Andrew Pilny
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Research Subjects
Andrew Pilny has a total of 78 co-authors. Here are the top co-authors by number of shared publications:
Marshall Scott Poole
4 shared publications- Loka Ashwood
University of Wisconsin-Madison
3 shared publications Amanda Slone
2 shared publications- C. Joseph Huber
Department of Communication, College of Communication & Information, University of Kentucky, Lexington, KY 40506, USA
2 shared publications - Iftekhar Ahmed
University of North Texas
2 shared publications Jeffrey D. Proulx
University of Illinois at Urbana–Champaign, USA
2 shared publications
Latest Publications
- Open Access
- 18 August 2026
- Published by Wiley in Journal of Agrarian Change
- Vol. 26 (4) , e70110
The Architecture of Consolidation: Nests of Corporate Control in the Agrochemical Network
Research on agricultural consolidation tends to equate corporate control with the dominance of a few large firms. We take an alternative approach that examines control through the explosive growth of legally distinct firms that fragment accountability. Using a new global dataset of 6982 entities linked to BASF, Bayer, Corteva, Dow, DuPont, FMC, COFCO, ChemChina and UPL, we use social network analysis to map the ownership and leadership networks behind agrochemical conglomerates. Employing community detection, we find that firms bind together in five highly clustered nests that are bridged by intermediaries like tax consultancies and directors. Firms, not only people, occupy leadership roles and a handful carries each nest's global reach. These structures expose the limits of market power measures like HHI and CR4 that are attenuated to parent firms and miss network level coordination. We offer nests as a relational way to see and potentially contest corporate control.
- 18 August 2026
- Published by Wiley in Journal of Agrarian Change
- Vol. 26 (4) , e70110
Open AccessThe Architecture of Consolidation: Nests of Corporate Control in the Agrochemical Network
Research on agricultural consolidation tends to equate corporate control with the dominance of a few large firms. We take an alternative approach that examines control through the explosive growth of legally distinct firms that fragment accountability. Using a new global dataset of 6982 entities linked to BASF, Bayer, Corteva, Dow, DuPont, FMC, COFCO, ChemChina and UPL, we use social network analysis to map the ownership and leadership networks behind agrochemical conglomerates. Employing community detection, we find that firms bind together in five highly clustered nests that are bridged by intermediaries like tax consultancies and directors. Firms, not only people, occupy leadership roles and a handful carries each nest's global reach. These structures expose the limits of market power measures like HHI and CR4 that are attenuated to parent firms and miss network level coordination. We offer nests as a relational way to see and potentially contest corporate control.
- 14 July 2026
- Published by Edward Elgar Publishing
- p. 19-37
Using large language models to code group interaction: some observations from early stages of the game
Collecting and processing interaction data is a labor-intensive but necessary endeavor if one is to understand the dynamics of group and dyadic communication. Generative artificial intelligence (AI), especially large language models (LLMs), has the potential to both streamline data processing and illuminate new research approaches/areas, but it is not without challenges and pitfalls. In this chapter, we describe our experiences using LLMs to code group and dyadic interaction. In short, we are largely impressed with how the LLM performs when coding data but sometimes puzzled by the coding choices the machine makes. We present a relatively nontechnical description of large language models and then describe how we have applied LLMs to various types of group and interaction data.
- 14 July 2026
- Published by Edward Elgar Publishing
- p. 19-37
Using large language models to code group interaction: some observations from early stages of the game
Collecting and processing interaction data is a labor-intensive but necessary endeavor if one is to understand the dynamics of group and dyadic communication. Generative artificial intelligence (AI), especially large language models (LLMs), has the potential to both streamline data processing and illuminate new research approaches/areas, but it is not without challenges and pitfalls. In this chapter, we describe our experiences using LLMs to code group and dyadic interaction. In short, we are largely impressed with how the LLM performs when coding data but sometimes puzzled by the coding choices the machine makes. We present a relatively nontechnical description of large language models and then describe how we have applied LLMs to various types of group and interaction data.
