Examining whether and how instructional coordination occurs within introductory undergraduate STE... more Examining whether and how instructional coordination occurs within introductory undergraduate STEM courses.
Status allocation is a process by which complex labor market processes allocate individuals with ... more Status allocation is a process by which complex labor market processes allocate individuals with varying credentials and background characteristics to occupational positions that vary systematically in prestige, autonomy, and compensation. Previous literature examines two principles of allocation, the meritocratic principle and the lottery or chance principle, to assess departures of an observed distribution from pure meritocratic versus pure chance distributions (Krauze and Słomczyński, 1985). In this paper, we model an observed distribution as a compound of elementary distributions, meritocratic and chance, governed by a mixing parameter that expresses, in effect, the proportion of the actual allocation attributable to one principle versus another. We generalize the constant mixing model to a differential mixing model in which the mixing parameters are origin-specific. We explore the insights our framework enables with two comprehensive data sets on educational origins and occupational destinations, one covering the EU for over 30 countries from 1992 to 2018 and the other from IPUMS for both EU and other nations. We find that the differential model typically fits better than the constant model, that mixing coefficient for the top origin category is much larger than the mixing coefficient for lower categories, that the allocation of women is generally more meritocratic than the allocation of men, that there is fluctuation over time in the degree to which an observed allocation fits a meritocratic pattern, and that countries differ widely how much meritocracy versus lottery contributes to status allocation.
2019 IEEE International Conference on Systems, Man and Cybernetics (SMC), 2019
This paper introduces a decomposition approach to address the problem of predicting different use... more This paper introduces a decomposition approach to address the problem of predicting different user activities at hour granularity over a long period of time. Our approach involves two steps. First, we used a temporal neural network ensemble to predict the number of each type of activity that occurred in a day. Second, we used a set of neural networks to assign the events to a user-repository pair in a particular hour. We focused this work on a subset of the public GitHub dataset that records the activities of over 2 million users on over 400,000 software repositories. Our experiments show we were able to predict hourly user-repo activity with reasonably low error. Our simulations are accurate for 1–3 weeks (168–504 hours) after inception, with accuracy gradually falling off. It was shown that activity on Twitter and Reddit increases the accuracy of activity prediction on GitHub for most events.
Researching and Enacting Change in Postsecondary Education, 2018
Network Measurement and Data Collection 75 network of individuals when it is not feasible to coll... more Network Measurement and Data Collection 75 network of individuals when it is not feasible to collect complete network data because of the size of the population or other constraints. This type of design is similar to typical surveys in that the addition of the network component is no more burdensome to participants than the addition of questions to a survey without a network component. In whole network studies, it is important that every tie between any pair of actors be captured. In other words, what must be measured is the complete ensemble of ties that connect a bounded group of interest such as all full-time faculty members in a chemistry department or all instructors teaching primarily freshman large enrollment courses. In addition to the standard demographic attributes of participants that are typically measured in surveys (e.g., age, gender, race/ethnicity, educational attainment, etc.), the researcher must, for each person in the study population, identify to whom they are connected by a type (or types) of social tie(s). Commonly, this task is accomplished by directly asking respondents; furthermore, it is not unusual to then ask for additional information, such as the frequency of contact or the nature of interactions between themselves and each named alter. Therefore, the response burden increases quickly with the size of the network of interest, and researchers must be judicious in design of data collection instruments to avoid survey fatigue. This chapter focuses on whole network studies for two reasons. First, whole network studies are the ones most relevant to researchers focused on change in higher education, and particularly with regard to change in instructional practices. It is axiomatic in such research that the academic department is the key locus of change (Austin, 2011); researchers must therefore understand how colleagues, within a department or across departments, relate to one another through their networks of collaboration, advice, friendship, material sharing, and discussion in order to understand how change occurs and how it can be constrained, diverted, or successfully implemented. Second, typical academic departments are not large and so it is feasible to collect the who-to-whom data on social ties needed for a whole network study. While ego-network designs have utility, they should only be the resort if one's target population is very large (all adults in the US) and automated monitoring of communications is not possible. 1 We organize the discussion in the next section around four basic data collection design considerations: study purpose, population, actor attributes, and-the most complex consideration unique to social network studies-identifi cation of the ties to be measured and comprehension of the nature of those ties. We follow up our discussion of tie identifi cation with a section that examines options for collecting network data such as surveys, interviews or focus groups, observations, and electronic records. Instrumentation and question wording are explored, with specifi c examples of questions used in the literature to elicit reports of social ties to others in educational settings in particular and work organizations in
Social media environments often foster the formation of communities promoted by users’ tendencies... more Social media environments often foster the formation of communities promoted by users’ tendencies toward homophily. These tendencies of connecting with similar users are solidified by the social media companies’ algorithmic and business practices, leading to polarized networks, where communities of different interests rarely interact. This paper investigates via simulations the adoption of a new convention promoted by a persistent minority in a network polarized into two communities. We perform experiments on two real-world networks and various synthetic networks with controlled properties. We discover that the position of the persistent minority has a greater impact on spreading new conventions than its relative size. We also show that although diffusion becomes harder as network polarization increases, a persistent minority can increase its effectiveness in promoting new conventions by targeting low-influence users from the opposite community.
