The Best Student Paper Award is given annually to a graduate student in recognition of research on social network analysis. The award is supported by royalties donated by the authors of the chapters in Models and Methods in Social Network Analysis edited by P.J. Carrington, J. Scott, and S. Wasserman, published by Cambridge University Press in 2005. INSNA wishes to thank the contributors for their generosity.
The award is for research by a graduate student and is administered by INSNA, the International Network for Social Network Analysis. For the 2018 competition, students should submit a paper (pdf file written in English) to the committee before 30 March 2018. The winner of the award will be announced at the 2018 Sunbelt conference. To be eligible, the student must be a member of INSNA and the first (or sole) author on the paper at the time of submission. The paper may be published or unpublished and must have been completed within two years of the submission deadline.
A letter of support from a supervisor, mentor or other faculty member should be submitted as well. The support letter should be aimed at helping the judges understand the significance of the paper. Papers are evaluated by a committee of three judges. Their decision is final and is based on the level of originality in the ideas and techniques, the possible applications and their treatment, and potential impact. The committee may arrive at the conclusion that none of the submitted papers merits the award.
The monetary prize for the best student paper is $2,500.
Please send all application materials to INSNA President, Steve Borgatti, firstname.lastname@example.org. Make sure to put INSNA Student Paper Award in the subject line to ensure your submission is not missed.
"Social influence on 5-year survival in a longitudinal chemotherapy ward co-presence network"
Jeffrey Lienert, Christopher Steven Marcum, John Finney, Felix Reed-Tsochas, and Laura Koehly
Chemotherapy is often administered in openly-designed hospital wards, where there is the possibility of social influence on health between patients. Previous research has found evidence that cancer patients’ health is impacted by social relationships; however, social influence has not been examined in patients while receiving treatment in a chemotherapy ward. In the current paper, we investigate the influence of co-presence vis-à-vis cancer patient outcomes on five-year survival in a chemotherapy ward. Using data on 4,691 cancer patients undergoing chemotherapy in Oxfordshire, UK, we construct a network of patients where edges between patients equal the Jaccard index of co-presence in the ward. Patients averaged 59.8 years of age, and 44% were male. We count the total Jaccard-weighted person-hours of overlap with focal patients’ immediate neighbors or those two nodes away who finish their chemotherapy cycle and survive 5 years or die within 5 years. Generalized Estimating Equations were used to evaluate the effect of neighbors’ outcomes on focal patient’s 5-year mortality. Each 1,000-unit increase in Jaccard-weighted person-hours (e.g. a hypothetical focal patient in the ward for 50 hours co-present with 100 patients, all with Jaccard indices of 0.2) with a patient dying within 5 years increases a patient’s mortality odds by 42% (β = 0.357 CI:0.204,0.510). Each 1,000-unit increase in Jaccard-weighted person-hours with a patient surviving 5 years reduces a patient’s odds of dying by 30% (β = -0.344, CI:-0.538,0.149). Our results suggest that social influence occurs in chemotherapy wards, which may need to be taken into account in chemotherapy delivery.
"Network Periphery, Group Boundaries, and the 40 Individuals in the 'Toronto 18' Terrorist Network"
Simon Fraser University
Simon Fraser University
Ouellet's study as a mixed-method approach to data collection using a combination of archival and informant sources (though, as mentioned, this is only a single slice through the cognitive social structure of these associations). The paper is well-written and accessible with a focus on a key problem in network studies, the boundary specification problem, as this problem appears in research into covert networks. The substantive contribution lies in its study of the process of radicalization at the group level using network data at multiple time points and highlighting the complex mix of motives of participants in the conspiracy.
The review committee also recommended that an honorable mention award be given to András Vörös for "Cluster Analysis of Multiplex Networks: Defining Composite Measures" co-authored with Tom A.B. Snijders. The committee saw much merit in the methodological advances proposed and illustrated in the paper.
"Inter-Ethnic Friendship adn Negative Ties in Secondary School"
Nuffield College, Oxford University, and MTA TK 'Lendület' Research Center for Educational and Network Studies
Institute for Sociology, Hungarian Academy of Sciences, Corvinus University of Budapest, and MTA TK 'Lendület' Research Center for Educational and Network Studies
Boda and Néray's paper presents a coherent theoretical foundation that informs a systematically constructed, longitudinal network study on inter-ethnic friendships and negative ties among secondary school students, with a focus on two ethnic groups: Roma and non-Roma Hungarian. It has the merit of collecting an extensive amount of data, including for example data on actors' perceptions, self-declared ethnicity and peer-based perception of ethnicity. Friendships and negative ties are modelled using cross-sectional exponential random graph models for sixteen classrooms separately, and then individual models are summarized using meta-analysis. Results emphasize the fact that different ethnicity aspects influence friendships and negative ties in different ways, and that inconsistencies in someone's ethnic categorization might play an important role in social rejection. The originality of looking at the impact of negative/positive ties combined with complex identifications indicates that the paper will make a significant contribution to the study of inter-ethnic relationships, and the use of meta-analysis provides a robust theoretical generalizability to the empirical results.
"Reciprocity, Transitivity, and the Mysterious Three-Cycle"
Nuffield College, Oxford University
Per Block shows that a common finding in much research on friendship networks, namely, the existence of a tendency against the formation of three-cycles, is spurious and the result of failing to control for differential tendencies for reciprocity within and between transitive groups. Specifically, an analysis of 30 friendship networks using stochastic actor-oriented models confirms theoretical expectations of a negative interaction effect between reciprocity and transitivity. The results mean that the common interpretation of three-cycle effects as indicating that unreciprocated friendships derive from hierarchical status differences between individuals must be re-thought.