- 22 January 2025
- Published by SAGE Publications in Small Group Research
- Vol. 56 (5) , 864-893
Coding Small Group Communication with AI: RNNs and Transformers with Context
This study compares the performance of recurrent neural networks (RNNs) and transformer-based models (DistilBERT) in classifying utterances as dialogue acts. The results show that transformers consistently outperform RNNs, highlighting their usefulness in coding small group interaction. Furthermore, the study explores the impact of incorporating context, in the form of preceding and following utterances. The findings reveal that adding context leads to modest improvements in model performance. Moreover, in some cases, adding context can lead to a slight decrease in performance. The study discusses the implications of these findings for small group researchers employing AI models for text classification tasks.
- 22 January 2025
- Published by SAGE Publications in Small Group Research
- Vol. 56 (5) , 864-893
Coding Small Group Communication with AI: RNNs and Transformers with Context
This study compares the performance of recurrent neural networks (RNNs) and transformer-based models (DistilBERT) in classifying utterances as dialogue acts. The results show that transformers consistently outperform RNNs, highlighting their usefulness in coding small group interaction. Furthermore, the study explores the impact of incorporating context, in the form of preceding and following utterances. The findings reveal that adding context leads to modest improvements in model performance. Moreover, in some cases, adding context can lead to a slight decrease in performance. The study discusses the implications of these findings for small group researchers employing AI models for text classification tasks.
- Open Access
- 1 August 2024
- Published by MDPI AG in Education Sciences
- Vol. 14 (8) , 837
Measuring and Comparing High School Teachers’ and Undergraduate Students’ Knowledge of Complex Systems
Complex systems (CSs) have garnered increasing attention in science education due to their prevalence in the natural world and their importance in addressing pressing issues such as climate change, pandemics, and biodiversity loss. However, the instruments for assessing one’s CS knowledge are limited, and the knowledge of CSs among in-service teachers remains underreported. Guided by the complexity framework, we launched a study to develop and validate a survey instrument for measuring the knowledge of CSs among high school teachers and undergraduate students, as well as delineating the contours of their knowledge. In this article, we present the development and validation of a Complex Systems Knowledge Survey (CSKS), and we use the survey to compare the CS knowledge among 252 high school teachers and 418 undergraduate students in the United States. Our key findings include that (1) both high school teachers and undergraduates exhibit relatively low knowledge of decentralization and stochasticity, (2) undergraduates, especially those in non-STEM majors, demonstrate moderate to low knowledge of emergence, (3) few teachers and undergraduates differentiate between complicated and complex systems, and (4) teachers and undergraduates recognize CS examples across natural and social systems. The implications of our findings are discussed.
- 1 August 2024
- Published by MDPI AG in Education Sciences
- Vol. 14 (8) , 837
Open AccessMeasuring and Comparing High School Teachers’ and Undergraduate Students’ Knowledge of Complex Systems
Complex systems (CSs) have garnered increasing attention in science education due to their prevalence in the natural world and their importance in addressing pressing issues such as climate change, pandemics, and biodiversity loss. However, the instruments for assessing one’s CS knowledge are limited, and the knowledge of CSs among in-service teachers remains underreported. Guided by the complexity framework, we launched a study to develop and validate a survey instrument for measuring the knowledge of CSs among high school teachers and undergraduate students, as well as delineating the contours of their knowledge. In this article, we present the development and validation of a Complex Systems Knowledge Survey (CSKS), and we use the survey to compare the CS knowledge among 252 high school teachers and 418 undergraduate students in the United States. Our key findings include that (1) both high school teachers and undergraduates exhibit relatively low knowledge of decentralization and stochasticity, (2) undergraduates, especially those in non-STEM majors, demonstrate moderate to low knowledge of emergence, (3) few teachers and undergraduates differentiate between complicated and complex systems, and (4) teachers and undergraduates recognize CS examples across natural and social systems. The implications of our findings are discussed.
