Homophilous Networks Research Articles

Networks and identity drive the spatial diffusion of linguistic innovation in urban and rural areas

Cultural innovation (e.g., music, beliefs, language) tends to be adopted regionally. The geographic area where innovation is adopted is often attributed to one of two factors: (i) speakers adopting new behaviors that signal their demographic identities (i.e., an identity effect), or (ii) these behaviors spreading through homophilous networks (i.e., a network effect). In this study, we show that network and identity play complementary roles in determining where new language is adopted; thus, modeling the diffusion of lexical innovation requires incorporating both network and identity. We develop an agent-based model of cultural adoption, and validate geographic properties in our simulations against a dataset of innovative words that we identify from a 10% sample of Twitter (e.g., fleeky, birbs, ubering). Using our model, we are able to directly test the roles of network and identity by comparing a model that combines network and identity against simulated network-only and identity-only counterfactuals. We show that both effects influence different mechanisms of diffusion. Specifically, network principally drives spread among urban counties via weak-tie diffusion, while identity plays a disproportionate role in transmission among rural counties via strong-tie diffusion. Diffusion between urban and rural areas, a key component in innovation spreading nationally, requires both network and identity. Our work suggests that models must integrate both factors in order to understand and reproduce the adoption of innovation.

Quality education for all in an ideas-informed society: exploring idea-engagement amongst friendship networks and how education can prevent idea “echo-chambers”

Purpose The purpose of this study is to begine to address this question. The concept of the ideas-informed society (IIS) represents a desired situation in which citizens actively and critically engage with new ideas, developments and claims to truth. Its successful actualisation is dependent on high-quality educational opportunity at all stages of the life course. Social networks represent our connections to one another. Features of our social networks impact on how we engage with ideas. For instance, homophily dictates that individuals form networks with others seen as being like themselves. A key question, however, is whether there are forms of homophily that, by the nature of those they bring together, promote ideas engagement by individuals and the implications of consequent networks for the IIS? Design/methodology/approach This study re-analysed survey data from 1,000 voting-age citizens in England. Focusing on friendship networks, the authors used a structural equation model approach to explore the existence and potency of homophilic friendship networks; whether such networks drive respondents’ ideas-engagement with friends; and whether ideas discussions with friends impacts on the importance respondents place on staying up to date. Findings Political homophily has the strongest influence on whether people discuss new ideas with their friends (ES = 0.326, p < 0.01). In turn, ideas discussion has a significant impact on the extent to which people value engaging with ideas (ES = −0.345, p < 0.01). Originality/value The authors consider whether ideas-related discussion within politically homophilous networks is problematic for the IIS and what is required from education systems if we are to build individuals’ capacity to engage with ideas while escaping echo chambers.

MAPPING TUMBLR THROUGH FANNISH HOMOPHILIES

Using survey data from a 2022 survey of fans that was distributed on Tumblr between February 22 and March 21, 2022, this paper speculatively maps the network of Tumblr. Tumblr is home to fandoms of all kinds, acting as a space where fans can gather, discuss theories, produce fan texts and media, and interact with fanworks, making them one of the key distinctive communities that use the site and a suitable community to investigate the shape of the network. First, we analyze the prevalence of particular demographic answers from the survey over time to find points at which a large number of responses have similar identities. Second, we use the baseline probability of any given answer within our data set to examine times when it becomes disproportionately prevalent compared to elsewhere in the data. Through looking at these correlations between time and probability in specific demographic answers, we can identify moments when the survey was traveling through a homophilic network within the platform, and therefore extract information about how homophily works on this platform. By mapping Tumblr through these survey homophiles, we can gain a greater understanding of how people gather and interact on the platform. This has implications both for questions of community and subcultural belonging, but also issues like the spread of mis- and disinformation.

