Homophily and Contagion Are Generically Confounded in Observational Social Network Studies

Focus on Authors

Recent studies suggest that obesity may be “contagious” between individuals in social networks. Social contagion (influence), however, may not be identifiable using traditional statistical approaches because they cannot distinguish contagion from homophily (the propensity for individuals to select friends who are similar to themselves) or from shared environmental influences. In this paper, we apply the stochastic actor-based model (SABM) framework developed by Snijders and colleagues to data on adolescent body mass index (BMI), screen time, and playing active sports. Our primary hypothesis was that social influences on adolescent body size and related behaviors are independent of friend selection. Employing the SABM, we simultaneously modeled network dynamics (friendship selection based on homophily and structural characteristics of the network) and social influence. We focused on the 2 largest schools in the National Longitudinal Study of Adolescent Health (Add Health) and held the school environment constant by examining the 2 school networks separately (N = 624 and 1151). Results show support in both schools for homophily on BMI, but also for social influence on BMI. There was no evidence of homophily on screen time in either school, while only one of the schools showed homophily on playing active sports. There was, however, evidence of social influence on screen time in one of the schools, and playing active sports in both schools. These results suggest that both homophily and social influence are important in understanding patterns of adolescent obesity. Intervention efforts should take into consideration peers’ influence on one another, rather than treating “high risk” adolescents in isolation.

Researchers have already observed social contagion effects in both in-person and online interactions. However, such studies have primarily focused on users’ beliefs, mental states, and interests. In this article, we expand the state of the art by exploring the impact of social contagion on social alignment, i.e., whether the decision to socially align oneself with the general opinion of the users on the social network is contagious to one’s connections on the network or not. The novelty of our work in this article includes: 1) unlike earlier work, this article is among the first to explore the contagiousness of the concept of social alignment on social networks; 2) our work adopts an instrumental variable approach to determine reliable causal relations between observed social contagion effects on the social network; and 3) our work expands beyond the mere presence of contagion in social alignment and also explores the role of population heterogeneity on social alignment contagion. Based on the systematic collection and analysis of data from two large social network platforms, namely, Twitter and Foursquare, we find that a user’s decision to socially align or distance from social topics and sentiments influences the social alignment decisions of their connections on the social network. We further find that such social alignment decisions are significantly impacted by population heterogeneity.

For more than a decade social media have connected people virtually, thereby creating digital social networks with different purposes and affordances. People typically co‐orient, leading to some content going viral and becoming “contagious” for other network members. Importantly, the complex contagion phenomenon can now be captured more easily using digital tools that became part of the method arsenal in computational social science. Conceptually, social contagion describes a nonpermanent event that, when it occurs, develops over time with specific social network dynamics. It can be assessed using both observational and controlled study designs that, importantly, also come with novel ethical challenges that need to be considered. Identifying social media influencers using centrality measures can be the starting point for measuring social media contagion, thus understanding the social contagion dynamics as “rippling effects” following a “three degrees of influence rule” within social networks of connected members, friends, and friends of friends. Over time, the contagion establishes itself as part of temporal clustering and as a result of a human tendency for assortative mixing and homophilic preferences. In combination with centrality measures and indicators for network homogeneity, the lifetime of social media content is a good measure to further describe not only the dynamics behind the adoption of viral content on social media, but also the emergence of echo chambers.

People often adjust their behaviours and preferences to align with others, a phenomenon termed social influence or social contagion. Social influence can have a profound effect on our economic preferences, with important implications for financial decision-making and making rational choices. Here we review recent findings in neuroeconomics, elucidating the psychological, computational, and neural mechanisms that underpin social influence during value-based decision-making. We examine how others’ decisions can shift our own economic preferences, regarding risk-taking, temporal discounting, effort discounting, and moral preferences. Additionally, we discuss the lifespan similarities and differences in susceptibility to social influence. Finally, we cover studies revealing the neural basis of susceptibility to social influence and highlight key common brain areas across studies that could underpin these effects. We end by considering the implications of strong evidence of susceptibility to social influence on economic preferences, the challenges and opportunities of this influence and how future studies could harness it to create positive changes across the lifespan.

