particularly youth in groups that include Black males – potentially drawing greater negative scrutiny from the local environment.

We draw on data from the 2014-16 Adolescent Health and Development in Context study to examine the in situ emotional responses of Black youth to spending time in advantaged neighborhoods. The AHDC uses a network name generator to collect up to 10 non-household friends or other people with which the youth spends the most time during a typical week. Youth participants report on age, race, sex, education, behaviors, risk behaviors, and relationship attributes for each named partner. After an initial in-home survey, youth participants are given a smartphone to complete a 7-day geographically-explicit ecological momentary assessment (EMA) that collects continuous GPS data along with smartphone-based brief surveys. The EMA module administers five daily, randomly-timed prompts to complete a mini-survey that collects current information on the presence of household and network partners, affect, activities, risk behaviors, and perceptions of the social climate of the current location at the time of the prompt. By asking which specific household and network partners are present in the moment, EMA responses can be linked to characteristics of the network partners present as well as aspects of the spatial environment.

Our dependent variables are youths’ EMA ratings of negative and positive affect. Youth report agreement at the time of the prompt on a scale from 1 to 5 with 1 being “not at all” and 5 being “extremely” to these 7 items: daring, happy, confident, relaxed, excited, cheerful, and energetic (Cronbach’s alpha = .86). Negative affect includes 9 items: angry, sad, stressed, bored, lonely, afraid, rejected, nervous, and irritable (Cronbach’s alpha = .80). The presence of Black friends that are male is a count of named network partners reported as present at the time of the EMA prompt, constructed as categories of 0 (reference), 1, 2, and 3 or more Black male friends present.

OLS regression models of emotional outcomes with youth-level fixed effects (N=1727 EMAs nested within 406 Black youth) offer evidence that the affluence level of the immediate neighborhood environment is negatively associated with positive mood for Black youth when accompanied by multiple Black male youth. We discuss the implications of everyday routine location effects on mood for understanding racial disparities in mental health and wellbeing as well as the broader benefits of incorporating network data into EMA data collection approaches.

Over the last few decades, the number of shops offering an alternative to the traditional supermarket model has grown in France. One of this shop's aims is to create social links between these consumers. The way the shop operates (participative or not), its clientele (homogeneous or heterogeneous), its place in the neighborhood, or the layout of the premises can be favorable or unfavorable to interaction between consumers.

What is the influence of the type of shop on the consumers’ interaction network structures and dynamics?

A non-participant observation protocol with an observation grid and a field notebook was set up on 3 sites: an organic franchise shop, a solidarity grocery shop and a cooperative supermarket. The grid is used to systematically record consumers’ attributes and their interactions in the shops to build a network per observation (25 sessions so far).

The first results show that the cashiers are at the center of the 3 networks, which is consistent with the way a shop operates. The organic store network has the biggest number of nodes (677) but the lowest density (1,6%) and 70% of interactions are between employees and customers. Only 4 edges (1,7%) are between customers who did not enter the shop together. In contrast, the solidarity grocery shop has a denser network (29%). Finally, the cooperative supermarket network has more interaction with cashiers than with other customers, but their exchanges are richer (edge weight of 2,43 out of 3) because the customer and the cashier are both members of the cooperative.

Over the past several years, sexual abuse scandals have erupted across organizations ranging from universities to workplaces to places of worship. Despite arrests and monetary payouts to victims, abuse often continued undetected, and covered-up, for decades. In this presentation, I argue that organizations systematically conceal misconduct by strategically moving workers accused of abuse to locations that are geographically distant and socially marginalized. I investigate this phenomenon by quantitatively examining the relocation of Catholic priests credibly accused of child sexual abuse within the Archdiocese of Boston. In the broader Catholic Church, movement of priests - the “geographic solution” - is a well-documented response to abuse allegations (Reynolds 2023; Wall 2019). Recent work suggests that dumping grounds for accused priests did not have to be geographically remote as long as they were socially marginalized (Reynolds 2023). It is less well understood how these forms of the “geographic solution” combined to create movement patterns of accused priests and how these patterns differed from those of non-accused priests. I investigate how the geographic and social location of parishes, in combination with the structure of organizational mobility networks, relate to abusive priest relocation. Using novel data collected from the Official Catholic Directory and BishopAccountability.org, I recreate the movement patterns of both non-accused and accused priests in Boston from 1930-1990 and convert them into networks in which the nodes are geocoded parishes and the edges are priest movements among them. I examine whether the movement of accused priests creates distinct network patterns reflective of strategic relocation, and I compare these movements to those of non-accused priests. I test whether the relocations of abusive priests were skewed towards socially and geographically distant parishes and influenced by the mobility network structure.