The committee's decision was made difficult by the high quality of all 15 papers submitted. In the final analysis, the choice came down to two submissions and the committee would like to recognize that fact and so recommends that the paper "Exponential Random Graph Models for Multilevel Networks" by Peng Wang with Garry Robins, Philippa Pattison, and Emmanuel Lazega be given honorable mention. It is an outstanding methodological paper associated with a research team that has done an enormous amount of work to advance statistical modeling of network data. Given the recent focus on multilevel models in the SNA community, it is sure to have a large number of possible applications.
“Macrostructure from Microstructure : Generating Whole Systems from Ego Networks”
Jeffrey A. Smith
In this paper Jeffrey Smith uses a simple algorithmic idea to estimate global networks from local network samples. He adjusts the parameters of an ERGM to build a model consistent with the entire distribution of observed local statistics in the network, thus making optimal use of the local network data. In doing so, he contributes to the field’s pressing search for good network sampling tools.
The committee was very pleased with the quality of the submissions. All 16 papers submitted were interesting scientific contributions to their specific field of studies, even though their sophistication in the use of specific network analytic tools may have varied considerably. Faced with a difficult choice, the committee ultimately decided to give priority to papers that combined a distinctive contribution to SNA with attention to a substantive problem, of broader interest to social scientists. The committee also recommends that the paper “Circular Migrations and HIV Transmission: A Comparison of Compartmental and Network Modeling,” by Aditya S. Khanna with Dobromir T. Dimitrov and Steven M. Goodreau be given honorable mention. In this paper Khanna and his co-authors investigate the impact of circular migrations on HIV transmission dynamics. They draw on the distinct mathematical frameworks of compartmental and network-based methods to study the interaction of sexual partnership structure with acute HIV infection. The paper illustrates how the effects of concurrency may be captured using network models.
“Social Selection and Peer Influence in an Online Social Network”
Harvard University and Santa Rosa Junior College
This paper by Mr. Lewis is a highly polished piece of research. The paper has a good balance between methodological sophistication and the capacity to address a substantive question that has plagued network science for years – social selection vs. social influence. Mr. Lewis conducts a very thorough analysis with a really rich dataset. The committee felt this paper has significant potential impact which will likely reach beyond the network community. The committee also recommends that the paper “Structuring Principles of Board Networks: Social Preferences and Board Capital as Determinants of Director Selection” by Julia Brennecke with Olaf N. Rank, and Anja Tuschke be given honorable mention. Ms. Brennecke does an excellent job articulating the theory that drives model choice and empirical analyses of the data. Further, she applies innovative network methodologies to analyze her data, resulting in a paper that has the potential to impact business research.
"The Co-evolution of Gossip and Friendship at Work: Studying the Dynamics of Multiplex Social Networks"
University of Groningen/ICS
University of Groningen/ICS
University of Groningen/ICS
This study investigates the co-evolution of friendship and gossip in organizations. Two contradicting theories are tested. Social capital theory predicts that friendship causes gossip between employees, defined as informal evaluative talking about absent colleagues. Evolutionary theory reverses this causality claiming that gossiping facilitates friendship. The data comprises of three observations of a complete organizational network, allowing longitudinal social network analyses and causal inferences. Gossip and friendship are modeled as both explanatory and outcome networks with Multiple SIENA. Results support evolutionary theory, as gossip increases friendship formation in dyads. However, high gossip activity decreases the number of friends in the group.
"Analysing Event Stream Dynamics in Two Mode Networks"
KIT Karlsruhe Institute of Technology
KIT Karlsruhe Institute of Technology
Exponential Random Graph Models (ERGMs) are widely used to describe structural network dynamics using panel data. In this paper the private communication behaviour within a question and answer community is analysed. Using this example, it is shown how event based extensions of the standard ERGM models can be applied on combined communication and affiliation networks. It is tested in how far communication patterns and affiliation patterns influence the choice of message receivers. To visualize change and evolution in event based networks, a sliding window approach is introduced and discussed.
"Respondent-Driven Sampling: An Assessment of Current Methodology"
Krista J. Gile
University of Washington
University of Washington
Respondent-Driven Sampling (RDS) employs a variant of a link-tracing network sampling strategy to collect data from hard-to-reach populations. By tracing the links in the underlying social network, the process exploits the social structure to expand the sample and reduce its dependence on the initial (convenience) sample. Current estimation focuses on estimating population averages in the hard-to-reach population. These estimates are based on strong assumptions allowing the sample to be treated as a probability sample. In particular, we focus on three critical sensitivities of the estimators: to the without-replacement structure of sampling, to bias induced by the initial sample, and to uncontrollable features of respondent behavior. First, estimates are based on a with-replacement random walk model, while the actual sampling is without replacement. We illustrate that when over half of the target population is sampled this approximation can lead to substantial bias in the resulting estimators. Previous research on the properties of RDS estimators has not considered this assumption. Second, we address the reduction of bias induced by the convenience sampling of the initial sample. RDS relies on many sample waves to create a type of mixing in the sampling process, much like the mixing in a Markov chain. We illustrate that the number of sample waves typically used in RDS is likely insufficient for the type of nodal mixing required to obtain the reputed asymptotic unbiasedness of the estimators. In some cases, however, we find that despite this, the resulting estimators are approximately unbiased, although this is highly sensitive to the degree of clustering in the population and the number of waves in the sample. Finally, we highlight the dependence of the estimators on characteristics of respondent behavior outside the control of the researcher. In particular, we illustrate the bias induced in the estimator by preferential coupon-passing behavior by respondents. We highlight the need to expand data collection to learn more about how respondents behave in RDS studies. This paper sounds a cautionary note for the users of RDS. While current RDS methodology is powerful and clever, the good statistical properties claimed for the current estimates are shown to be heavily dependent on often unrealistic assumptions.