Coloring in the Links

Abstract

The richness that characterizes relationships is often absent when they are modeled using computational methods in network science. Typically, relationships are represented simply as links, perhaps with weights. The lack of finer granularity is due in part to the fact that, aside from linkage and strength, no fundamental or immediately obvious dimensions exist along which to categorize relationships. Here we propose a set of dimensions that capture major components of many relationships -derived both from relevant academic literature and people's everyday descriptions of their relationships. We first review prominent findings in sociology and social psychology, highlighting dimensions that have been widely used to categorize social relationships. Next, we examine the validity of these dimensions empirically in two crowd-sourced experiments. Ultimately, we arrive at a set of ten major dimensions that can be used to categorize relationships: similarity, trust, romance, social support, identity, respect, knowledge exchange, power, fun, and conflict. These ten dimensions, while not dispositive, offer higher resolution than existing models. Indeed, we show that one can more accurately predict missing links in a social graph by using these dimensions than by using a state-of-theart link embeddedness method. We also describe tinghy.org, an online platform we built to collect data about how social media users perceive their online relationships, allowing us to examine these dimensions at scale. Overall, by proposing a new way of modeling social graphs, our work aims to contribute both to theory in network science and practice in the design of social-networking applications. S. Deri et al.most visited website in the world, behind only Google and Youtube [7]; and, on average, users were estimated to spend almost an hour per day on the site [84].Unsurprisingly, researchers have spent a great deal of time trying to capture and model social relationships both online and offline [2]. One of the dominant paradigms that has emerged is to think of and model social relationships in terms of tie strength [48]. Such approaches have no doubt been fruitful and certainly capture a key element of social relationships. For example, weak ties with acquaintances are critical in finding a job or generating creative ideas, while strong ties with family or close friends provide crucial emotional and social support. Tie strength is even predictive of subjective well-being [20].Yet, the focus on tie strength as a way of understanding and modeling social relationships has its limitations. Not all ties of the same strength are created equal [93]. Weak ties can be the way through which new job opportunities are discovered by professionals but not by blue collar workers; for people of low status, weak ties of a similar status are not generally useful or far reaching. Meanwhile, strong ties with parents offer financial support but rarely companionship, while strong ties with friends offer the opposite. Many other factors such as trust, power, and gr...

The anatomy of a population-scale social network

“…Previous work pointed towards the necessity of distinguishing edge types in order to adequately understand overlapping community structure of social networks 1 , or the measurement and modeling of weak ties and their importance in network cohesion 43 , 45 . For automatically collected large-scale networks, it is sometimes possible to infer edge types from the communication content 26 or to capture different interactions between nodes by the data collection design 36 , 53 . However, these examples remain exceptions and limited to the particular measurement setting used in these studies, or include only a small number of nodes (see 6 , 51 and 27 , Chapter 3).…”

The Persuasive Power of Large Language Models

“…The theory posits that a speaker can enhance their chances of changing the hearer's mind by loading their arguments with the appropriate intent, for example by conveying trust and willingness to share knowledge (Habermas 1979). Prior research has identified universal dimensions of social pragmatics (Deri et al 2018), and developed a transformer-based tool to reliably capture the presence of these dimensions in conversational language (Choi et al 2020). The tool was tested on online debates for which a ground truth of successful arguments was available, showing that the most persuasive arguments are characterized by the dimensions indicated by the theory: factual knowledge, expression of trust, and appeals to the similarity between points of view (Monti et al 2022).…”

On the Conversational Persuasiveness of Large Language Models: A Randomized Controlled Trial

“…Social Dimensions. We then consider as features the social dimensions introduced by [79], a set of universal categories of social pragmatics. Previous research has analyzed the presence of these dimensions in language and conversations, finding them to be highly predictive of opinion change in online debates [70,80,81].…”