Evaluating the method: Twitter data

Jannie Phienboupha

Riots and Twitter: connective politics, social media and framing discourses in the digital public sphere
Summarize briefly each study: what are the researchers trying to accomplish?  Where do they get their data? What are their findings? What are they arguing?
The researchers in this article are trying to fill in the gap between internet activism and mobilization. Data was based around the comparative hashtags of #RiotCleanUp and #OperationCupofTea. The researchers found that clean-up movements were discursive political acts where celebrities played a huge influential role in framing the discourse. The #RiotCleanUp was found to not explain mobilization compared to the more emotive #OperationCupOfTea. The researchers also argued that ambience, the engagement of people in discourses that translates into an action frame, must be a product of systemic interaction among software, users, and text.

How are they using twitter data?  What methods are they employing? What is their methodological approach to coding? (inductive? Deductive?)
Content coding and close textual reading was used to characterize discourse with major riot-related hashtags and then metrics were used for assessing and comparing influence. There were three elements to identify, analyze, and compare hashtag discourse: 1) “it must establish an overview of discourse, which means identifying the dominant hashtags during the relevant period, 2) it must be able to differentiate between these hashtags in a way that supports a critical analysis of their relative influence. In doing so, it must recognize and allow for the interactive dynamics of the Twitter system, 3) Third, it must provide a mechanism for interpreting this influence in terms of connective action – in other words, it must reveal clues as to why some discourses energize group mobilization and others do not”. Since the researchers were pulling 1000 tweets to analyze, their approach was inductive. They were looking for a common pattern/occurrence.

What are the strengths and weaknesses of the methods that they are using?  What do they capture well? What do you think they are missing out on? If you were to conduct this study, would you do anything differently?  What and why or why not?
One of the weaknesses of the approach mentioned in the article is that it does not describe discourse in the riot public, the researcher has to be able to differentiate between discourses that can explain connective influence in their own way. Furthermore, the researchers collected 1000 tweets with 7 of the hashtags, thats approximately 142 tweets for each tweet. It would be beneficial to widen the range of tweets collected by increasing the collection of each hashtag or decreasing the amount of hashtags to increase the number of tweets that can be collected.

Understanding a digital movement of opinion: the case of #RefugeesWelcome
Summarize briefly each study: what are the researchers trying to accomplish?  Where do they get their data? What are their findings? What are they arguing?
The researchers in this study are trying to analyze the digital force combining properties of social movements and public opinion. In other words, they are trying to conceptualize and capture particular cases where citizens voices rise from social media is combined with traditional notions of public opinions and social movements. They call this “digital movement of opinion” (DMO). They use the hashtag #RefugeesWelcome on Twitter to collect their data. The researchers found that the hashtag, with a prominent number of elites followed by the masses, created a powerful digital voice that provided legitimacy to the refugee crisis and pro-refugee movement in the government.

How are they using twitter data?  What methods are they employing? What is their methodological approach to coding? (inductive? Deductive?)
The researchers used a triangulation of methods and qualitative text-based validations. They pulled tweets using the #RefugeesWelcome to look how the hashtag connected to DMOs, therefore the study was inductive.

What are the strengths and weaknesses of the methods that they are using? What do they capture well? What do you think they are missing out on? If you were to conduct this study, would you do anything differently?  What and why or why not?
Compared to the other article, this research only used 1 hashtag compared to 7 to analyze. Is this too little? I feel like there should be at least 2 to create a more diverse sampling.

2 thoughts on “Evaluating the method: Twitter data

  1. Hi, Jannie! Good job on your summaries of the articles. They were very comprehensive. I agree with you that for the first study that it would be beneficial to widen the range of tweets collected by increasing the collection of each hashtag or decreasing the amount of hashtags to increase the number of tweets that can be collected. For the second study, I also think more hashtags should be analyzed to have a more diverse sample.

  2. Jannie,
    Great summaries of the research. I like how your first sections were concise and provided a very brief but detailed enough write up to understand what the research was about- I could grasp the purpose of the research and could have gotten a basic understanding even if I had not read the article. Then in your section two writings you provide enumerated methods for how the data was codified/extracted. I agree with and made similar comments regarding the strengths and weaknesses of the data collection methods for both studies- having too few Tweets provides less than sufficient volume of Tweets to understand the phenomena in question, although time and other resources area always a consideration. In research study II, the opposite problem is present, there may be too many Tweets collected because where over a million is effective in making a representative sample it can cause an overwhelming demand on coding and data analysis- although I believe that starting small and scaling up helps to understand what methodologies can be most effective in big data research. Thanks again for posting, I really enjoyed reading.

Leave a Reply