Research in Social Stratification and Mobility, 2023
Status allocation is a process by which complex labor market processes allocate individuals with ... more Status allocation is a process by which complex labor market processes allocate individuals with varying credentials and background characteristics to occupational positions that vary systematically in prestige, autonomy, and compensation. Previous literature examines two principles of allocation, the meritocratic principle and the lottery or chance principle, to assess departures of an observed distribution from pure meritocratic versus pure chance distributions (Krauze and Słomczyński, 1985). In this paper, we model an observed distribution as a compound of elementary distributions, meritocratic and chance, governed by a mixing parameter that expresses, in effect, the proportion of the actual allocation attributable to one principle versus another. We generalize the constant mixing model to a differential mixing model in which the mixing parameters are origin-specific. We explore the insights our framework enables with two comprehensive data sets on educational origins and occupational destinations, one covering the EU for over 30 countries from 1992 to 2018 and the other from IPUMS for both EU and other nations. We find that the differential model typically fits better than the constant model, that mixing coefficient for the top origin category is much larger than the mixing coefficient for lower categories, that the allocation of women is generally more meritocratic than the allocation of men, that there is fluctuation over time in the degree to which an observed allocation fits a meritocratic pattern, and that countries differ widely how much meritocracy versus lottery contributes to status allocation.
Expectation states theory suggests that status orders emerge in task groups via the activation of... more Expectation states theory suggests that status orders emerge in task groups via the activation of exogenous status differences and the development of endogenous behavior interchange patterns. The formal e-state model developed by Skvortez and Fararo postulates three specific mechanisms—status activation, interaction initiation, and bystander effect—that drive the emergence conceptualized as networks of complete asymmetric digraphs, also known as “tournaments.” The author assesses this model with data from three sources: 77 six-person groups in which external status differences are experimentally created to span all possible initial status combinations from one high and five low to five high and one low status actor, 31 six-person and 51 four-person groups in which status combinations leverage existing status differences in the population of subjects (gender and year in college). Overall consistency with model prediction is noted, but challenges with deep evaluation remain, and these are addressed in the final section of the chapter.
Limitations: Participants in our study are divided into two groups based on a criterion that is l... more Limitations: Participants in our study are divided into two groups based on a criterion that is likely to have been too weak to induce strong group identity. Measures that enhance the sense of group identity need to be introduced in future studies.
... Computer Simulation for Theory Development: An Evolving Component of Sociological Research Pr... more ... Computer Simulation for Theory Development: An Evolving Component of Sociological Research Programs 47 Barbara F. Meeker and Robert K. Leik ... and Shawn Donnelly Toward a Formal Theory of Equilibrium in Network Exchange Systems 303 Tadeusz Sozanski Strategy in ...