Alzheimer’s disease is associated with increased modularity and assortativity: Evidence from structural and metabolic connectivity

AbstractBackgroundAlzheimer’s disease (AD) is a progressive disorder associated with cognitive dysfunction that alters brain connectivity and assessing these altered connectivity patterns has become a topic of increasing interest. In this study, we examine the structural and metabolic connectivity using cortical thickness and [18F] Flouro deoxyglucose (FDG)‐positron emission tomography (PET) to identify the network modification in patient with AD when compared to mild cognitive impairment (MCI). To examine our hypothesis; we quantified a relatively newer graph theory measures including assortativity and modularity. We aimed to unravel the multimodal network alterations using simultaneous MRI‐PET acquisitions.MethodsPatients who underwent simultaneous FDG‐PET and MRI study on a 3T MRI were recruited for this exploratory study. After appropriate pre‐processing of structural‐MRI and PET‐metabolic data, graph measures were obtained using the BRAPH toolbox, and subsequently; global and nodal network measures were used to derive group‐level differences (corrected p<0.05). Regions that revealed significant differences were correlated with the cognitive scores (ACE‐III).Results83 demographically matched patients [MCI (n = 27), and AD (n = 56)] were analysed. There were no significant differences between the groups in the mean age (MCI 63.52 ±9.19 years and AD 63.73 ±9.60 years), gender (M: F) MCI 20:7 and AD 26:30, or the mean duration of disease (MCI 2.23±1.92 and AD 2.95 years) other than ACE scores (MCI‐87.30±5.53 and AD 46.02 ±20.61, p value <0.001) and the CDR (MCI‐very mild and AD‐Moderate and Severe). Patients with AD revealed significantly increased assortativity at the global level in both structural and metabolic connectivity analysis. In addition, metabolic connectivity revealed increased modularity and decreased participation coefficient in patients with AD. Cortical thickness and SUV abnormalities in several regions in bilateral parietotemporal lobes were proportional with the cognitive scores (ACE‐III).ConclusionProbable erroneous and homophilic network segregation paralleling the aging associated loss of whole brain integration was found proportional to the memory scores. Further studies in larger cohorts are required to ascertain the clinical relevance of these findings as an imaging biomarker.

The old-new epistemology of digital journalism: how algorithms and filter bubbles are (re)creating modern metanarratives

In journalism studies, the advent of the World Wide Web and the rise of online journalism are generally associated with going beyond the objective, normative paradigm associated with the principles of philosophical and scientific modernity towards a postmodern paradigm centred on subjectivity and relativism. This article offers an alternative reading of the epistemology of online journalism: the fragmentation of audiences into homophilic networks, the formation of ideological bubbles, and the growing polarisation caused by algorithms make the contents circulating online a reintroduction of modernity’s metanarratives. These metanarratives in no way correspond to the principles typical of postmodernism, such as the equivalence of interpretations and openness to dialogue. Journalistic content also comes under this charge: although it conveys narratives that are subjective, they are perceived as absolute truths inside the information bubbles in which they circulate. This phenomenon is caused by “information platformization” processes. Based on these premises, a new definition of online journalism is proposed: rather than “postmodern”, it can be better understood as a fulfilment of the foundational principles of modernism, but in a subjective form.

Partisan Memes as a Catalyst for Homophilous Networks

ABSTRACT Although diverse political networks are seen as democratically valuable, online social networks enable the construction and maintenance of networks that are less diverse. In this study, we explore the cultivation of like-minded networks through blocking those sharing counter-attitudinal partisan memes and engaging with pro-attitudinal partisan memes. We then test the efficacy of an intervention to reduce the spread of homophily-inducing partisan memes. We present four experiments. Study 1 establishes that people react differently to partisan memes than to partisan news. Studies 1–4 confirm that people react differently to pro- and counter-attitudinal memes. Studies 3 and 4 provide limited evidence that reminding people of the diversity of their online networks can reduce digital behaviors that produce more homophilous networks. The results provide initial evidence that partisan memes may give rise to a spiral of homophily.

Birds of a Feather Get Angrier Together: Social Media News Use and Social Media Political Homophily as Antecedents of Political Anger

A significant body of literature within political communication revolves around the constructive political virtues and blighting social and democratic consequences of political anger. For the most part, studies have focused on identifying the primary causes and antecedents of political anger. However, within the context of social media, fewer efforts have been devoted to clarifying how and what infuriates people about politics. Does social media news use relate to increased or reduced levels of political anger? Do social media political homophilic networks explain political anger? And to what extent does political homophily influence the potential effect of social media news use on citizens’ political anger levels—moderating effect? Results drawing on a two-wave U.S. survey dataset show that the frequency of social media news use alone has no direct effect on people’s increased political anger, whereas interacting in homophilic discussion and information networks on social media positively associates with anger. Furthermore, the relationship between social media news use and political anger is contingent upon social media political homophily. Those who report high levels of social media news use and very low levels of social media political homophily end up being less angry over time. Limitations and steps for future research are discussed in the manuscript.

