Homophily explains perception biases in social networks.
Year: 2017
Type: preprint
Abstract: Individual's perceptions about the prevalence of attributes in their social networks is commonly skewed by the limited information available to them. Filter bubbles -- being exposed to other like-minded people -- and majority illusion -- overestimation of minorities in social networks -- are two examples of how perception biases can manifest. In this paper, we show how homophily and disproportionate group sizes influence the emergence of perception biases in social networks. Using a generative network model with adjustable homophily and group size, we demonstrate analytically and numerically under which conditions and to what extent perception biases can emerge. We compare these theoretical results with empirical investigations of perception biases in six real-world networks with various levels of homophily and group sizes. Our results show (i) that perception biases can emerge in social networks with high homophily or high heterophily and unequal group sizes, (ii) that these effects are highly related to the asymmetric nature of homophily in networks and that (iii) the perception of nodes is not systematically distorted or enhanced by their degree. Finally, we explore under which structural conditions individuals can reduce their perception bias by taking the perception of their direct neighbors into account. These results advance our understanding of the impact of network structure on perception biases and offer a quantitative approach to address this issue in society.
Source: arXiv (Cornell University)
Language: English
Subfield: Statistical and Nonlinear Physics
Field: Physics and Astronomy
Domain: Physical Sciences
Open Access status: green
OpenAlex ID: W2765924671