Research Reflection

Hello and welcome back!

Thirteen weeks later, my research on extracting data about feminism from Twitter is complete – but before I celebrate my last submitted assessment of the semester with a swagalicious, Obama-esque mic drop, I’m going to reflect on what I have learned about research practice in BCM212 this semester.

giphy (8)
source: giphy

A Brief Recap

Last semester in BCM240 I conducted a research task on a topic I picked at random at the last minute. I didn’t properly understand it and I had no interest in it whatsoever. This was reflected in the work I produced and in the grade I received. This semester, I was determined to not make the same mistake again.

A reading on GamerGate I studied for another subject outlined how to issue-map a controversial issue over social media platforms. I was fascinated by the design of the research because I hadn’t come across it before and thus I decided to recreate it on a much smaller scale, using feminism instead of GamerGate on the basis of this curiosity. Why draw up a survey and get, say 100 responses at best, when I have the entire twitter database at my disposal? View my research proposal and research update if you’d like to learn more.

It was a risky project to choose in that I’m more researching a form of research design – my aim is to figure out how much information I can extract from one hashtag – than a topic in particular. It’s really paid off in both my enjoyment of the task and the marks I have received so far. It made it really simple for me to follow my project plan (pictured below) and complete each task on time and to the best of my ability.

plan

This has been a valuable lesson to me. As someone who is not generally comfortable thinking outside the box and trying new things, I have discovered it can be really rewarding and insightful to walk down a new path.

Research Values

Socially responsible research design was difficult to comprehend in my research because I used social media data from Twitter to analyse a conversation. This meant I could not offer a privacy disclaimer, nor obtain consent from each user whose data I analysed (Buchanan et. al 2016). Internet data is a relatively new form of research. It has somewhat blurred the line between public and private data and, as such, there are no uniform principles on how to undertake it (ibid). All I had to go from was Twitter’s Privacy Policy documentwhich states that a user’s data, once posted to Twitter, may be stored and disseminated for a variety of uses (Twitter 2016). A user may opt to have their tweets and information kept private, but the default setting is public. If the user has taken no steps to change this, the data is considered public (Buchanan et. al 2016) and I have fulfilled my legal requirements as a researcher. However, this does not necessarily make my research ethical (University of Western Australia 2017).

To combat this I wrote a blog post on the difficulties of obtaining consent in a mass-social media data study and invited any user who did not wish for their data to be read to contact me regarding their concerns. Although I posted in on Twitter and tagged it “#feminism”, the same hashtag I was studying data from, I understand it is unrealistic to believe that all, or even most, participants will view it. This meant that my social responsibility value was still not fulfilled. I thus had to ensure identities remained protected, throughout my research I did not once mention any individual’s name or Twitter account specifically, all the data was discussed collectively. This is due to several factors; a user may be underage or come from an environment which is culturally sensitive now, or which may become so in the future (which is impossible to predict), in which case publishing or otherwise identifying their Tweets could jeopardise their liberty, or even their life in some extreme cases (Markham 2012). This is not something I had considered prior to beginning this project, I had assumed that once somebody posted data online it was public, but I have learned this doesn’t necessarily mean it can be used without careful consideration.

I strived to remain flexible in collecting and analysing my data. This involves remaining open-minded to surprise twists in the story of data collection (Grafanaki 2007). There were certain results which surprised me. For example, I had assumed that the underlying content discussed under #feminism would be the feminism argument in itself, but I discovered that (in the studied period) the conversation was dominated by public affairs or events which related to feminism in some way.

Overall?

 I feel that this project was successful. Despite being relatively small-scale, only collecting three days’ worth of data, over 5000 tweets were analysed collectively and I managed to pull some fascinating information out of Twitter which revealed a lot more about the feminism conversation than one might anticipate it would. In future subjects I hope to adopt similar methodology on a larger scale and develop more knowledge in the area of social media data research, because it’s a growing area with a growing interest from academic researchers and marketers alike.

