Before: Theme 4

For this week's theme, I have chosen the article Everyone's an influencer: quantifying influence on twitter, published in WSDM '11.

Which quantitative method or methods are used in the paper? Which are the benefits and limitations of using these methods?
As the name of the article suggests, the main quantitative method used is the collection of millions of tweets over a time period of two months. Right here, we have the first benefit - it is very easy to get a hold of data, and the speed of collection and data creation is astounding as there are millions of tweets created every day. The authors do not have to create their own data, only read and collect it. This also brings us to another point which could be seen as both a benefit and a limitation. The authors do not have control over the data created. This aspect is great as the data is pure and objective, however they can not directly control the kind of data they get, as you would be able to in a user study where you could formulate the questions yourself.

The purpose of the study was to find tweets containing a link-shortened URL's to try and see how that URL spreads over Twitter. To do this, they analyzed the followers of every person doing a tweet, to see if they retweeted the URL, and then repeating the same operation until they reached the end of that URL. One limitation they bring up is that they only analyzed active users, which makes their results not reliably representable for the entire Twitter user base.

What did you learn about quantitative methods from reading the paper?
I have learned that using existing data like tweets is highly desirable when conducting empirical studies of diffusion such as this one. It is a great way to collect a lot of data without having to manually create a set of questions for research, which is hard to do in a objective way.

Which are the main methodological problems of the study? How could the use of the quantitative method or methods have been improved?
The authors do not try to predict how a ”word of mouth” campaign could spread on social media in general, they only analyze how it spreads on Twitter. The Twitter ecosystem when is comes to followers and friends does not exist on every social media platform out there, which makes this study quite unique and bound to Twitter. To make a more general approach, they could have done the same study on several social media platforms to try and get a more general understanding of how things spread on social media.

While two months is a long time when it comes to social media, the gathering of the data could have been conducted during a longer time period, or perhaps during time periods with an interval such as every other month for a year.


Drumming in Immersive Virtual Reality: The Body Shapes the Way We Play
The main point of the paper was to see if people act differently when their body suddenly is transformed into another in an immersive VR environment. The first thing I reacted to was that the entire test group were caucasians. Is this something the authors did because of technical limitations, or did they only want to study how caucasian people act?

The results demonstrate that full body ownership illusions can lead to substantial behavioral and possibly cognitive changes depending on the appearance of the virtual body. I find this extremely interesting from a story-telling/gaming perspective. Only in rare cases do I actually "connect" with a game character enough to let the character in a way decide how the game will progress or how I will act, and in those cases the story has been extremely well done. This makes me wonder if you could use a VR body illusion as a substitute for a stronger story? Would a player connect more with the game character if they in VR actually inhabited the body of the character? If we are to believe the results of this study, that would probably be the case!

Which are the benefits and limitations of using quantitative methods?
One of the largest benefits of quantitative methods, to me, is that when using large enough data sets, we can generalize the results to fit an entire population. The immediate limitation to this is that the data set you use for your quantitative research has to be random enough in order to get data to represent a population. Another benefit is that you choose to only research a subset of a population, you could also hand-pick your data to fit said subset. Quantitative methods are also a lot less time-consuming when it comes to gathering and analyzing the data as compared to qualitative methods.


Which are the benefits and limitations of using qualitative methods?
Using qualitative methods, you can get a deeper understanding of the data as you could form the questions you ask and thereby the data you collect. The drawback here is that you really can not get the same amount of data as you could with quantitative methods, without putting down lots and lots of time into it. However, this makes understanding why the data looks like it does much easier as you could easily ask questions such as "why?". Naturally, as the amount of data you collect is usually much less compared to quantitative research, it is hard to draw any general conclusions about a population.

Of course, by combining both qualitative and quantitative methods, you could get the best of both worlds as the benefits and limitations are usually connected to each other across the different methods.