Inhoudsopgave
Week 1...................................................................................................................................................................... 2
Literature ............................................................................................................................................................. 2
When communication meets computation: opportunities, challenges, and pitfalls in Computational
communication science. Wouter van Atteveldt & Tai-Quan Peng (2018........................................................ 2
Experimental evidence of massive-scale emotional contagion through social networks (Kramer et al.,
2014) ............................................................................................................................................................... 4
R steps and important functions........................................................................................................................... 5
Week 2...................................................................................................................................................................... 5
Literature ............................................................................................................................................................. 5
Text analysis in R (2017) ................................................................................................................................ 5
Political migration discourses on social media: comparative perspective on visibility and sentiment across
political Facebook accounts in Europe ........................................................................................................... 8
Sourcing pandemic news: A cross-national computational analysis of mainstream media coverage of
COVID-19 on Facebook, Twitter & Instagram............................................................................................. 10
SUMMARY SLIDES LIT............................................................................................................................. 13
R steps and important functions......................................................................................................................... 13
Week 3.................................................................................................................................................................... 14
Literature ........................................................................................................................................................... 14
The validity of sentiment analysis: comparing manual annotation, crowd-coding, dictionary approaches,
and machine learning algorithms (van Atteveldt et al., 2021) ...................................................................... 14
Uncivil and personal? Comparing patterns of incivility in comments on the Facebook pages of news
outlets ............................................................................................................................................................ 16
R steps and important functions......................................................................................................................... 20
Extra information ............................................................................................................................................... 21
Naive Bayes:............................................................................................................................................. 21
o Processing Information: Naive Bayes is a probabilistic model based on Bayes' theorem. It assumes that
the features used for classification are conditionally independent given the class label. Despite its "naive"
assumption, Naive Bayes often performs wel in practice for text classification tasks. ..................................... 21
o How It Works:........................................................................................................................................... 21
Week 4.................................................................................................................................................................... 22
Literature ........................................................................................................................................................... 22
Advancing Automated Content Analysis for a New Era of Media Effects Research: The Key Role of
Transfer Learning (Kroon et al., 2023) ......................................................................................................... 22
Limitations of traditional content analysis techniques.................................................................................... 23
R steps and important functions......................................................................................................................... 27
Extra information ............................................................................................................................................... 28
1
,Week 1
Literature
When communication meets computation: opportunities, challenges, and pitfalls in Computational
communication science. Wouter van Atteveldt & Tai-Quan Peng (2018
The role of computational methods in communication science
The recent acceleration in the promise and use of computational methods for communication science
is primarily fueled by the confluence of at least three developments:
• A deluge of digitally available data
• Improved tools to analyze data
• The emergence of powerful and cheap processing power and easy to use computing
infrastructure for processing these data
• Many of these new data sets contain communication artifacts such as tweets etc.
• Many communication theories are cited as major backbones in many computational studies
• Computational communication science studies generally involve:
o Large and complex data sets
o Consisting of digital traces and other naturally occurring data
o Requiring algorithmic solutions to analyze
o Allowing the study of human communication by applying and testing communication
theory
• Computational methods are an expansion and enhancement of the existing methodological
toolbox
Opportunities offered by computational methods
• From self-report to real behavior
o These data allow us to measure actual behavior in an unobtrusive way rather than self-
reported attitudes or intentions
o This can help overcome social desirability problems, and more importantly it does not
rely on peoples imperfect estimate of their own desires and intentions
o Can also help overcome the problems of linking content data to survey data. We can
now trace news consumption in real-time and combine it with survey data to get a
more sophisticated measurement of news consumption and effects.
• From lab experiments to studies of the actual social environment
o Actual environment
o It's no longer difficult to recruit thousands or even millions of diverse or specialized
subjects from crowdsourcing platforms
o The implementation is no easy task however. Social media companies will be very
selective on their collaborators and on research topics because of the fear of losing
reputation
o Furthermore, how to adequately address ethical concerns involved in online
experiments has become a pressing ethical issue in scientific research
• From small N to large N
o Increasing the scale of measurement can also enable us to study more subtle relations
or effects in smaller subpopulations than possible with the sample sizes normally
available in communication research.
o by measuring messages and behavior in real time rather than in daily or weekly (or
yearly) surveys, much more fine-grained time series can be constructed, alleviating
the problems of
simultaneous correlation and making a stronger case for finding causal mechanisms.
o In order to leverage the more complex models because of the larger data sets, we need to
change the way we build and test our models.
• From solitary to collaborative research
2
, o Digital data and computational tools make it easier to share and reuse these resources.
o Fostering the interdisciplinary collaboration needed to deal with larger data sets and
more complex computational techniques.
