Computational Communication Research

Authors: Sorin A. Matei, Kerk F. Kee
Affiliations:
- Brian Lamb School of Communication, Purdue University, West Lafayette, Indiana
- College of Media & Communication, Texas Tech University, Lubbock, Texas
Grant Information: College of Liberal Arts, Purdue University, Grant/Award Number: 0001
Contact: Sorin A. Matei, 100 N. University Street, BRNG 1290P, West Lafayette, IN 47907. Email: smatei@purdue.edu
The article proposes a new framework for computational communication research, suggesting it should not solely rely on inductive methods but instead blend theoretical and data-driven approaches.
Emphasizes the need for communication theory to center on human needs and abilities, drawing from established theories in mass and interpersonal communication.
Practical implications of the proposed definition are noted, calling for specific processes within computational communication research, including:
- Processes: Communication production, behavior, effects.
- Domains: Collaboration, trust, data storytelling, journalism.
- Methods: Content and network analysis.
Utilizes the Johari window as a framework for discussing computational approaches.
Provides a typology of computational communication research through a blend of foundational texts and leading research summaries.
Concludes by addressing the strengths and potential opportunities within this evolving field.

Computational Science Definition: A broad field encompassing data collection and analysis to address discovery problems across various disciplines, especially social sciences.
Computational Social Science: Described as a new field of inquiry proposed by Lazer et al. (2009), stressing the innovative capacities of combining large data analysis to understand collective human behavior.
Critiques previous definitions of computational social science for failing to clarify its uniqueness compared to non-computational methods, especially concerning:
- Units of analysis
- Domains of inquiry
- Types of questions addressed
Proposed method for defining computational communication research, emphasizing the requisite balance between inductive and deductive approaches, and calling for a delineation of what “computational” means in this context.

Computational research should be distinguished from mere digital or big data work, emphasizing its unique data harnessing capabilities compared to traditional methods.
Discovery through computational means is posited to transcend existing theoretical boundaries and focus on empirical questions lacking prior theoretical frameworks.
Introduces the Johari window model to discuss different knowledge types that researchers navigate:
- Known-knowns: Facts we verify.
- Known-unknowns: Acknowledged gaps in knowledge.
- Unknown-knowns: Hindsight knowledge.
- Unknown-unknowns: Potential discoveries yet unrecognized.

Population Size Advantage: Complete datasets from large online platforms enable analysis of rich interaction and relationships in ways traditional sampling cannot achieve.
Emphasizes the importance of studying autocorrelation and the interconnectedness of data in communication science.
Notes the exponential growth in data connections necessitating advanced computation methods for analysis, raising questions about feasibility and cost in managing large networks.

Surveys various aspects of computational communication research, focusing on:
- Processes (production, behavior, effects)
- Domains (collaboration, trust, journalism)
- Methods (content and network analysis)
Highlights the need for theoretical frameworks to drive effective computational research.

An increasing body of foundational literature within communication establishes computational approaches, such as the formation of the Computational Methods Interest Group (2017) in the International Communication Association.
Several open questions are recommended by Cappella (2017), focusing on the structures of communication networks and the interplay of mass and interpersonal effects.

Need for a unified framework to balance inductive and deductive methodologies while addressing issues like generalizability and ethics in computational communication.
Sorts current literature into segments reflecting production, behavior, effects, collaboration, trust, and journalistic domains.

Studies significant contributions to understanding user behaviors in communication through computational methods.

Key Questions: How does computational research explain user contributions to social media?
Historical perspectives on social capital and its influence on engagement and production within social platforms.
Empirical literature remains somewhat limited in scope, needing to build more comprehensive theoretical models to substantiate findings.

Research on information cascades highlights the roles of influencers within social networks, signaling the complexities involved in understanding user interactions.

Highlights foundational studies of media effects; notably the ethical challenges stemming from manipulating user emotion on social media.
Calls for rigorous theoretical integration to avoid fallacies in data-driven conclusions regarding media effects.

Highlights distinct domains where computational communication offers rich insights: social organization, trust and credibility, and data journalism.

Examines how platforms such as Facebook and Wikipedia revolutionize collaboration and content generation.
Emphasizes the importance of understanding the emergent and adhocratic nature of these online communities.

Investigates the challenges posed by user-generated content regarding trust and credibility in digital environments.
Research on systems that assess and ensure quality and truthfulness in content.

Engages with the role of data-driven analysis in reshaping journalistic practices and storytelling methods.

Differentiates between basic natural language processing and more nuanced content analysis methods to understand deeper meanings within texts.

Focuses on how network structures can predict behaviors and outcomes within communication environments, emphasizing the technical demands of network analysis research.

Concludes with a call for more integrated approaches that blend computational methods with robust theoretical foundations.

Underlines the continual evolution of methods and inquiries in computational communication and underscores the importance of situating computational approaches in deeper theoretical contexts and real-world applications.
Conflict of Interest: None declared.

Comprehensive list of cited works providing foundational texts, methodological discussions, and significant research contributions to the field of computational communication.