Summary Computational Analysis in Digital Communication

Summary Computational Analysis of Digital Communication
November/December 2021



Literature

1. 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.

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.

3. Mellado, C., Hallin, D., Cárcamo, L., Alfaro, R., Jackson, D., Humanes, M. L., ... &
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, 1-25.

4. Van Atteveldt, W., van der Velden, M. A., & 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.

5. 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.




1

, 1. Van Atteveldt & Peng (2018): When communication meets
computation: Opportunities, challenges, and pitfalls in
computational communications science

The role of computational methods in communication science

Development for the use of computational methods for communication science:
1. Deluge of digitally available data, ranging from social media messages and other
‘digital traces’ to web archives and newly digitized newspaper and other historical
archives
2. Improved tools to analyze this data, including:
o Network analysis methods
o Automatic text analysis methods:
 Supervised text classification
 Topic modelling
 Word embeddings
 Syntactic methods
3. Emerge of powerful and cheap processing power, and easy to use computing
infrastructure for processing these data, including:
o Scientific and commercial cloud computing
o Sharing platforms = Github. Dataverse
o Crowd coding platforms = Amazon MTurk, Crowdflower

Three developments:
- Potential to give an unprecedented boost to progress in communication science
- Overcome technical, social, and ethical challenges presented by developments

Computational communication science studies involve:
1. Large and complex datasets
2. Consisting of digital traces and other ‘naturally occurring’ data
3. Requiring algorithmic solutions to analyze
4. Allowing the study of human communication by applying and testing communication
theory

Computational methods:
- Not replacement of methodological approaches = complementation
- Distinction between classical and computational methods = boundaries are fuzzy

Opportunities offered by computational methods

Computational methods allow us to change our discipline in four ways:
1. From self-report to real behavior
o Help overcome social desirability problems
o It does not rely on people’s imperfect estimate of their own desires and
intentions
o Help overcome problems of linking content data to survey data:
 Bias in media self-reports
 News consumers cherry-pick articles from multiple sites
2. From lab experiments to studies of the actual social environment

2

, o Emergence of social media:
 Facilitates design and implementation of experiment research
 Crowd surfing platforms lowers obstacles in research subject
recruitment
o Fear of losing reputation = companies are reluctant in sharing data
o Coordination of experiments on social media = time-consuming
o Ethical issue = how to address these concerns involved in online experiments
3. From small N to large N
o Possibility to study subtle relations in smaller subpopulations
o Measuring messages/behavior in real-time:
 More fine-grained time series can be constructed
 Alleviating problems of simultaneous correlation
 Stronger case for finding casual mechanisms
o Machine learning research:
 Model selection and model shrinkage
 Examples of models = penalized (lasso) regression and cross-
validation
 Advantages:
 More parsimonious models
 Alleviate problems of overfitting that can occur with large
datasets
o Modeling of network and group dynamics:
 Exponential Random Graph Modeling (ERGM)
 Relational Event Modeling (REM)
4. From solitary to collaborative research:
o Digital data and computational tools:
 Easier to share and reuse resources
 Very hard for one researcher to do all steps of computational research
alone
o Advantages:
 Be more rigorous in defining operationalizations
 Documenting data and analysis process
 Furthering transparency and reproducibility of research
o Digital methods = bring quantitative and qualitative research closer together

Challenges and pitfalls in computational methods

Challenges and pitfalls of new methods:
1. How do we keep research datasets accessible?
o Privileged access = only researchers with a lot of connections get access to the
data
o Importance of open and transparent datasets:
 Stimulation of sharing and publishing datasets
 AmCAT = helps alleviate copyright restrictions
 Working with funding agencies and data providers to make
standardized datasets available
2. Is ‘big’ data always good data?
o Limitations of bigger data:
 It is found, while survey data is made:

3