Challenges and pitfalls in computational methods
• How do we keep research datasets accessible?
o many of the “big data” sets are proprietary ones which are highly demanding to access
for most communication researchers. The privileged access to big data by a small
group of researchers will make researchers with the access “enjoy an unfair amount of
attention at the expense of equally talented researchers without these connections”
o Studies connected to these actors are generally based only on a single platform, which
makes it challenging to develop a panoramic understanding of users behavior on
social media.
o Such access to big data will thwart the reproducibility of the research which serves as
the minimum standard
o the sampling, aggregation, and other transformation imposed on the released data is a
black box, which poses great challenges for communication researchers to evaluate
the quality and representativeness of the data and then assess the external validity of
their findings derived from such data.
o A corpus management tool can help alleviate copyright restrictions. By allowing data
to be queried and analyzed even if the full text of the data set cannot be published.
o Or we could work with funding agencies and data providers to make standardized data
sets available for all researchers
• Is big data always good data?
o Most of the big data are secondary and intended for other primary uses most of which
have little relevance to academic research.
• The gap between the primary purpose intended for big data and the secondary
purpose found for big data will pose threat to the validity of design
measurement, and analysis in computational communication research.
o Big does not mean that it is representative for a certain population. People do not
randomly select into social media platforms
o This also means that p-values are less meaningful as a measure of validity.
• Are computational measurement methods valid and reliable?
o The unobtrusiveness of social media data makes them less vulnerable to traditional
measurement bias, however, this does not imply that they are free of measurement
errors.
o Measurement errors can be introduced when text mining techniques are employed to
identify semantic features in user-generated content.
o It should also be noted that classical methods of manual content analysis are also no
guarantee of valid or reliable data.
o The validity of a method or tool is dependent on the context in which it is used, so
even if a researcher uses an existing off-the-shelf tool with published validity results,
it is vital to show how well it performs in a specific domain and on a specific task.
• What is responsible and ethical conduct in computational communication research?
o the scientific community and the general public have expressed growing concern on
ethical (mis)conduct in computational social science.
o Different steps
• How to get informed consent from the subjects?
• To what extent should the data be anonymized and sanitized for the sake of
privacy protection, how can we achieve balance between privacy and
research?
• How do we deal with the findings that our digital traces can reveal a lot of
very personal traits
• How do we get the needed skills and infrastructure?
o We need to invest in skills (1), infrastructure (2) and institutions (3)
3
, o Many digital traces and other big data are textual rather than the numerical data most
communication scholars are trained for and used to, and will require us to hone our
skills in natural language processing.
• We expect that doing research in communication science will increasingly
demand at least some level of computational literacy.
o Move to a culture of sharing and reusing tools and data. Bigger teams have more skills
and resources, but there is no need for all steps of the process to be taken within the
same team or project
o It is important that researchers and institutions give credit to development and sharing
of tools and data.
Experimental evidence of massive-scale emotional contagion through social networks (Kramer et al.,
2014)
• Emotional states can be transferred to other via emotional contagion, leading people to
experience the same emotions without their awareness.
• Experiment on Facebook, tested whether emotional contagion occurs outside of in-person
interaction between individuals by reducing the amount of emotional content in the News
Feed.
• When positive expressions were reduced, people produced fewer positive posts and more
negative posts
• When negative expressions were reduced, the opposite pattern occurred.
• These results indicate that emotions expressed by others on Facebook influence our own
emotions, constituting experimental evidence for massive-scale contagion via social networks
• This work also suggests that, in contrast to prevailing assumptions, in-person interaction and
non-verbal cues are not strictly necessary for emotional contagion, and that the observation of
others' positive experiences constitutes a positive experience for people.
Why
• Data from a large, real-world social network collected over a 20y period suggests that longer-
lasting moods can be transferred through networks as well
o However, this interpretation has come under scrutiny due to the study's correlational
nature, including concerns over misspecification of contextual variables or failure to
account for shared experiences.
o An experimental approach can address this scrutiny directly, however, methods used
in controlled experiments have been criticized.
• Maybe the negative emotion come from the unpleasant interaction instead of
the emotion
o Prior studies have also failed to address whether nonverbal cues are necessary for
contagion to occur, or if verbal cues alone suffice.
Previous research
• Emotional contagion occurs via text-based computer-mediated communication
• Contagion of psychological and physiological qualities has been suggested based on
correlational data for social networks generally
• Peoples emotional expressions on Facebook predict friends' emotional expressions, even days
later
To date, however, there is no experimental evidence that emotions or moods are contagious in the
absence of direct interaction between experiencer and target.
Experiment
• Manipulated the extent to which people were exposed to emotional expressions in their News
Feed. This tested whether exposure to emotions led people to change their own posting
behaviors, in particular whether it led to post content that was consistent with the exposure.