References

Burgess, J 2016, Mapping sociocultural controversies across digital media platforms: one week of #gamergate on Twitter, YouTube and Tumblr, Communication Research and Practice, Vol. 2 (1) (http://www.tandfonline.com/doi/abs/10.1080/22041451.2016.1155338?journalCode=rcrp20)

Grafanaki, S 2007, How research can change the researcher: The need for sensitivity, flexibility and ethical boundaries in conducting qualitative research in counselling/psychotherapy, Taylor & Francis, vol. 24 (3) (http://www.tandfonline.com/doi/abs/10.1080/03069889608253017

Markham, A 2012, Ethical Decision-Making and Internet Research 2.0: Recommendations from the AoIR Ethics Working Committee, Association of Internet Researchers, accessed 16.04.17, https://aoir.org/aoir_ethics_graphic_2016/

Twitter 2016, Twitter Privacy Policy, Twitter, accessed 16.04.17, https://twitter.com/privacy

University of Western Australia 2017, Research Data Management Toolkit: Ethics, Privacy, Consent and Legal Issues, University of Western Australia, accessed 4 June 2017, http://guides.library.uwa.edu.au/c.php?g=325196&p=2178575

Obtaining Consent from an Invisible Audience

Obtaining consent is an ethical requirement of undertaking research, yet there is a gap in research ethics in terms of collecting online data. My research project (see my proposal here) involves the mass collection of tweets containing “#feminism” over three days to issue map feminism online. Unlike most other students in BCM212, who are setting surveys to collect data, setting a simple consent disclaimer is not something I can realistically achieve. The vast majority of my respondents will not be aware their data is being used for my research.

giphy-downsized (2).gif
source

The line between public and private data has been fading rapidly since the birth of big data in 2012. Therefore it is difficult to set an ethical standard in collecting online data. There is no industry standard, so I have used this guide on ethics from the Association of Internet Researchers to draw my own strategy.

I will not store, nor publish, any data which identifies individuals. I will be focusing on collective mass data to investigate attitudes toward feminism. I will post examples below. This is for three primary reasons:

  1. Protect the identity of minors;
  2. Protect the identity of those who come from a sensitive cultural context, or those who may find themselves in such an environment in the future; and
  3. Protect the identity of those who simply do not wish to be exposed.

In addition, if you see any aspect of my research which you feel identifies you in some way you do not appreciate, drop me a line in the comments and I will fix it for you. Alternatively you can reach me via Twitter (@claireee096) or via email (cf207@uowmail.edu.au).

Examples of Data Collected

This slideshow requires JavaScript.

Although it is unrealistic to aim to reach everybody in the feminist conversation on Twitter, I will regularly post updates on the thread so that anyone can keep up with my research.

Thanks for your interest in my research!

-Claire

Issue-Mapping Feminism on Twitter: Research Update

MEMEbcm
Original Image

I have always been slightly frustrated at the notion of a ‘trending topic‘ on Twitter. Lots of people are tweeting about the same thing? SO WHAT? What does that tell us? What does it contribute to?

BOOM! Just like that, my research topic was born.

Recently I published this research proposal about issue-mapping the feminism discussion on twitter. After receiving very positive feedback, I have enthusiastically continued forward on my venture to discover what lies beyond the mere “trending” Twitter conversation.

giphy-downsized.gif
source

 

I have navigated around several tools designed to measure tweets through demographics, sentiment analysis and across various social media platforms. I have decided to use keyhole, a piece of software which tracks hashtags in real time. As per Susan’s feedback I have reduced the scope of my research. I have decided to record just three days worth of twitter data. This will give me ample time to scrounge through all the facets of the vast amount of data I expect to collect.

Something I have found fascinating thus far is the neutrality of sentiment (a measure of the attitude of a piece of writing) in tweets about feminism. Generally, when I picture a feminist, I see someone extremely passionate or emotional about a particular topic –

timentforblog.JPG
source: Sentiment of #feminism at as 3.12pm 23.04.17

and I picture the same for someone tweeting against feminism. Despite this, over 60% of tweets containing “#feminism” have been classified as neutral. Of course, the system is far from perfect, but it definitely was not the result I was expecting. I began my three-day data collection period yesterday, so it will be interesting to see if this fluctuates and to determine why.

