Summary Computational Analysis of Digital Communication
***
Table of contents
| Lecture 1: Introduction to computational methods in Communication Science ................................................................... 2
| Article 1: Kramer, A. D. I., Guillory, J. E, & Hancock, J. (2014). Experimental Evidence of Massive-Scale Emotional Contagion
Through Social Networks. Proceedings of the National Academy of Sciences, 111(24), 8788-8790. ...........................................11
| Article 2: Van Atteveldt, W., & Peng, T.-Q. (2018). When communication meets computation: Opportunities, challenges, and
pitfalls in computational communication science. Communication Methods and Measures, 12(2-3), 81–92.
https://doi.org/10.1080/19312458.2018.1458084 ...................................................................................................................12
| Lecture 2: Automated Text Analysis and Dictionary Approaches........................................................................................ 14
| Article 1: Welbers, K., Van Atteveldt, W., & Benoit, K. (2017). Text analysis in R. Communication Methods and Measures, 11(4),
245-265. ................................................................................................................................................................................32
| Article 2: Heidenreich, T., Eberl, J.-M., Lind, F. & Boomgaarden, H. (2020). Political migration discourses on social media: a
comparative perspective on visibility and sentiment across political Facebook accounts in Europe. Journal of Ethnic and
Migration Studies, (46)7, 1261-1280, https://doi.org/10.1080/1369183X.2019.1665990 ............................................................34
| Article 3: Mellado, C., Hallin, D., Cárcamo, L. Alfaro, R. … & Ramos, A. (2021) Sourcing Pandemic News: A Cross-National
Computational Analysis of Mainstream Media Coverage of COVID-19 on Facebook, Twitter, and Instagram. Digital Journalism,
9(9), 1271-1295, https://doi.org/10.1080/21670811.2021.1942114 ..........................................................................................35
| Lecture 3: Text Classification using machine learning ....................................................................................................... 37
| Article 1: van Atteveldt, W., van der Velden, M. A. C. G., & Boukes, M.. (2021). The Validity of Sentiment Analysis: Comparing
Manual Annotation, Crowd-Coding, Dictionary Approaches, and Machine Learning Algorithms. Communication Methods and
Measures, (15)2, 121-140, https://doi.org/10.1080/19312458.2020.1869198 ...........................................................................57
| Article 2: Su, L. Y.-F., Xenos, M. A., Rose, K. M., Wirz, C., Scheufele, D. A., & Brossard, D. (2018). Uncivil and personal?
Comparing patterns of incivility in comments on the Facebook pages of news outlets. New Media & Society, 20(10), 3678–3699.
https://doi.org/10.1177/1461444818757205 ...........................................................................................................................59
| Lecture 4: Transformers and Large Language Models ........................................................................................................ 60
| Article 1: Balluff, P., Eberl, J-M., Oberhänsli, S. J., Bernhard-Harrer, J., & Huber, M. (2024). The Austrian Political Advertisement
Scandal: Patterns of “Journalism for Sale”. International Journal of Press/Politics. https://doi.org/10.1177/19401612241285672
.............................................................................................................................................................................................74
, Cycle 1: Introduction to computational methods and automated text analysis
***
| Lecture 1 : Introduction to computational methods in Communi cation Science
Much of what we know about human behavior is based on what people tell us:
• in self-report measures in surveys
• in responses in experimental research
• in qualitative interviews
Note: Although valuable, such information can be biased. A lot of (mass)communication looks like
(online)news or is based on user-generated content. We need to understand how news/ information
changes how people think.
What is computational Social Science?
In 2009, researchers studied wealth and poverty in Rwanda. They conducted a traditional survey with
1,000 mobile phone customers to collect demographic, social, and economic data. Additionally,
they analyzed call records from 1.5 million people. Using the survey data, they trained a machine
learning model to predict individuals’ wealth based on their phone usage patterns and estimated
their locations using geographic data from the call records.