Regulating Social Media as a Public Good: Limiting Epistemic Segregation

ABSTRACT The rise of social media has correlated with an increase in political polarization, which many perceive as a threat to public discourse and democratic governance. This paper presents a framework, drawing on social epistemology and the economic theory of public goods, to explain how social media can contribute to polarization, making us collectively poorer, even while it provides a preferable media experience for individual consumers. Collective knowledge and consensus is best served by having richly connected networks that are epistemically integrated: individuals with diverse levels of expertise should be relatively well connected to each other. In epistemically segregated networks, by contrast, we have reason to predict collective epistemic failures. Expert knowledge will be isolated from the majority, leading average opinion to be less informed than is socially optimal, and entrenching disagreements. Because social media enables users to very easily adopt homophilous network connections – connections to those with similar opinions, education levels, and social backgrounds – it is likely to have increased epistemic segregation compared to older media platforms. The paper explains the theoretical foundations of these predictions, and sketches regulatory measures – such as taxes – that might be employed to preserve the public good of a well integrated social media network.

Graph structure learning based on feature and label consistency

Graph Neural Networks (GNNs) have achieved remarkable success in graph-related tasks by combining node features and graph topology elegantly. Most GNNs assume that the networks are homophilous, which is not always true in the real world, i.e., structure noise or disassortative graphs. Only a few works focus on generalizing graph neural networks to heterophilous or low homophilous networks, where connected nodes may have different labels. In this paper, we design a simple and effective Graph Structure Learning strategy based on Feature and Label consistency (GSLFL) to increase the homophilous level of networks for generalizing any existing GNNs to heterophilous networks. Specifically, we first introduce a method to learn graph structure based on node features and then modify the graph structure based on label consistency. Further, we combine the GSLFL with three existing GNNs to learn node representations and graph structure together. And we design a self-training method to iteratively train models and modify graph structure with pseudo-labels. Finally, our empirical results on 6 public networks with homophily or heterophily, and structure attacks show that our methods outperform the state-of-the-art methods in most cases.

Bridging the polarization gap: Maximizing diffusion among dissimilar communities

Polarized networks, composed of weakly connected and self-reinforcing groups, can limit the diffusion of ideas, behaviors, and innovations. Here, we use a complex contagion model, in which diffusion depends on both the connectivity and the similarity of individuals, to ask how to optimally build bridges and enhance diffusion in networks characterized by fragmentation and homophily. First, we show that the problem is NP-hard. Then, we explore the space of solutions using heuristics, finding that connecting high degree nodes, or hubs, is an ineffective strategy to accelerate diffusion in fragmented and homophilous networks. We show that in these networks, diffusion is more effectively accelerated by connecting similar but low degree nodes. These results tell us that, in the presence of homophily and polarization, connecting communities through their most central actors may impede rather than facilitate diffusion. Instead, strategies to accelerate the diffusion of innovation, behaviors, and ideas should focus on creating links among the most similar members of different communities. These findings shed light on the diffusion of ideas and innovations in polarized networks. CCS Concepts: • Mathematics of computing → Network optimization; • Information systems → Social networks

Reproducing Hierarchies at the APSA Annual Meeting: Patterns of Panel Attendance by Gender, Race, and Ethnicity

ABSTRACTResearch on the political science profession has shown that homophilous research networks—that is, those organized along the lines of gender and race/ethnicity—reproduce hierarchies. Research networks composed of white men experience the most prestige and lead to the most opportunities. This study documents homophilous networks in a setting where they likely are nurtured: academic conferences. Drawing data from the 2019 Annual Meeting of the American Political Science Association, we examine the correspondence between the gender and the racial/ethnic composition of section members, panelists, and audience members for four research sections: Political Methodology; Political Psychology; Race, Ethnicity, and Politics; and Women and Politics. We find that attendees’ and panelists’ gender and racial/ethnic identity largely mirror the dominant gender and racial/ethnic group in their section. These findings indicate that homophily manifests at academic conferences and that efforts to diversify research networks should consider who listens to whom in these settings.