International Journal of STEM Education, Aug 4, 2022
Background: Change strategies may leverage interpersonal relationships and conversations to sprea... more Background: Change strategies may leverage interpersonal relationships and conversations to spread teaching innovations among science faculty. Knowledge sharing refers to the process by which individuals transfer information and thereby spread innovative ideas within an organization. We use knowledge sharing as a lens for identifying factors that encourage productive teaching-related conversations between individuals, characterizing the context and content of these discussions, and understanding how peer interactions may shape instructional practices. In this study, we interview 19 science faculty using innovative teaching practices about the teaching-focused conversations they have with different discussion partners. Results: This qualitative study describes characteristics of the relationship between discussion partners, what they discuss with respect to teaching, the amount of help-seeking that occurs, and the perceived impacts of these conversations on their teaching. We highlight the role of office location and course overlap in bringing faculty together and characterize the range of topics they discuss, such as course delivery and teaching strategies. We note the tendency of faculty to seek out partners with relevant expertise and describe how faculty perceive their discussion partners to influence their instructional practices and personal affect. Finally, we elaborate on how these themes vary depending on the relationship between discussion partners. Conclusions: The knowledge sharing framework provides a useful lens for investigating how various factors affect faculty conversations around teaching. Building on this framework, our results lead us to propose two hypotheses for how to promote sharing teaching knowledge among faculty, thereby identifying productive directions for further systematic inquiry. In particular, we propose that productive teaching conversations might be cultivated by
Studies in computational intelligence, Dec 5, 2018
Real network datasets provide significant benefits for understanding phenomena such as informatio... more Real network datasets provide significant benefits for understanding phenomena such as information diffusion or network evolution. Yet the privacy risks raised from sharing real graph datasets, even when stripped of user identity information, are significant. When nodes have associated attributes, the privacy risks increase. In this paper we quantitatively study the impact of binary node attributes on node privacy by employing machine-learning-based re-identification attacks and exploring the interplay between graph topology and attribute placement. Our experiments show that the population's diversity on the binary attribute consistently degrades anonymity.
Preface 1: Macrostructural Concepts 2: Formal Theory of Population Structure 3: Testing Theoretic... more Preface 1: Macrostructural Concepts 2: Formal Theory of Population Structure 3: Testing Theoretical Implications 4: Occupational Chances 5: Structural Context and Organizations 6: Social Exchange 7: Historical Developments Bibliography Author Index Subject Index
Preface and acknowledgments 1. Introduction and overview 2. Social exchange and power 3. Punishme... more Preface and acknowledgments 1. Introduction and overview 2. Social exchange and power 3. Punishment and coercion 4. An experimental setting for studying power in exchange relations 5. The early research: experimental tests and theoretical puzzles 6. The structural determination of power use 7. Dependence and risk: structural constraints on strategic power use 8. Injustice and risk: normative constraints on strategic power use 9. The effects of coercion: compliance or conflict? 10. A theory of coercion in social exchange 11. Conclusions and implications Appendix I. Definitions of basic concepts of social exchange Appendix II. The experimental instructions for the standardized setting Notes References.
We evaluate four theories that predict the distribution of power in exchange networks. All four t... more We evaluate four theories that predict the distribution of power in exchange networks. All four theories-core theory, equidependence theory, exchange-resistance theory, and expected value theory-assume actors rationally pursue self-interests. Three of the theories add social psychological assumptions that place the pursuit of self interest in an interactive context. Predictions of exchange earnings by the four theories are evaluated against data from eight experimental networks, including types of networks not previously studied. These networks vary conditions that affect the chances that a position can be excluded from exchange. We find that when the theories base predictions on a network position's structural potential for exclusion, exchange-resistance theory provides the bestfit, but when predictions are based on actual experiences of exclusion, expected value theory fits best. Our discussion focuses on the distinction between the a priori potential for exclusion versus experienced exclusion as factors in the genesis of power
Network exchange theory predicts relative profits from negotiations among actors in social exchan... more Network exchange theory predicts relative profits from negotiations among actors in social exchange networks (Markovsky et al. 1993; Markovsky, Willer & Patton 1988). Here we extend the theory to allow exact predictions, rather than merely ordinal, for actors' exchange profits. This is accomplished by integrating two important factors. First, a resistance model predicts bilateral negotiation outcomes within a given set of network constraints. It does so by weighing actors' interests in gaining the best possible exchanges against their desires to avoid the worst. Second, the resistance model predictions are modified by actors' profit expectations. In particular, we incorporate two factors that affect such expectations, both common features of ongoing exchange relations: the number of other actors to whom one is directly connected in the network, and the likelihood of one's completing exchanges with them. We derive hypotheses from the theory and test them in two very different experimental settings. We find that the theory's predictions are more accurate than those of previous versions of the theory and those offive alternative theories.