Computational social science
A field of social science that uses algorithms and big data to study human and social behavior. It
adds to traditional research methods rather than replacing them. The goal is to use digital tools to
collect and analyze data in new ways.
Common methods include:
• Data mining: collecting large amounts of data from online sources
• Software tools: creating programs for social science experiments
• Text analysis: studying written content, such as sentiment or keywords
• Image analysis: recognizing faces or identifying visual themes
• Machine learning: making predictions or classifying information
• Agent-based modeling: simulating how people behave or how information spreads
Why is this important now?
• There is an enormous amount of digital data available — from social media posts and online
activity to web archives and digitized historical records.
• Big data about people and businesses is being created constantly.
• Computing power has become cheaper and easier to use, making large-scale data
processing accessible.
• New analytical tools are available, such as network analysis, automatic text analysis, topic
modeling, word embeddings, and large language models (LLMs).
2
,Ten characteristics of big data
# Characteristic Description
The scale or volume of some current data sets is often impressive. However, big data
sets are not an end in themselves, but they can enable certain kinds of research
including the study of rare events, the estimation of heterogeneity, and the detection of
1 Big small differences
Many big data systems are constantly collecting data and thus enable to study
2 Always-on unexpected events and allow for real-time measurement
Participants are generally not aware that their data are being captured or they have
3 Nonreactive become so accustomed to this data collection that it no longer changes their behavior.
Most big data sources are incomplete, in the sense that they don’t have the
information that you will want for your research. This is a common feature of data that
4 Incomplete were created for purposes other than research.
5 Inaccessible Data held by companies and governments are difficult for researchers to access.
Most big datasets are nonetheless not representative of certain populations. Out-of-
6 Nonrepresentative sample generalizations are hence difficult or impossible.
Many big data systems are changing constantly, thus making it difficult to study long-
7 Drifting term trends
Algorithmically Behavior in big data systems is not natural; it is driven by the engineering goals of the
8 confounded systems.
9 Dirty Big data often includes a lot of noise (e.g., junk, spam, spurious data points…)
10 Sensitive Some of the information that companies and governments have is sensitive.
Pro’s and Con’s of Computational Methods
Opportunities Pitfalls
• We can study actual behavior instead • Techniques often (rather)
of simply self-reports complicated
• We can study human beings in their • Data is often proprietary (not shared
social context instead of in an openly)
artificial lab setting • Samples are often biased
• We can increase our N (higher power) • Often, data have only insufficient
• Potential to uncover patterns and metadata
insights that we couldn’t investigate • Risks of no longer understanding the
before models we use (black box)
Definition: “Computational Communication Science (CCS) is the label applied to the emerging
subfield that investigates the use of computational algorithms to gather and analyze big and often
semi- or unstructured data sets to develop and test communication science theories”
Typical research areas
Computational communication science studies thus usually involve:
1. large and complex data set
2. consisting of digital traces and other “naturally occurring” data
3. requiring algorithmic solutions to analyze (e.g., machine learning, LLMs)
4. allowing the study of human communication by applying and testing communication theory
3
, Examples of research areas
• Political Communication
o Democratization and Polarization
o Hate Speech
• Social Media Use
o Tracking of actual social media use
o Spreading of behavior, information, or emotions
• Health Communication
o Prevalence of health information online
• (Online) Journalism
o News coverage across decades
o Gender equality
Example 1: Analyzing news coverage – Jacobi et al. (2016)
• Studied New York Times coverage of nuclear technology from 1945–
2014.
• Analyzed 51,528 news articles (headlines and leads) — too many for
manual coding.
• Used LDA topic modeling in R to identify hidden topics and track how
they changed over time.
• The results showed a shift in the types of nuclear-related news
published across the decades.
Example 2: Facebook data to predict personality – Kosinski et al. (2013)
• Used data from over 58,000 Facebook users who shared their Likes, demographic details,
and results from personality tests.
• Showed that it’s possible to predict personal characteristics and personality traits based on
Facebook Likes.
• The results demonstrated that even simple online behaviors (like what people “like”) can
reveal things such as gender and sexual orientation.