Preference or Persistence? The Origins and Implications of Homophily in Entrepreneurial Networks

Homophily in social networks is common across a range of social settings. However, the particular processes through which people construct homophilous social networks are not well understood. We theorize that two processes could lead to homophily in an individual’s social networks and test them using data from an entrepreneurial mentoring program. In one possible model, people are more likely to initiate relationships with similar others because they prefer those relationships. Alternatively, people could find that of the relationships they begin, homophilous relationships are more valuable, leading homophilous relationships to persist longer. Entrepreneurs in this program were given the opportunity to request which mentors they worked with and did not request mentors of their gender more frequently, indicating that were not more likely to initiate homophilous relationships. However, relationships between entrepreneurs and mentors of the same gender become more persistent working relationships and lead to more referrals on the entrepreneurs’ behalf, indicating that homophilous relationships were more valuable. Men mentors disproportionately made fewer referrals on behalf of female entrepreneurs. The results indicate that homophilous networks could emerge endogenously as entrepreneurs seek relationships that can offer valuable resources and provide insight into inequality in access to network-based resources.

Echo chambers and information transmission biases in homophilic and heterophilic networks

We study how information transmission biases arise by the interplay between the structural properties of the network and the dynamics of the information in synthetic scale-free homophilic/heterophilic networks. We provide simple mathematical tools to quantify these biases. Both Simple and Complex Contagion models are insufficient to predict significant biases. In contrast, a Hybrid Contagion model—in which both Simple and Complex Contagion occur—gives rise to three different homophily-dependent biases: emissivity and receptivity biases, and echo chambers. Simulations in an empirical network with high homophily confirm our findings. Our results shed light on the mechanisms that cause inequalities in the visibility of information sources, reduced access to information, and lack of communication among distinct groups.

Political homophily, bifurcated social reality, and perceived legitimacy of the 2020 US presidential election results: A four‐wave longitudinal study

AbstractWe report the results from a four‐wave longitudinal study (N = 951–1349) investigating perceptions of the 2020 US Presidential Election as legitimate. Specifically, the role of homophily (“love of same”) was examined in creating politically homogenous social and informational networks that facilitate and amplify polarized perceptions of the election's legitimacy and belief in voter fraud across time. Consistent with our hypotheses, Trump voters’ susceptibility to misinformation about voter fraud and refusals to accept the election results as legitimate drastically increased to the extent to which they were entrenched in politically homophilic networks. Biden voters modestly strengthened their views of the election results as legitimate when entrenched in politically homophilic networks. Moreover, the influence of politically homophilic networks in facilitating this polarization grew stronger with the passage of time. Longitudinal modeling revealed that, among Trump voters, politically homophilic networks decreased perceptions of election legitimacy through increased belief in voter fraud.

Simulating systematic bias in attributed social networks and its effect on rankings of minority nodes

Network analysis provides powerful tools to learn about a variety of social systems. However, most analyses implicitly assume that the considered relational data is error-free, and reliable and accurately reflects the system to be analysed. Especially if the network consists of multiple groups (e.g., genders, races), this assumption conflicts with a range of systematic biases, measurement errors and other inaccuracies that are well documented in the literature. To investigate the effects of such errors we introduce a framework for simulating systematic bias in attributed networks. Our framework enables us to model erroneous edge observations that are driven by external node attributes or errors arising from the (hidden) network structure itself. We exemplify how systematic inaccuracies distort conclusions drawn from network analyses on the task of minority representations in degree-based rankings. By analysing synthetic and real networks with varying homophily levels and group sizes, we find that the effect of introducing systematic edge errors depends on both the type of edge error and the level of homophily in the system: in heterophilic networks, minority representations in rankings are very sensitive to the type of systematic edge error. In contrast, in homophilic networks we find that minorities are at a disadvantage regardless of the type of error present. We thus conclude that the implications of systematic bias in edge data depend on an interplay between network topology and type of systematic error. This emphasises the need for an error model framework as developed here, which provides a first step towards studying the effects of systematic edge-uncertainty for various network analysis tasks.