4
Week 1...................................................................................................................................................................... 2
Literature ............................................................................................................................................................. 2
When communication meets computation: opportunities, challenges, and pitfalls in Computational
communication science. Wouter van Atteveldt & Tai-Quan Peng (2018........................................................ 2
Experimental evidence of massive-scale emotional contagion through social networks (Kramer et al.,
2014) ............................................................................................................................................................... 4
R steps and important functions........................................................................................................................... 5
Week 2...................................................................................................................................................................... 5
Literature ............................................................................................................................................................. 5
Text analysis in R (2017) ................................................................................................................................ 5
Political migration discourses on social media: comparative perspective on visibility and sentiment across
political Facebook accounts in Europe ........................................................................................................... 8
Sourcing pandemic news: A cross-national computational analysis of mainstream media coverage of
COVID-19 on Facebook, Twitter & Instagram............................................................................................. 10
SUMMARY SLIDES LIT............................................................................................................................. 13
R steps and important functions......................................................................................................................... 13
Week 3.................................................................................................................................................................... 14
Literature ........................................................................................................................................................... 14
The validity of sentiment analysis: comparing manual annotation, crowd-coding, dictionary approaches,
and machine learning algorithms (van Atteveldt et al., 2021) ...................................................................... 14
Uncivil and personal? Comparing patterns of incivility in comments on the Facebook pages of news
outlets ............................................................................................................................................................ 16
R steps and important functions......................................................................................................................... 20
Extra information ............................................................................................................................................... 21
Naive Bayes:............................................................................................................................................. 21
o Processing Information: Naive Bayes is a probabilistic model based on Bayes' theorem. It assumes that
the features used for classification are conditionally independent given the class label. Despite its "naive"
assumption, Naive Bayes often performs wel in practice for text classification tasks. ..................................... 21
o How It Works:........................................................................................................................................... 21
Week 4.................................................................................................................................................................... 22
Literature ........................................................................................................................................................... 22
Advancing Automated Content Analysis for a New Era of Media Effects Research: The Key Role of
Transfer Learning (Kroon et al., 2023) ......................................................................................................... 22
Limitations of traditional content analysis techniques.................................................................................... 23
R steps and important functions......................................................................................................................... 27
Extra information ............................................................................................................................................... 28
1
,Week 1
Literature
When communication meets computation: opportunities, challenges, and pitfalls in Computational
communication science. Wouter van Atteveldt & Tai-Quan Peng (2018
The role of computational methods in communication science
The recent acceleration in the promise and use of computational methods for communication science
is primarily fueled by the confluence of at least three developments:
• A deluge of digitally available data
• Improved tools to analyze data
• The emergence of powerful and cheap processing power and easy to use computing
infrastructure for processing these data
• Many of these new data sets contain communication artifacts such as tweets etc.
• Many communication theories are cited as major backbones in many computational studies
• Computational communication science studies generally involve:
o Large and complex data sets
o Consisting of digital traces and other naturally occurring data
o Requiring algorithmic solutions to analyze
o Allowing the study of human communication by applying and testing communication
theory
• Computational methods are an expansion and enhancement of the existing methodological
toolbox
Opportunities offered by computational methods
• From self-report to real behavior
o These data allow us to measure actual behavior in an unobtrusive way rather than self-
reported attitudes or intentions
o This can help overcome social desirability problems, and more importantly it does not
rely on peoples imperfect estimate of their own desires and intentions
o Can also help overcome the problems of linking content data to survey data. We can
now trace news consumption in real-time and combine it with survey data to get a
more sophisticated measurement of news consumption and effects.
• From lab experiments to studies of the actual social environment
o Actual environment
o It's no longer difficult to recruit thousands or even millions of diverse or specialized
subjects from crowdsourcing platforms
o The implementation is no easy task however. Social media companies will be very
selective on their collaborators and on research topics because of the fear of losing
reputation
o Furthermore, how to adequately address ethical concerns involved in online
experiments has become a pressing ethical issue in scientific research
• From small N to large N
o Increasing the scale of measurement can also enable us to study more subtle relations
or effects in smaller subpopulations than possible with the sample sizes normally
available in communication research.
o by measuring messages and behavior in real time rather than in daily or weekly (or
yearly) surveys, much more fine-grained time series can be constructed, alleviating
the problems of
simultaneous correlation and making a stronger case for finding causal mechanisms.
o In order to leverage the more complex models because of the larger data sets, we need to
change the way we build and test our models.
• From solitary to collaborative research
2
, o Digital data and computational tools make it easier to share and reuse these resources.
o Fostering the interdisciplinary collaboration needed to deal with larger data sets and
more complex computational techniques.