- 12 March 2024
- Published by Taylor & Francis in Communication Research Reports
- Vol. 41 (2) , 61-70
From manual to machine: assessing the efficacy of large language models in content analysis
- 12 March 2024
- Published by Taylor & Francis in Communication Research Reports
- Vol. 41 (2) , 61-70
From manual to machine: assessing the efficacy of large language models in content analysis
- 3 November 2023
- Published by Taylor & Francis
- p. 94-106
Groups as Systems
This chapter introduces the groups as systems perspective. While conventional thinking views groups as a set of independent actors that need optimizing, systems thinking views groups as a set of dynamic and interdependent actors that interact to create emergent outcomes. The first part of the chapter contrasts systems thinking and conventional thinking across dimensions of causality, accountability, outcomes, and time. The second part introduces how to recognize and tell system stories, using the television show The Wire as an example. Finally, the third part provides practical recommendations for life inside groups, maintenance (e.g., soliciting feedback) and adapting (e.g., collective sensemaking).
- 3 November 2023
- Published by Taylor & Francis
- p. 94-106
Groups as Systems
This chapter introduces the groups as systems perspective. While conventional thinking views groups as a set of independent actors that need optimizing, systems thinking views groups as a set of dynamic and interdependent actors that interact to create emergent outcomes. The first part of the chapter contrasts systems thinking and conventional thinking across dimensions of causality, accountability, outcomes, and time. The second part introduces how to recognize and tell system stories, using the television show The Wire as an example. Finally, the third part provides practical recommendations for life inside groups, maintenance (e.g., soliciting feedback) and adapting (e.g., collective sensemaking).
- 9 August 2022
- Published by Springer Nature in Agriculture and Human Values
- Vol. 39 (4) , 1435
Correction: From Big Ag to Big Finance: a market network approach to power in agriculture
- 9 August 2022
- Published by Springer Nature in Agriculture and Human Values
- Vol. 39 (4) , 1435
Correction: From Big Ag to Big Finance: a market network approach to power in agriculture
- 12 July 2022
- Published by Springer Nature in Agriculture and Human Values
- Vol. 39 (4) , 1421-1434
From Big Ag to Big Finance: a market network approach to power in agriculture
- 12 July 2022
- Published by Springer Nature in Agriculture and Human Values
- Vol. 39 (4) , 1421-1434
From Big Ag to Big Finance: a market network approach to power in agriculture
- 28 May 2021
- Published by SAGE Publications in Communication Research
- Vol. 49 (5) , 703-732
What Influences Relationship Formation in a Global Civil Society Network? An Examination of Valued Multiplex Relations
Recent interorganizational communication research has taken up the question: why are networks structured the way they are? This line of inquiry has advanced communication network research by helping explain how and why networks take on certain structures or why certain organizations become positioned advantageously (or not). Yet, those studies assume relationships among organizations are either present or absent. This study considers how the strength of ties and multiplex relationships among organizations may reveal a more complex explanation for why networks take on certain structures. Our results challenge some long held assumptions about the mechanisms that influence network formation. For instance, our results offer important insights into the consequences of closure mechanisms, the applicability of preferential attachment to real-world networks, and the nuances of homophily in network formation on multidimensional relationships in a communication network. Implications for interorganizational networks are discussed.
- 28 May 2021
- Published by SAGE Publications in Communication Research
- Vol. 49 (5) , 703-732
What Influences Relationship Formation in a Global Civil Society Network? An Examination of Valued Multiplex Relations
Recent interorganizational communication research has taken up the question: why are networks structured the way they are? This line of inquiry has advanced communication network research by helping explain how and why networks take on certain structures or why certain organizations become positioned advantageously (or not). Yet, those studies assume relationships among organizations are either present or absent. This study considers how the strength of ties and multiplex relationships among organizations may reveal a more complex explanation for why networks take on certain structures. Our results challenge some long held assumptions about the mechanisms that influence network formation. For instance, our results offer important insights into the consequences of closure mechanisms, the applicability of preferential attachment to real-world networks, and the nuances of homophily in network formation on multidimensional relationships in a communication network. Implications for interorganizational networks are discussed.