AboutSectionsView PDF ToolsAdd to favoritesDownload CitationsTrack CitationsPermissionsReprints ShareShare onFacebookTwitterLinked InEmail Go to Section HomeMarketing ScienceVol. 30, No. 2 Commentary—Invited Comment on “Opinion Leadership and Social Contagion in New Product Diffusion”David GodesDavid GodesPublished Online:10 Dec 2010https://doi.org/10.1287/mksc.1100.0605 Previous Back to Top Next FiguresReferencesRelatedInformationCited byImpacts of normative and hedonic motivations on continuous knowledge contribution in virtual community: the moderating effect of past contribution experience10 February 2023 | Information Technology & People, Vol. 22Relationships among immigrant consumers' cultural orientation, innovativeness and opinion leadership15 November 2021 | International Marketing Review, Vol. ahead-of-print, No. ahead-of-printConsumer susceptibility to social influence in new product diffusion networks: how does network location matter?31 December 2020 | European Journal of Marketing, Vol. 55, No. 5Social Media Firm Specific Advantages as Enablers of Network Embeddedness of International Entrepreneurial VenturesJournal of World Business, Vol. 56, No. 3Leaders that bind: the role of network position and network density in opinion leaders' responsiveness to social influenceAsia Pacific Journal of Marketing and Logistics, Vol. ahead-of-print, No. ahead-of-printCreating Social Contagion Through Firm-Mediated Message Design: Evidence from a Randomized Field ExperimentTianshu Sun, Siva Viswanathan, Elena Zheleva5 August 2020 | Management Science, Vol. 67, No. 2Improving the value of the retailer brand through social media equity18 April 2020 | Journal of Brand Management, Vol. 27, No. 5Social interactions impact on product and service development27 July 2020 | Proceedings of the International Conference on Business Excellence, Vol. 14, No. 1Who is tech savvy? Exploring the adoption of smartphones and tablets: An empirical investigationThe Journal of High Technology Management Research, Vol. 30, No. 2Upward intergenerational influences on parents' innovativeness and innovation adoption: A comparative study of single‐ and multiple‐child families21 June 2019 | Journal of Consumer Behaviour, Vol. 18, No. 4Individual-level social influence identification in social media: A learning-simulation coordinated methodEuropean Journal of Operational Research, Vol. 273, No. 3Social Media Communications and Marketing Strategy: A Taxonomical Review of Potential Explanatory Approaches8 February 2019 | Journal of Internet Commerce, Vol. 18, No. 1Can members’ WeChat engagement affect relational outcomes in community-supported agriculture? Empirical evidence from China5 November 2018 | Agribusiness, Vol. 35, No. 1I Spot, I Adopt! A Discrete Choice Analysis on Peer Effects in Solar Photovoltaic System Adoption of HouseholdsSSRN Electronic Journal, Vol. 65Socially Nudged: A Quasi-Experimental Study of Friends’ Social Influence in Online Product RatingsChong (Alex) Wang, Xiaoquan (Michael) Zhang, Il-Horn Hann10 May 2018 | Information Systems Research, Vol. 29, No. 3Preaching to the Choir: The Chasm Between Top-Ranked Reviewers, Mainstream Customers, and Product SalesElham Yazdani, Shyam Gopinath, Steve Carson22 August 2018 | Marketing Science, Vol. 37, No. 5How Customer Referral Programs Turn Social Capital into Economic Capital1 February 2018 | Journal of Marketing Research, Vol. 55, No. 1Projecting service quality: The effects of social media reviews on service perceptionInternational Journal of Hospitality Management, Vol. 69When are influentials equally influenceable? The strength of strong ties in new product adoptionJournal of Business Research, Vol. 82Social Dollars in Online Communities: The Effect of Product, User, and Network Characteristics1 January 2018 | Journal of Marketing, Vol. 82, No. 1Marketing on TumblrFrom conformity to reactance: Contingent role of network centrality in consumer-to-consumer influenceJournal of Business Research, Vol. 75Marketing on TumblrIndicators of opinion leadership in customer networks: self-reports and degree centrality28 April 2015 | Marketing Letters, Vol. 27, No. 3Adaptive personalization using social networks1 May 2015 | Journal of the Academy of Marketing Science, Vol. 44, No. 1From Social to Sale: The Effects of Firm-Generated Content in Social Media on Customer Behavior1 January 2016 | Journal of Marketing, Vol. 80, No. 1Social networks in marketing research 2001–2014: a co-word analysis15 August 2015 | Scientometrics, Vol. 105, No. 1Word of Mouth and Interpersonal CommunicationSocial Contagion in New Product Trial and RepeatRaghuram Iyengar, Christophe Van den Bulte, Jae Young Lee19 January 2015 | Marketing Science, Vol. 34, No. 3Nonmonotonic Status Effects in New Product AdoptionYansong Hu, Christophe Van den Bulte13 May 2014 | Marketing Science, Vol. 33, No. 4Dynamic Effects of Social Influence and Direct Marketing on the Adoption of High-Technology Products1 March 2014 | Journal of Marketing, Vol. 78, No. 2Predicting Individual Behavior with Social NetworksSharad Goel, Daniel G. Goldstein24 October 2013 | Marketing Science, Vol. 33, No. 1Creating Social Contagion Through Firm Mediated Message Design: Evidence from a Randomized Field ExperimentSSRN Electronic Journal, Vol. 106Leveraging Peer-to-Peer Networks in Pharmaceutical Marketing26 October 2013Research Commentary —Information in Digital, Economic, and Social NetworksInformation Systems Research, Vol. 24, No. 4Rejoinder—Further Reflections on Studying Social Influence in New Product DiffusionRaghuram Iyengar, Christophe Van den Bulte, Thomas W. Valente, 10 December 2010 | Marketing Science, Vol. 30, No. 2Marketing on Tumblr Volume 30, Issue 2March-April 2011Pages 195-388 Article Information Metrics Information Received:June 24, 2010Accepted:June 27, 2010Published Online:December 10, 2010 Copyright © 2011, INFORMSCite asDavid Godes, (2010) Commentary—Invited Comment on “Opinion Leadership and Social Contagion in New Product Diffusion”. Marketing Science 30(2):224-229. https://doi.org/10.1287/mksc.1100.0605 Keywordsdiffusion of innovationsopinion leadershipsocial contagionsocial networksPDF download