We extend network exchange theory (Markovsky, Willer, and Patton 1988) to accommodate a new class... more We extend network exchange theory (Markovsky, Willer, and Patton 1988) to accommodate a new class of power phenomena. Previous theory and research have shown that structural configurations in some networks promote or inhibit exchange opportunities, leading to robust power and resource differentials. The extension identifies a structural basis for subtler forms of differentiation. Using computer simulations and laboratory experiments, we show that the degree to which this "weak power" is manifested in resource accumulations is conditioned by local and global network patterns, and by the experience and strategies of actors in the network. Experimental tests corroborate the predicted weak power effects and the consequences of variations in actors' negotiating experiences. NTetwork exchange theory (NET) was developed to predict negotiated distributions of resources in a class of networks consisting of interrelated individual or corporate actors (Markovsky, Willer, and Patton 1988). These networks provide a fertile context for addressing issues of power and exchange, social dynamics, structural transformation, micro-macro connections, and other questions central to sociology. A recent issue of the journal Social Networks (Vol. 14 (3-4), Sept.-Dec. 1992) is devoted to locating power in exchange networks. We report on a newly discovered phenomenon in exchange networks, weak power. Informally, we define weak power as a condition that promotes significant advantages for some network positions, but severely restricts those advantages relative to their poten-* Direct all correspondence to
Accurately modeling information di usion within and across social media platforms has many practi... more Accurately modeling information di usion within and across social media platforms has many practical applications, such as estimating the size of the audience exposed to a particular narrative or testing intervention techniques for addressing misinformation. However, it turns out that real data reveal phenomena that pose significant challenges to modeling: events in the physical world a ect in varying ways conversations on di erent social media platforms; coordinated influence campaigns may swing discussions in unexpected directions; a platform's algorithms direct who sees which message, which a ects in opaque ways how information spreads. This article describes our research e orts in the SocialSim program of the Defense Advanced Research Projects Agency. As formulated by DARPA, the intent of the SocialSim research program was "to develop innovative technologies for high-fidelity computational simulation of online social behavior ... [focused] specifically on information spread and evolution." In this article we document lessons we learned over the + years of the recently concluded project. Our hope is that an accounting of our experience may prove useful to other researchers should they attempt a related project.
Forecasting social media activity can be of practical use in many scenarios, from understanding t... more Forecasting social media activity can be of practical use in many scenarios, from understanding trends, such as which topics are likely to engage more users in the coming week, to identifying unusual behavior, such as coordinated information operations or PumpNDump efforts. To evaluate a new approach to forecasting, it is important to have baselines against which to assess performance gains. We experimentally evaluate the performance of four baselines for forecasting activity in several social media datasets that record discussions related to three different geo-political contexts synchronously taking place on two different platforms, Twitter and YouTube. Experiments are done over hourly time periods. Our evaluation identifies the baselines which are most accurate for particular metrics and thus provide guidance for future work in social media modeling.
A frequently used measure of scientific (or research) collaboration is co-authorship in scholarly... more A frequently used measure of scientific (or research) collaboration is co-authorship in scholarly publications. Less frequently used are joint grant proposals and patents. Many scholars believe that the use of co-authorship as the sole measure of research collaboration is insufficient because collaboration between researchers might not result in co-authorship. Collaborations involve informal communication (i.e., conversational exchange) between researchers. Using self-reports from 100 tenured/tenure-track faculty in the College of Engineering at the University of South Florida, researchers' networks are constructed from their communication relationships and also from collaborations in three areas: joint publications, joint grant proposals, and joint patents. The data collection: (1) covers both researchers' in-progress and completed collaborative outputs, (2) yields a rating from the researchers on the importance of a relationship to them and (3) obtains multiple types of relationship ties between researchers allowing for the comparison of multiple networks. Exponential random graph models (ERGMs) results show that the more communication researchers have the more likely they produce collaborative outputs. Furthermore, we find that joint grant proposals tend to have mixed gender teams and that researchers of the same race are more likely to publish together, but those demographic attributes have no additional explanatory power regarding joint patents.
Examining whether and how instructional coordination occurs within introductory undergraduate STE... more Examining whether and how instructional coordination occurs within introductory undergraduate STEM courses.