4
***
Table of contents
| Lecture 1: Introduction to computational methods in Communication Science ................................................................... 2
| Article 1: Kramer, A. D. I., Guillory, J. E, & Hancock, J. (2014). Experimental Evidence of Massive-Scale Emotional Contagion
Through Social Networks. Proceedings of the National Academy of Sciences, 111(24), 8788-8790. ...........................................11
| Article 2: Van Atteveldt, W., & Peng, T.-Q. (2018). When communication meets computation: Opportunities, challenges, and
pitfalls in computational communication science. Communication Methods and Measures, 12(2-3), 81–92.
https://doi.org/10.1080/19312458.2018.1458084 ...................................................................................................................12
| Lecture 2: Automated Text Analysis and Dictionary Approaches........................................................................................ 14
| Article 1: Welbers, K., Van Atteveldt, W., & Benoit, K. (2017). Text analysis in R. Communication Methods and Measures, 11(4),
245-265. ................................................................................................................................................................................32
| Article 2: Heidenreich, T., Eberl, J.-M., Lind, F. & Boomgaarden, H. (2020). Political migration discourses on social media: a
comparative perspective on visibility and sentiment across political Facebook accounts in Europe. Journal of Ethnic and
Migration Studies, (46)7, 1261-1280, https://doi.org/10.1080/1369183X.2019.1665990 ............................................................34
| Article 3: Mellado, C., Hallin, D., Cárcamo, L. Alfaro, R. … & Ramos, A. (2021) Sourcing Pandemic News: A Cross-National
Computational Analysis of Mainstream Media Coverage of COVID-19 on Facebook, Twitter, and Instagram. Digital Journalism,
9(9), 1271-1295, https://doi.org/10.1080/21670811.2021.1942114 ..........................................................................................35
| Lecture 3: Text Classification using machine learning ....................................................................................................... 37
| Article 1: van Atteveldt, W., van der Velden, M. A. C. G., & Boukes, M.. (2021). The Validity of Sentiment Analysis: Comparing
Manual Annotation, Crowd-Coding, Dictionary Approaches, and Machine Learning Algorithms. Communication Methods and
Measures, (15)2, 121-140, https://doi.org/10.1080/19312458.2020.1869198 ...........................................................................57
| Article 2: Su, L. Y.-F., Xenos, M. A., Rose, K. M., Wirz, C., Scheufele, D. A., & Brossard, D. (2018). Uncivil and personal?
Comparing patterns of incivility in comments on the Facebook pages of news outlets. New Media & Society, 20(10), 3678–3699.
https://doi.org/10.1177/1461444818757205 ...........................................................................................................................59
| Lecture 4: Transformers and Large Language Models ........................................................................................................ 60
| Article 1: Balluff, P., Eberl, J-M., Oberhänsli, S. J., Bernhard-Harrer, J., & Huber, M. (2024). The Austrian Political Advertisement
Scandal: Patterns of “Journalism for Sale”. International Journal of Press/Politics. https://doi.org/10.1177/19401612241285672
.............................................................................................................................................................................................74
, Cycle 1: Introduction to computational methods and automated text analysis
***
| Lecture 1 : Introduction to computational methods in Communi cation Science
Much of what we know about human behavior is based on what people tell us:
• in self-report measures in surveys
• in responses in experimental research
• in qualitative interviews
Note: Although valuable, such information can be biased. A lot of (mass)communication looks like
(online)news or is based on user-generated content. We need to understand how news/ information
changes how people think.
What is computational Social Science?
In 2009, researchers studied wealth and poverty in Rwanda. They conducted a traditional survey with
1,000 mobile phone customers to collect demographic, social, and economic data. Additionally,
they analyzed call records from 1.5 million people. Using the survey data, they trained a machine
learning model to predict individuals’ wealth based on their phone usage patterns and estimated
their locations using geographic data from the call records.
Computational social science
A field of social science that uses algorithms and big data to study human and social behavior. It
adds to traditional research methods rather than replacing them. The goal is to use digital tools to
collect and analyze data in new ways.
