Before Theme 4: Quantitative Research

(N.B. This is the study blog for the course Theory and Method for Media Technology. More detailed information regarding the theme as well as the course is available on the course page [1].)

First of all, I would like to clarify some difficulties I go through for this assignment. Initially it is not easy to identify quantitative methods clearly in the research papers with a focus on media technology. Based on my search, I notice that most of those papers are applied research paper instead of theoretical research papers. And theory and methods are not always clear stated in applied research paper as the ones within theoretical research area, since mostly applied research discusses more on framework, proposed architecture, and so on. Maybe this is just limited in my sight or we share the same issue here. If you have found any tips on this, please do share with me. Thank you in advance. Secondly, I am confused about how to look for the impact factor for a particular paper. It seems that the impact factor for the journal is available, but only citation number is available for the paper. In this case, my strategy is to choose a journal with high impact factor, then look for an example paper following most cited order.

As a result, The selected research paper is Sensing Trending Topics in Twitter by Aiello L.M et al [2], which published on the journal IEEE Transaction on Multimedia. The journal is of high quality with an impact factor of 2.303. And it also claims that the article influence score is 1.01. The paper compares six topic detection methods on three Twitter datasets related to major events, which differ in their time scale and topic churn rate. So I think it is using quantitative methods, and in a good way. 

Which quantitative method or methods are used in the paper? Which are the benefits and limitations of using these methods?

In the paper, the authors started that "The methods we test cover three different classes: proba- bilistic models (Latent Dirichlet Allocation), classical Topic De- tection and Tracking (a common document-pivot approach) and feature-pivot methods." For me, it is more like data collection and analysis by using topic detection algorithms. When analyzing data in a countable way, it is quantitative method. It is particularly useful to sorting big data and conducting research with a large amount of data. With the considerable data, the results will be more reliable for the research. 

What did you learn about quantitative methods from reading the paper?

The paper presents a good use of quantitative method for big data analysis. In particular I feel impressed for the algorithm part. It is something I never look into before. I feel it is very technology way, and I hope I am on the right track. 

Which are the main methodological problems of the study? How could the use of the quantitative method or methods have been improved?

I have not found any main methodological problems of the study. But I think the use of the quantitative method can be improved with combination to the qualitative method. Take content analysis as an example, quantitative methods are valuable and useful to mapping and categorizing data. As Lacey and Luff (2001) [3] argued, “the quantitative data can be beneficial because it is categorized quantitatively, and subjected to statistical analysis”. In this way, it enables the researcher to easily summarize empirical data by using statistics or graphs for further analyses. However, as Neuendorf (2002) [4] identified in her book, “although quantitative content analysis uses a broader brush to provide a general outline for the research, it is typically less in-depth and less detailed”. In media landscape, Robertson and Levin (2010) [5] also argued that “purely quantitative approaches fail to gain analytical purchase on the meaning-making properties of media texts”. Therefore qualitative methods are useful here to combat this challenge, as “quantitative and qualitative research are common to be used at the same and be viewed as different ways of examining the same research problem” (Gray and Densten, 1998) [6].

Those are all for today’s blog. Thanks for reading. Your valuable comments are welcome. Please let me know if you find anything interesting or want to have a further discussion. I am looking forward to more discussions in seminar with you all.

---

Sources:

[1] https://www.kth.se/social/course/DM2572/page/theme-4-quantitative-research-2/

[2] Aiello, L. M., Petkos, G., Martin, C., Corney, D., Papadopoulos, S., Skraba, R., ... & Jaimes, A. (2013). Sensing trending topics in Twitter. Multimedia, IEEE Transactions on, 15(6), 1268-1282.

[3] Lacey, A., & Luff, D. (2001). Qualitative data analysis (pp. 320-357). Sheffield: Trent Focus.

[4] Neuendorf, K. A. (2002). The content analysis guidebook. Sage.

[5] Robertson, A., & Levin, P. (2010). Europe as Other: Difference in global media discourse. Statsvetenskaplig tidskrift, 112(1).

[6] Gray, J. H., & Densten, I. L. (1998). Integrating quantitative and qualitative analysis using latent and manifest variables. Quality and Quantity, 32(4), 419-431.