Status allocation is a process by which complex labor market processes allocate individuals with ... more Status allocation is a process by which complex labor market processes allocate individuals with varying credentials and background characteristics to occupational positions that vary systematically in prestige, autonomy, and compensation. Previous literature examines two principles of allocation, the meritocratic principle and the lottery or chance principle, to assess departures of an observed distribution from pure meritocratic versus pure chance distributions (Krauze and Słomczyński, 1985). In this paper, we model an observed distribution as a compound of elementary distributions, meritocratic and chance, governed by a mixing parameter that expresses, in effect, the proportion of the actual allocation attributable to one principle versus another. We generalize the constant mixing model to a differential mixing model in which the mixing parameters are origin-specific. We explore the insights our framework enables with two comprehensive data sets on educational origins and occupational destinations, one covering the EU for over 30 countries from 1992 to 2018 and the other from IPUMS for both EU and other nations. We find that the differential model typically fits better than the constant model, that mixing coefficient for the top origin category is much larger than the mixing coefficient for lower categories, that the allocation of women is generally more meritocratic than the allocation of men, that there is fluctuation over time in the degree to which an observed allocation fits a meritocratic pattern, and that countries differ widely how much meritocracy versus lottery contributes to status allocation.
2019 IEEE International Conference on Systems, Man and Cybernetics (SMC), 2019
This paper introduces a decomposition approach to address the problem of predicting different use... more This paper introduces a decomposition approach to address the problem of predicting different user activities at hour granularity over a long period of time. Our approach involves two steps. First, we used a temporal neural network ensemble to predict the number of each type of activity that occurred in a day. Second, we used a set of neural networks to assign the events to a user-repository pair in a particular hour. We focused this work on a subset of the public GitHub dataset that records the activities of over 2 million users on over 400,000 software repositories. Our experiments show we were able to predict hourly user-repo activity with reasonably low error. Our simulations are accurate for 1–3 weeks (168–504 hours) after inception, with accuracy gradually falling off. It was shown that activity on Twitter and Reddit increases the accuracy of activity prediction on GitHub for most events.
Researching and Enacting Change in Postsecondary Education, 2018
Network Measurement and Data Collection 75 network of individuals when it is not feasible to coll... more Network Measurement and Data Collection 75 network of individuals when it is not feasible to collect complete network data because of the size of the population or other constraints. This type of design is similar to typical surveys in that the addition of the network component is no more burdensome to participants than the addition of questions to a survey without a network component. In whole network studies, it is important that every tie between any pair of actors be captured. In other words, what must be measured is the complete ensemble of ties that connect a bounded group of interest such as all full-time faculty members in a chemistry department or all instructors teaching primarily freshman large enrollment courses. In addition to the standard demographic attributes of participants that are typically measured in surveys (e.g., age, gender, race/ethnicity, educational attainment, etc.), the researcher must, for each person in the study population, identify to whom they are connected by a type (or types) of social tie(s). Commonly, this task is accomplished by directly asking respondents; furthermore, it is not unusual to then ask for additional information, such as the frequency of contact or the nature of interactions between themselves and each named alter. Therefore, the response burden increases quickly with the size of the network of interest, and researchers must be judicious in design of data collection instruments to avoid survey fatigue. This chapter focuses on whole network studies for two reasons. First, whole network studies are the ones most relevant to researchers focused on change in higher education, and particularly with regard to change in instructional practices. It is axiomatic in such research that the academic department is the key locus of change (Austin, 2011); researchers must therefore understand how colleagues, within a department or across departments, relate to one another through their networks of collaboration, advice, friendship, material sharing, and discussion in order to understand how change occurs and how it can be constrained, diverted, or successfully implemented. Second, typical academic departments are not large and so it is feasible to collect the who-to-whom data on social ties needed for a whole network study. While ego-network designs have utility, they should only be the resort if one's target population is very large (all adults in the US) and automated monitoring of communications is not possible. 1 We organize the discussion in the next section around four basic data collection design considerations: study purpose, population, actor attributes, and-the most complex consideration unique to social network studies-identifi cation of the ties to be measured and comprehension of the nature of those ties. We follow up our discussion of tie identifi cation with a section that examines options for collecting network data such as surveys, interviews or focus groups, observations, and electronic records. Instrumentation and question wording are explored, with specifi c examples of questions used in the literature to elicit reports of social ties to others in educational settings in particular and work organizations in
Social media environments often foster the formation of communities promoted by users’ tendencies... more Social media environments often foster the formation of communities promoted by users’ tendencies toward homophily. These tendencies of connecting with similar users are solidified by the social media companies’ algorithmic and business practices, leading to polarized networks, where communities of different interests rarely interact. This paper investigates via simulations the adoption of a new convention promoted by a persistent minority in a network polarized into two communities. We perform experiments on two real-world networks and various synthetic networks with controlled properties. We discover that the position of the persistent minority has a greater impact on spreading new conventions than its relative size. We also show that although diffusion becomes harder as network polarization increases, a persistent minority can increase its effectiveness in promoting new conventions by targeting low-influence users from the opposite community.
