SoylentNews is people
SoylentNews
The internet dominates our world and each one of us is leaving a larger digital footprint as more time passes. Those footprints are ripe for studying, experts say.
In a recently published paper, a group of Stanford sociology experts encourage other sociologists and social psychologists to focus on developing online research studies with the help of big data in order to advance the theories of social interaction and structure.
[...] In the new study, the researchers make a case for "online field experiments" that could be embedded within the structure of existing communities on the internet.
The researchers differentiate online field experiments from online lab experiments, which create a controlled online situation instead of using preexisting environments that have engaged participants.
"The internet is not just another mechanism for recruiting more subjects," Parigi said. "There is now space for what we call computational social sciences that lies at the intersection of sociology, psychology, computer science and other technical sciences, through which we can try to understand human behavior as it is shaped and illuminated by online platforms."
As part of this type of experiment, researchers would utilize online platforms to take advantage of big data and predictive algorithms. Recruiting and retaining participants for such field studies is therefore more challenging and time-consuming because of the need for a close partnership with the platforms.
But online field experiments allow researchers to gain an enhanced look at certain human behaviors that cannot be replicated in a laboratory environment, the researchers said.
For example, theories about how and why people trust each other can be better examined in the online environments, the researchers said, because the context of different complex social relationships is recorded. In laboratory experiments, researchers can only isolate the type of trust that occurs between strangers, which is called "thin" trust.
Is Big Data the path to respectability for the social sciences?
More information: Paolo Parigi et al. Online Field Experiments, Social Psychology Quarterly (2017). DOI: 10.1177/0190272516680842
(Score: 2) by looorg on Wednesday April 05 2017, @01:47AM (4 children)
In some sense this sounds just like scaling up what sociologist already do, you could see the current usage of data-analysis as small scale (compared to big data). Still it's data, it usually only lends it self to certain things - it rarely explains why people do things, just that they do them, how they feel or what they think. Never the why, which is something that most social scientists like to know and explain. That X thinks Y is never very interesting. Making the data set BIG isn't going to change anything in that regard, things won't become more accurate (to a certain degree it might - but going from thousands to hundreds of thousands or millions of data points usually doesn't add much) or better - just like things usually don't become better just cause you read a few more books or do another round of interviews. There is usually a limit to new information before reaching saturation, adding more just isn't necessarily going to be better.
For example, for my masters in Sociology I did data-analysis on four years of a national wide survey , about 40k respondents per year. Out of the entire survey, several hundred questions, I only used six questions and half of them was just to classify the respondents by gender, age and political orientation - the rest being the actual questions. I don't think my results would have changed in any meaningful way if I had had 100k respondents or more. So adding more just isn't going to be better automatically. There might have been in my case if the entire population had answered, then there wouldn't have to be any need for random sampling or any such statistical work since it would have been a total survey.
As mentioned data analysis lends itself very well for studies that ask questions such as what are they thinking - not how are they thinking or why are they thinking like they do. You could give then options to the why and how but that is then limited by you - the sender of the survey. The problem is usually that there is no room for such open questions. Doing things online probably won't change that very much. Even if you have a big text box where people can explain the WHY they rarely do. People don't like to explain themselves. So you are usually left with having to do complementary studies where you interview or send more forms to smaller groups of people to try and figure out the how and the why.
I really don't see big data analysis changing much of this. It will be a niche area in a niche area. If you ask me the reason why people look down on social sciences is that there is way to much focus spent on really soft studies (that get more publicity then they deserve), usually about feelings or some really far out topic that most people don't give a shit about (hello gender- and race studies!) and this just can't be quantifiable in any kind of meaningful way. There is already more Qualitative studies compared to Quantitative studies and I don't see this changing. The big data pull will only interest the quantitative scientists so the will be a small field in an already small field.
(Score: 0) by Anonymous Coward on Wednesday April 05 2017, @03:53AM (1 child)
As mentioned data analysis lends itself very well for studies that ask questions such as what are they thinking
No, it only tells you what they claim they're thinking. Nothing more.