Social network analysis has been postulated as a tool to study potential pathogen transmission in wildlife but is resource-intensive to quantify. Networks based on bacterial genotypes have been proposed as a cost-effective method for estimating social or transmission network based on the assumption that individuals in close contact will share commensal bacteria. However, the use of network analysis to study wild populations requires critical evaluation of the assumptions and parameters these models are founded on. We test (a) whether networks of commensal bacterial sharing are related to hosts' social associations and hence could act as a proxy for estimating transmission networks, (b) how the parameters chosen to define host associations and delineate bacterial genotypes impact inference and (c) whether these relationships change across time. We use stochastic simulations to evaluate how uncertainty in parameter choice affects network structure. We focused on a well-studied population of eastern grey kangaroos (Macropus giganteus), from Sundown National Park, Australia. Using natural markings, each individual was identified and its associations with other kangaroos recorded through direct field observations over 2years to construct social networks. Faecal samples were collected, Escherichia coli was cultured and genotyped using BOX-PCR, and bacterial networks were constructed. Two individuals were connected in the bacterial network if they shared at least one E.coli genotype. We determined the capacity of bacterial networks to predict the observed social network structure in each year. We found little support for a relationship between social association and dyadic commensal bacterial similarity. Thresholds to determine host associations and similarity cut-off values used to define E.coli genotypes had important ramifications for inferring links between individuals. In fact, we found that inferences can show opposite patterns based on the chosen thresholds. Moreover, no similarity in overall bacterial network structure was detected between years. Although empirical disease transmission data are often unavailable in wildlife populations, both bacterial networks and social networks have limitations in representing the mode of transmission of a pathogen. Our results suggest that caution is needed when designing such studies and interpreting results.

Social network influences of alcohol and marijuana cognitive associations

Online communities have experienced burgeoning popularity over the last decade and have become a key platform for users to share information and interests, and to engage in social interactions. Drawing on the social contagion literature, the authors examine the effect of online social connections on users’ product purchases in an online community. They assess how product, user, and network characteristics influence the social contagion effect in users’ spending behavior. The authors use a unique large-scale data set from a popular massively multiplayer online role-playing game community—consisting of users’ detailed gaming activities, their social connections, and their in-game purchases of functional and hedonic products—to examine the impact of gamers’ social networks on their purchase behavior. The analysis, based on a double-hurdle model that captures gamers’ decisions of playing and spending levels, reveals evidence of “social dollars,” whereby social interaction between gamers in the community increases their in-game product purchases. Interestingly, the results indicate that social influence varies across different types of products. Specifically, the effect of a focal user's network ties on his or her spending on hedonic products is greater than the effect of network ties on the focal user's spending on functional products. Furthermore, the authors find that user experience negatively moderates social contagion for functional products, whereas it positively moderates contagion for hedonic products. In addition, dense networks enhance contagion over functional product purchases, whereas they mitigate the social influence effect over hedonic product purchases. The authors perform a series of tests and robustness checks to rule out the effect of confounding factors. They supplement their econometric analyses with dynamic matching techniques and estimate average treatment effects. The results of the study have implications for both theory and practice and help provide insights on how managers can monetize social networks and use social information to increase user engagement in online communities.