Research in Social Stratification and Mobility, 2023
Status allocation is a process by which complex labor market processes allocate individuals with ... more Status allocation is a process by which complex labor market processes allocate individuals with varying credentials and background characteristics to occupational positions that vary systematically in prestige, autonomy, and compensation. Previous literature examines two principles of allocation, the meritocratic principle and the lottery or chance principle, to assess departures of an observed distribution from pure meritocratic versus pure chance distributions (Krauze and Słomczyński, 1985). In this paper, we model an observed distribution as a compound of elementary distributions, meritocratic and chance, governed by a mixing parameter that expresses, in effect, the proportion of the actual allocation attributable to one principle versus another. We generalize the constant mixing model to a differential mixing model in which the mixing parameters are origin-specific. We explore the insights our framework enables with two comprehensive data sets on educational origins and occupational destinations, one covering the EU for over 30 countries from 1992 to 2018 and the other from IPUMS for both EU and other nations. We find that the differential model typically fits better than the constant model, that mixing coefficient for the top origin category is much larger than the mixing coefficient for lower categories, that the allocation of women is generally more meritocratic than the allocation of men, that there is fluctuation over time in the degree to which an observed allocation fits a meritocratic pattern, and that countries differ widely how much meritocracy versus lottery contributes to status allocation.
Expectation states theory suggests that status orders emerge in task groups via the activation of... more Expectation states theory suggests that status orders emerge in task groups via the activation of exogenous status differences and the development of endogenous behavior interchange patterns. The formal e-state model developed by Skvortez and Fararo postulates three specific mechanisms—status activation, interaction initiation, and bystander effect—that drive the emergence conceptualized as networks of complete asymmetric digraphs, also known as “tournaments.” The author assesses this model with data from three sources: 77 six-person groups in which external status differences are experimentally created to span all possible initial status combinations from one high and five low to five high and one low status actor, 31 six-person and 51 four-person groups in which status combinations leverage existing status differences in the population of subjects (gender and year in college). Overall consistency with model prediction is noted, but challenges with deep evaluation remain, and these are addressed in the final section of the chapter.
Limitations: Participants in our study are divided into two groups based on a criterion that is l... more Limitations: Participants in our study are divided into two groups based on a criterion that is likely to have been too weak to induce strong group identity. Measures that enhance the sense of group identity need to be introduced in future studies.
... Computer Simulation for Theory Development: An Evolving Component of Sociological Research Pr... more ... Computer Simulation for Theory Development: An Evolving Component of Sociological Research Programs 47 Barbara F. Meeker and Robert K. Leik ... and Shawn Donnelly Toward a Formal Theory of Equilibrium in Network Exchange Systems 303 Tadeusz Sozanski Strategy in ...