Explaining classification performance and bias via network structure and sampling technique

Social networks are very important carriers of information. For instance, the political leaning of our friends can serve as a proxy to identify our own political preferences. This explanatory power is leveraged in many scenarios ranging from business decision-making to scientific research to infer missing attributes using machine learning. However, factors affecting the performance and the direction of bias of these algorithms are not well understood. To this end, we systematically study how structural properties of the network and the training sample influence the results of collective classification. Our main findings show that (i) mean classification performance can empirically and analytically be predicted by structural properties such as homophily, class balance, edge density and sample size, (ii) small training samples are enough for heterophilic networks to achieve high and unbiased classification performance, even with imperfect model estimates, (iii) homophilic networks are more prone to bias issues and low performance when group size differences increase, (iv) when sampling budgets are small, partial crawls achieve the most accurate model estimates, and degree sampling achieves the highest overall performance. Our findings help practitioners to better understand and evaluate their results when sampling budgets are small or when no ground-truth is available.

The paradox of second-order homophily in networks

Homophily—the tendency of nodes to connect to others of the same type—is a central issue in the study of networks. Here we take a local view of homophily, defining notions of first-order homophily of a node (its individual tendency to link to similar others) and second-order homophily of a node (the aggregate first-order homophily of its neighbors). Through this view, we find a surprising result for homophily values that applies with only minimal assumptions on the graph topology. It can be phrased most simply as “in a graph of red and blue nodes, red friends of red nodes are on average more homophilous than red friends of blue nodes”. This gap in averages defies simple intuitive explanations, applies to globally heterophilous and homophilous networks and is reminiscent of but structually distinct from the Friendship Paradox. The existence of this gap suggests intrinsic biases in homophily measurements between groups, and hence is relevant to empirical studies of homophily in networks.

Inference and influence of network structure using snapshot social behavior without network data.

Population behavior, like voting and vaccination, depends on the structure of social networks. This structure can differ depending on behavior type and is typically hidden. However, we do often have behavioral data, albeit only snapshots taken at one time point. We present a method jointly inferring a model for both network structure and human behavior using only snapshot population-level behavioral data. This exploits the simplicity of a few parameter model, geometric sociodemographic network model, and a spin-based model of behavior. We illustrate, for the European Union referendum and two London mayoral elections, how the model offers both prediction and the interpretation of the homophilic inclinations of the population. Beyond extracting behavior-specific network structure from behavioral datasets, our approach yields a framework linking inequalities and social preferences to behavioral outcomes. We illustrate potential network-sensitive policies: How changes to income inequality, social temperature, and homophilic preferences might have reduced polarization in a recent election.

Chasm in Hegemony: Explaining and Reproducing Disparities in Homophilous Networks

In networks with a minority and a majority community, it is well-studied that minorities are under-represented at the top of the social hierarchy. However, researchers are less clear about the representation of minorities from the lower levels of the hierarchy, where other disadvantages or vulnerabilities may exist. We offer a more complete picture of social disparities at each social level with empirical evidence that the minority representation exhibits two opposite phases: at the higher rungs of the social ladder, the representation of the minority community decreases; but, lower in the ladder, which is more populous, as you ascend, the representation of the minority community improves. We refer to this opposing phenomenon between the upper-level and lower-level as the chasm effect. Previous models of network growth with homophily fail to detect and explain the presence of this chasm effect. We analyze the interactions among a few well-observed network-growing mechanisms with a simple model to reveal the sufficient and necessary conditions for both phases in the chasm effect to occur. By generalizing the simple model naturally, we present a complete bi-affiliation bipartite network-growth model that could successfully capture disparities at all social levels and reproduce real social networks. Finally, we illustrate that addressing the chasm effect can create fairer systems with two applications in advertisement and fact-checks, thereby demonstrating the potential impact of the chasm effect on the future research of minority-majority disparities and fair algorithms.

Prevalence of mental disorders in a German kidney transplant population – results of a KTx360°-substudy

The spreading of epidemics in complex networks has been a subject of renewed interest of several scientific branches. In this regard, we have focused our attention on the study of the susceptible–infected–susceptible (SIS) model, within a Monte Carlo numerical simulation approach, representing the spreading of epidemics in a clustered homophilic network. The competition between infection and recovery that drives the system either to an absorbing or to an active phase is analyzed. We estimate the static critical exponents β∕ν, 1∕ν and γ∕ν, through finite-size scaling (FSS) analysis of the order parameter ρ and its fluctuations, showing that they differ from those associated with the contact process on a scale-free network, as well as those predicted by the heterogeneous mean-field theory.