(Score: 2) by looorg on Wednesday April 05 2017, @05:01PM
No, it only tells you what they claim they're thinking. Nothing more.
That isn't a problem that is somehow unique for social science. As a matter of fact it's an entire field of study - thoughts, ideas and opinions people claim to have in and towards various groups. How they might be projecting ideas when they believe they are anonymous on the internet compared to in real life, or with close friends and family compared to strangers.
After all unless we develop actual mind reading techniques this will always be the case. The analysis done is done on the data you have and it's assumed to be true unless proven to be false and proving that it is false is usually a really hard thing to do, most likely borderline impossible. The study conducted is then based on the assumption that observations are true and it's generalized from that. If it turns out at some later stage that you fixed the data or that every single respondent lied then it all falls with that, but could probably be used in somewhat interesting manners by or for something else.
That said it usually doesn't matter very much if a few people lie on their questionnaire or in an interview cause a single person is usually not very interesting. On this kind of analysis level it's always about how groups of people think and do things and never on an individual level.
Doing it Big Data style online isn't going to change any of this, people lie online - hide behind "anonymity" or even go as far as to try and obscure their actions by trying to fill it with "noise" just to make the pattern harder to detect. These people with probably be outlier and they'll be removed, ignored or placed in some special category from the data-set anyway for being to out of place.
(Score: 2) by TGV on Wednesday April 05 2017, @05:50AM (1 child)
Sounds right to me. Even worse is that it's nearly impossible to translate any meaningful theory into a test on a questionnaire without making a huge amount of assumptions, and this strategy seems to add a bias in the sampled group and a dependency on the actual tool used. That's far from ideal.
An example. A university teacher recently proposed to have students use Twitter and "sentiment analysis" to do some kind of research. He didn't know the tool, and didn't even realize that a Twitter search yields different results for different users. The group that discussed this proposal was rather international, and most people got similar search results and analyses, but some got results that were clearly tailored to specific countries. Still most people there thought it would be a good tool for the students. The proposal above seems like just a bigger version of the same problem.
(Score: 2) by looorg on Wednesday April 05 2017, @05:23PM
One of the issues they don't seem to touch on is where are you going to get the data, are the people with the data going to want to share or are these Sociologists going to have to get a job at like Facebook or Google and if they do those reports are going to be like smoking reports put out by tobacco companies - everything is fine and dandy etc.
Christian Rudder is some big cheese over at OkCupid and he used their data to write a few books on data analysis called Dataclysm: who we are when we think no one's looking, it showed that you can do interesting things with the data but at the same time he is some big cheese over at the company so it's doubtful anyone else could have done the things he did. It also shows that regardless of or to the book title someone is looking - he and their system was. Question about his study then becomes are these things unique for OkCupid users or are they somehow universal? Would other dating sites show the same patters? Unless they are universal they are almost worthless - it's like being told that Coke drinkers like Coke over Pepsi.
I read the first of Rudders book, but not the second one. It was entertaining and interesting but at the same time somewhat pointless result-wise: People like to date other people with similar interests to their own, in their age group and of their own race. Hardly surprising. Not something you need big data to figure out or hadn't been noticed before. So it's just another tool and as stated in my previous post I do believe this will be a very small and marginal area of Sociology -- mostly cause most of them that I have met are really bad at maths and can't or won't program so these data analysis packages will have to be really damn intuitive and simple to use if they hope to get any users at all cause these people won't write their own stuff in R, mine any social platforms on their own or work with Hadoop and similar.
Also there seems to be some idea that Big Data is somehow immune to all the issues with normal data. Like all the data gathered is going to be golden somehow and not filled to the brim with shit that you can't do anything with or will require so much massaging you could open up your own massage-parlor - there will be happy endings for all involved after that.
https://www.amazon.co.uk/Dataclysm-When-Think-Ones-Looking/dp/0008101000 [amazon.co.uk]
https://www.amazon.com/Dataclysm-Identity-What-Online-Offline-Selves/dp/0385347391 [amazon.com]