International Journal of STEM Education, Aug 4, 2022
Background: Change strategies may leverage interpersonal relationships and conversations to sprea... more Background: Change strategies may leverage interpersonal relationships and conversations to spread teaching innovations among science faculty. Knowledge sharing refers to the process by which individuals transfer information and thereby spread innovative ideas within an organization. We use knowledge sharing as a lens for identifying factors that encourage productive teaching-related conversations between individuals, characterizing the context and content of these discussions, and understanding how peer interactions may shape instructional practices. In this study, we interview 19 science faculty using innovative teaching practices about the teaching-focused conversations they have with different discussion partners. Results: This qualitative study describes characteristics of the relationship between discussion partners, what they discuss with respect to teaching, the amount of help-seeking that occurs, and the perceived impacts of these conversations on their teaching. We highlight the role of office location and course overlap in bringing faculty together and characterize the range of topics they discuss, such as course delivery and teaching strategies. We note the tendency of faculty to seek out partners with relevant expertise and describe how faculty perceive their discussion partners to influence their instructional practices and personal affect. Finally, we elaborate on how these themes vary depending on the relationship between discussion partners. Conclusions: The knowledge sharing framework provides a useful lens for investigating how various factors affect faculty conversations around teaching. Building on this framework, our results lead us to propose two hypotheses for how to promote sharing teaching knowledge among faculty, thereby identifying productive directions for further systematic inquiry. In particular, we propose that productive teaching conversations might be cultivated by
Studies in computational intelligence, Dec 5, 2018
Real network datasets provide significant benefits for understanding phenomena such as informatio... more Real network datasets provide significant benefits for understanding phenomena such as information diffusion or network evolution. Yet the privacy risks raised from sharing real graph datasets, even when stripped of user identity information, are significant. When nodes have associated attributes, the privacy risks increase. In this paper we quantitatively study the impact of binary node attributes on node privacy by employing machine-learning-based re-identification attacks and exploring the interplay between graph topology and attribute placement. Our experiments show that the population's diversity on the binary attribute consistently degrades anonymity.
Preface 1: Macrostructural Concepts 2: Formal Theory of Population Structure 3: Testing Theoretic... more Preface 1: Macrostructural Concepts 2: Formal Theory of Population Structure 3: Testing Theoretical Implications 4: Occupational Chances 5: Structural Context and Organizations 6: Social Exchange 7: Historical Developments Bibliography Author Index Subject Index
Preface and acknowledgments 1. Introduction and overview 2. Social exchange and power 3. Punishme... more Preface and acknowledgments 1. Introduction and overview 2. Social exchange and power 3. Punishment and coercion 4. An experimental setting for studying power in exchange relations 5. The early research: experimental tests and theoretical puzzles 6. The structural determination of power use 7. Dependence and risk: structural constraints on strategic power use 8. Injustice and risk: normative constraints on strategic power use 9. The effects of coercion: compliance or conflict? 10. A theory of coercion in social exchange 11. Conclusions and implications Appendix I. Definitions of basic concepts of social exchange Appendix II. The experimental instructions for the standardized setting Notes References.
We evaluate four theories that predict the distribution of power in exchange networks. All four t... more We evaluate four theories that predict the distribution of power in exchange networks. All four theories-core theory, equidependence theory, exchange-resistance theory, and expected value theory-assume actors rationally pursue self-interests. Three of the theories add social psychological assumptions that place the pursuit of self interest in an interactive context. Predictions of exchange earnings by the four theories are evaluated against data from eight experimental networks, including types of networks not previously studied. These networks vary conditions that affect the chances that a position can be excluded from exchange. We find that when the theories base predictions on a network position's structural potential for exclusion, exchange-resistance theory provides the bestfit, but when predictions are based on actual experiences of exclusion, expected value theory fits best. Our discussion focuses on the distinction between the a priori potential for exclusion versus experienced exclusion as factors in the genesis of power
Network exchange theory predicts relative profits from negotiations among actors in social exchan... more Network exchange theory predicts relative profits from negotiations among actors in social exchange networks (Markovsky et al. 1993; Markovsky, Willer & Patton 1988). Here we extend the theory to allow exact predictions, rather than merely ordinal, for actors' exchange profits. This is accomplished by integrating two important factors. First, a resistance model predicts bilateral negotiation outcomes within a given set of network constraints. It does so by weighing actors' interests in gaining the best possible exchanges against their desires to avoid the worst. Second, the resistance model predictions are modified by actors' profit expectations. In particular, we incorporate two factors that affect such expectations, both common features of ongoing exchange relations: the number of other actors to whom one is directly connected in the network, and the likelihood of one's completing exchanges with them. We derive hypotheses from the theory and test them in two very different experimental settings. We find that the theory's predictions are more accurate than those of previous versions of the theory and those offive alternative theories.