There are many studies in the marketing and diffusion literature of the conditions in which social contagion affects adoption processes. Yet most of these studies assume that social interactions do not change over time, even though actors in social networks exhibit different likelihoods of being influenced across the diffusion period. Rooted in physics and epidemiology theories, this study proposes a Susceptible Infectious Susceptible (SIS) model to assess the role of social contagion in adoption processes, which takes changes in social dynamics over time into account. To study the adoption over a span of ten years, the authors used detailed data sets from a community of consumers and determined the importance of social contagion, as well as how the interplay of social and non-social influences from outside the community drives adoption processes. Although social contagion matters for diffusion, it is less relevant in shaping adoption when the study also includes social dynamics among members of the community. This finding is relevant for managers and entrepreneurs who trust in word-of-mouth marketing campaigns whose effect may be overestimated if marketers fail to acknowledge variations in social interactions.

Social influence and reduction of animal protein consumption among young adults: Insights from a socio-psychological model

Although employees increasingly need support to reconcile work and family, many lack a thorough knowledge of work-life practices such as flexible work arrangements, leaves, and dependent care programs, or they hesitate to use them. Building on social network and social contagion research, this paper argues that employees assess work-life practices not in isolation but through relational processes of social priming, social influence, and social comparison. I delineate six dimensions along which employees assess work-life practices–visibility, relevance, employer's motivations, instrumentality, fairness, and relative generosity–and analyze social contagion processes in networks of strong and weak ties, expressive and instrumental ties, within and outside the organization. I then examine how the national context may intervene in these processes by making the information that flows across ties more or less gender normative and by setting employees’ expectations for employer work-life support.

This study extends earlier research suggesting that board network ties may reflect the strategic and/or political concerns of top managers by considering how the managerial objectives that drive the formation and maintenance of board interlock ties may be subject to social influence. The particular form of social influence examined in this study derives from the social network research process itself. Specifically, we draw from research on social information processing and the framing of information to suggest how the administration of social network surveys can influence managers’ perceptions about their relationship to directors and the potential benefits to be derived from director network ties, thus affecting their subsequent selection of board members in ways that change the firm’s board interlock ties.We also consider how this social influence effect may diffuse beyond the actual survey respondents to create a more pervasive influence on the actions of managers at other firms in the board interlock network. We test our theoretical argument with an original quasiexperiment in which CEOs are randomly assigned to different versions of a survey questionnaire that have the potential to prime different schemata about the possible benefits to be derived from board network ties. Beyond addressing the potential for social influence in the formation and maintenance of board network ties, our study also addresses the potential for unintended reactive measurement effects in social network research, wherein network surveys influence the very ties that they are designed to measure.

We describe and validate a system for monitoring social contagions on Twitter: social movements, rumors, and emotional outbursts that spread from person to person in a viral manner. We use Twitter streams to monitor the spread of these phenomena through human social and information networks. This system, the contagion monitor, parses Twitter posts to identify emerging phenomena, as captured in hashtags, URLs, words and phrases, or account-handles, and then determines the extent to which a particular phenomenon spreads via the social network (in contrast to its spread via news broadcasts or independent adoption) and locates the contagion within Twitter communities. The monitor approximates the adoption threshold of a social contagion by measuring the fraction of Twitter users who were infected by the contagion (e.g., joined a particular social movement) after more than one of their friends had done so. Finally, the monitor makes a judgment about whether the phenomenon has reached critical mass, which is defined as the point where a social contagion begins spreading rapidly and breaches the social boundaries of its early adopter group. We test our prototype monitor on two data sources --- an ongoing stream of tweets grouped by user-added hashtags and a collection of posts by a monitored set of Nigerian Twitter users --- before productionalizing. We use the Amazon Mechanical Turk platform to evaluate the performance on both data sources. In both cases, we find that our approach successfully distinguishes between high-threshold and low-threshold social contagions.

Social networks are becoming more and more popular with the advent of numerous online social networking services. In this paper, we explore social rating networks, which record not only social relations but also user ratings for items. We analyze and model the effects of social influence and correlational influence in such networks, based on influence coefficients that measure the degree of influence in a network. We distinguish two types of user behavior: adopting an item and adopting a rating value for that item. We propose models to analyze and measure the influence of neighbors on both item and rating adoption behavior of users. Our experiments demonstrate that social influence has a much stronger impact on user behavior than correlational influence. Social and correlational influence are global effects in the entire network. However, there are local differences, i.e. certain users have a stronger social influence than others. To model this effect, we introduce the novel concept of social authority of individual users. We also propose an objective way to evaluate the social authority measure by injecting it into a simple recommender system.