Common methods include:
• Data mining: collecting large amounts of data from online sources
• Software tools: creating programs for social science experiments
• Text analysis: studying written content, such as sentiment or keywords
• Image analysis: recognizing faces or identifying visual themes
• Machine learning: making predictions or classifying information
• Agent-based modeling: simulating how people behave or how information spreads
Why is this important now?
• There is an enormous amount of digital data available — from social media posts and online
activity to web archives and digitized historical records.
• Big data about people and businesses is being created constantly.
• Computing power has become cheaper and easier to use, making large-scale data
processing accessible.
• New analytical tools are available, such as network analysis, automatic text analysis, topic
modeling, word embeddings, and large language models (LLMs).
2
,Ten characteristics of big data
# Characteristic Description
The scale or volume of some current data sets is often impressive. However, big data
sets are not an end in themselves, but they can enable certain kinds of research
including the study of rare events, the estimation of heterogeneity, and the detection of
1 Big small differences
Many big data systems are constantly collecting data and thus enable to study
2 Always-on unexpected events and allow for real-time measurement
Participants are generally not aware that their data are being captured or they have
3 Nonreactive become so accustomed to this data collection that it no longer changes their behavior.
Most big data sources are incomplete, in the sense that they don’t have the
information that you will want for your research. This is a common feature of data that
4 Incomplete were created for purposes other than research.
5 Inaccessible Data held by companies and governments are difficult for researchers to access.
Most big datasets are nonetheless not representative of certain populations. Out-of-
6 Nonrepresentative sample generalizations are hence difficult or impossible.
Many big data systems are changing constantly, thus making it difficult to study long-
7 Drifting term trends
Algorithmically Behavior in big data systems is not natural; it is driven by the engineering goals of the
8 confounded systems.
9 Dirty Big data often includes a lot of noise (e.g., junk, spam, spurious data points…)
10 Sensitive Some of the information that companies and governments have is sensitive.
Pro’s and Con’s of Computational Methods
Opportunities Pitfalls
• We can study actual behavior instead • Techniques often (rather)
of simply self-reports complicated
• We can study human beings in their • Data is often proprietary (not shared
social context instead of in an openly)
artificial lab setting • Samples are often biased
• We can increase our N (higher power) • Often, data have only insufficient
• Potential to uncover patterns and metadata
insights that we couldn’t investigate • Risks of no longer understanding the
before models we use (black box)
Definition: “Computational Communication Science (CCS) is the label applied to the emerging
subfield that investigates the use of computational algorithms to gather and analyze big and often
semi- or unstructured data sets to develop and test communication science theories”
Typical research areas
Computational communication science studies thus usually involve:
1. large and complex data set
2. consisting of digital traces and other “naturally occurring” data
3. requiring algorithmic solutions to analyze (e.g., machine learning, LLMs)
4. allowing the study of human communication by applying and testing communication theory
3
, Examples of research areas
• Political Communication
o Democratization and Polarization
o Hate Speech
• Social Media Use
o Tracking of actual social media use
o Spreading of behavior, information, or emotions
• Health Communication
o Prevalence of health information online
• (Online) Journalism
o News coverage across decades
o Gender equality
Example 1: Analyzing news coverage – Jacobi et al. (2016)
• Studied New York Times coverage of nuclear technology from 1945–
2014.
• Analyzed 51,528 news articles (headlines and leads) — too many for
manual coding.
• Used LDA topic modeling in R to identify hidden topics and track how
they changed over time.
• The results showed a shift in the types of nuclear-related news
published across the decades.
Example 2: Facebook data to predict personality – Kosinski et al. (2013)
• Used data from over 58,000 Facebook users who shared their Likes, demographic details,
and results from personality tests.
• Showed that it’s possible to predict personal characteristics and personality traits based on
Facebook Likes.
• The results demonstrated that even simple online behaviors (like what people “like”) can
reveal things such as gender and sexual orientation.
4