We extend network exchange theory (Markovsky, Willer, and Patton 1988) to accommodate a new class... more We extend network exchange theory (Markovsky, Willer, and Patton 1988) to accommodate a new class of power phenomena. Previous theory and research have shown that structural configurations in some networks promote or inhibit exchange opportunities, leading to robust power and resource differentials. The extension identifies a structural basis for subtler forms of differentiation. Using computer simulations and laboratory experiments, we show that the degree to which this "weak power" is manifested in resource accumulations is conditioned by local and global network patterns, and by the experience and strategies of actors in the network. Experimental tests corroborate the predicted weak power effects and the consequences of variations in actors' negotiating experiences. NTetwork exchange theory (NET) was developed to predict negotiated distributions of resources in a class of networks consisting of interrelated individual or corporate actors (Markovsky, Willer, and Patton 1988). These networks provide a fertile context for addressing issues of power and exchange, social dynamics, structural transformation, micro-macro connections, and other questions central to sociology. A recent issue of the journal Social Networks (Vol. 14 (3-4), Sept.-Dec. 1992) is devoted to locating power in exchange networks. We report on a newly discovered phenomenon in exchange networks, weak power. Informally, we define weak power as a condition that promotes significant advantages for some network positions, but severely restricts those advantages relative to their poten-* Direct all correspondence to
Accurately modeling information di usion within and across social media platforms has many practi... more Accurately modeling information di usion within and across social media platforms has many practical applications, such as estimating the size of the audience exposed to a particular narrative or testing intervention techniques for addressing misinformation. However, it turns out that real data reveal phenomena that pose significant challenges to modeling: events in the physical world a ect in varying ways conversations on di erent social media platforms; coordinated influence campaigns may swing discussions in unexpected directions; a platform's algorithms direct who sees which message, which a ects in opaque ways how information spreads. This article describes our research e orts in the SocialSim program of the Defense Advanced Research Projects Agency. As formulated by DARPA, the intent of the SocialSim research program was "to develop innovative technologies for high-fidelity computational simulation of online social behavior ... [focused] specifically on information spread and evolution." In this article we document lessons we learned over the + years of the recently concluded project. Our hope is that an accounting of our experience may prove useful to other researchers should they attempt a related project.
Forecasting social media activity can be of practical use in many scenarios, from understanding t... more Forecasting social media activity can be of practical use in many scenarios, from understanding trends, such as which topics are likely to engage more users in the coming week, to identifying unusual behavior, such as coordinated information operations or PumpNDump efforts. To evaluate a new approach to forecasting, it is important to have baselines against which to assess performance gains. We experimentally evaluate the performance of four baselines for forecasting activity in several social media datasets that record discussions related to three different geo-political contexts synchronously taking place on two different platforms, Twitter and YouTube. Experiments are done over hourly time periods. Our evaluation identifies the baselines which are most accurate for particular metrics and thus provide guidance for future work in social media modeling.
A frequently used measure of scientific (or research) collaboration is co-authorship in scholarly... more A frequently used measure of scientific (or research) collaboration is co-authorship in scholarly publications. Less frequently used are joint grant proposals and patents. Many scholars believe that the use of co-authorship as the sole measure of research collaboration is insufficient because collaboration between researchers might not result in co-authorship. Collaborations involve informal communication (i.e., conversational exchange) between researchers. Using self-reports from 100 tenured/tenure-track faculty in the College of Engineering at the University of South Florida, researchers' networks are constructed from their communication relationships and also from collaborations in three areas: joint publications, joint grant proposals, and joint patents. The data collection: (1) covers both researchers' in-progress and completed collaborative outputs, (2) yields a rating from the researchers on the importance of a relationship to them and (3) obtains multiple types of relationship ties between researchers allowing for the comparison of multiple networks. Exponential random graph models (ERGMs) results show that the more communication researchers have the more likely they produce collaborative outputs. Furthermore, we find that joint grant proposals tend to have mixed gender teams and that researchers of the same race are more likely to publish together, but those demographic attributes have no additional explanatory power regarding joint patents.
Students' social relationships and the knowledge and resources available through those connection... more Students' social relationships and the knowledge and resources available through those connections (Lin, 2002) Social capital Cultural models of engineering success (CMES) cultural beliefs about how students can succeed in their engineering programs (D'Andrade, 1995) Cultural Models
Uploads
Papers by John Skvoretz