Challenges and pitfalls in computational methods
• How do we keep research datasets accessible?
o many of the “big data” sets are proprietary ones which are highly demanding to access
for most communication researchers. The privileged access to big data by a small
group of researchers will make researchers with the access “enjoy an unfair amount of
attention at the expense of equally talented researchers without these connections”
o Studies connected to these actors are generally based only on a single platform, which
makes it challenging to develop a panoramic understanding of users behavior on
social media.
o Such access to big data will thwart the reproducibility of the research which serves as
the minimum standard
o the sampling, aggregation, and other transformation imposed on the released data is a
black box, which poses great challenges for communication researchers to evaluate
the quality and representativeness of the data and then assess the external validity of
their findings derived from such data.
o A corpus management tool can help alleviate copyright restrictions. By allowing data
to be queried and analyzed even if the full text of the data set cannot be published.
o Or we could work with funding agencies and data providers to make standardized data
sets available for all researchers
• Is big data always good data?
o Most of the big data are secondary and intended for other primary uses most of which
have little relevance to academic research.
• The gap between the primary purpose intended for big data and the secondary
purpose found for big data will pose threat to the validity of design
measurement, and analysis in computational communication research.
o Big does not mean that it is representative for a certain population. People do not
randomly select into social media platforms
o This also means that p-values are less meaningful as a measure of validity.
• Are computational measurement methods valid and reliable?
o The unobtrusiveness of social media data makes them less vulnerable to traditional
measurement bias, however, this does not imply that they are free of measurement
errors.
o Measurement errors can be introduced when text mining techniques are employed to
identify semantic features in user-generated content.
o It should also be noted that classical methods of manual content analysis are also no
guarantee of valid or reliable data.
o The validity of a method or tool is dependent on the context in which it is used, so
even if a researcher uses an existing off-the-shelf tool with published validity results,
it is vital to show how well it performs in a specific domain and on a specific task.
• What is responsible and ethical conduct in computational communication research?
o the scientific community and the general public have expressed growing concern on
ethical (mis)conduct in computational social science.
o Different steps
• How to get informed consent from the subjects?
• To what extent should the data be anonymized and sanitized for the sake of
privacy protection, how can we achieve balance between privacy and
research?
• How do we deal with the findings that our digital traces can reveal a lot of
very personal traits
• How do we get the needed skills and infrastructure?
o We need to invest in skills (1), infrastructure (2) and institutions (3)
3
, o Many digital traces and other big data are textual rather than the numerical data most
communication scholars are trained for and used to, and will require us to hone our
skills in natural language processing.
• We expect that doing research in communication science will increasingly
demand at least some level of computational literacy.
o Move to a culture of sharing and reusing tools and data. Bigger teams have more skills
and resources, but there is no need for all steps of the process to be taken within the
same team or project
o It is important that researchers and institutions give credit to development and sharing
of tools and data.
Experimental evidence of massive-scale emotional contagion through social networks (Kramer et al.,
2014)
• Emotional states can be transferred to other via emotional contagion, leading people to
experience the same emotions without their awareness.
• Experiment on Facebook, tested whether emotional contagion occurs outside of in-person
interaction between individuals by reducing the amount of emotional content in the News
Feed.
• When positive expressions were reduced, people produced fewer positive posts and more
negative posts
• When negative expressions were reduced, the opposite pattern occurred.
• These results indicate that emotions expressed by others on Facebook influence our own
emotions, constituting experimental evidence for massive-scale contagion via social networks
• This work also suggests that, in contrast to prevailing assumptions, in-person interaction and
non-verbal cues are not strictly necessary for emotional contagion, and that the observation of
others' positive experiences constitutes a positive experience for people.
Why
• Data from a large, real-world social network collected over a 20y period suggests that longer-
lasting moods can be transferred through networks as well
o However, this interpretation has come under scrutiny due to the study's correlational
nature, including concerns over misspecification of contextual variables or failure to
account for shared experiences.
o An experimental approach can address this scrutiny directly, however, methods used
in controlled experiments have been criticized.
• Maybe the negative emotion come from the unpleasant interaction instead of
the emotion
o Prior studies have also failed to address whether nonverbal cues are necessary for
contagion to occur, or if verbal cues alone suffice.
Previous research
• Emotional contagion occurs via text-based computer-mediated communication
• Contagion of psychological and physiological qualities has been suggested based on
correlational data for social networks generally
• Peoples emotional expressions on Facebook predict friends' emotional expressions, even days
later
To date, however, there is no experimental evidence that emotions or moods are contagious in the
absence of direct interaction between experiencer and target.
Experiment
• Manipulated the extent to which people were exposed to emotional expressions in their News
Feed. This tested whether exposure to emotions led people to change their own posting
behaviors, in particular whether it led to post content that was consistent with the exposure.
4