 

To stay in touch or give feedback for my research project, you can:

(a) Follow & comment on my blog

(b) Follow me on Twitter: @claireee096

(c) Email me at cf207@uowmail.edu.au

 

Thanks for your interest!

-Claire

 

 

 

Take A Number

“If you would take a man’s life, you owe it to him to look into his eyes and hear his final words. And if you cannot bear to do that, then perhaps the man does not deserve to die” (George R.R. Martin, Game of Thrones)

1944

The Nazi Germans of World War II assigned serial numbers to the Jews they kept in camps. This number was sewn to their prison uniforms, and often tattooed on their bodies, ridding them of their identity as humans. Because the identities of Jews were hidden behind so many numbered punchcards“, it was a morally simpler task for the Nazis to exterminate them. It is extremely difficult (in a psychological sense) to kill another human personally. Therefore in dehumanising the Jews, it became possible for great atrocities to take place.

jew1
source

***

2017

It’s final exam week. I walk into a large hall, dubbed ‘building nine‘. I sit at the table marked ‘A38’, as instructed by an automated email sent two weeks prior. This number will be my identity for the next three hours; one hundred and eighty minutes; 10 800 seconds. I write my student number, a seven-digit code I’ve learned to memorise (and which means more to the university than my name), at the top of each exam paper. I sneak a look at my surroundings; hundreds of tables span in neat lines, like a perfect digital binary code. When the clock strikes 9am, I scribble furiously with my pen until my wrist aches. I fill in little dots to mark the slabs of content I’ve been forced to memorise over the past thirteen-or-so-weeks. When they tell me to stop, I leave the hall. I go home and wait for the number to appear on my screen; the number which will tell me whether or not I have to repeat the subject next time.

hall.jpg
Typical exam environment (source)

They may not be legitimising genocide anymore, but numbers are still identifying and dividing us.

In high school, every report card, every exam and assessment was ranked from the top student to the bottom. My maths teacher would post a ‘top 10’ ranking on the wall of the classroom after each mathematics assessment. I made the list every time except once, and everyone noticed. And wondered why. Aloud. They asked both me and the teacher, and made me more disappointed in myself than I really should have been. What happened, Claire? OMG Claire’s not on the list. And (my personal favourite) I BEAT CLAIREIt wasn’t that my mark was bad, it was my ranking. I felt like shit, though. That was the first time I realised the issue with ranking in schools; it’s all well and good until you aren’t at the top.

Another teacher said the best way to ensure a great ATAR (Australian Tertiary Admissions Rank) was to get a good rank in every class. Here’s the issue with that useless piece of advice; not everyone can rank well, by definition. Conversely, if that teacher had said the best way to ensure a great ATAR was to average an 80% + grade in each subject, every student could (in theory) achieve that goal. It doesn’t matter how hard the whole cohort tries in a ranking system; there will always be a bottom group – and that’s just a really shitty system. But welcome to the ATAR. If you’re ranked first in your class, you’re guaranteed the top mark in the HSC exam produced by your cohort. If you’re ranked 12th, you will receive the 12th highest exam mark, and so on.

This actually ended up working well in my favour. My maths rank wasn’t the best towards the end of year 12, but there was one assessment left: a poster summary. All we had to do was summarise a topic from the previous year and present it on cardboard. It was me in a class full of wannabe engineers; super-smart boys who were fluent in the language of projectiles and trigonometry, but who didn’t know glitter glue from ribbon. My teacher therefore loved my super sexy maths poster, and I jumped to rank five in the course and therefore received an exam mark which I was very pleased with, but probably didn’t deserve.

glitter
source

As big data approaches the mainstream axis of information technology, the issue of using numbers in classifying and coding humans is as relevant as ever. At the same time there is a push from researchers to adopt qualitative data techniques to combat the effects of this.

-Claire

Research Proposal: Social Data Analysis of the Feminist Movement

hash.jpg
source

My passion for digital media lies in its limitless nature. Entire conversations about a topic exist publicly online. There’s so much to be learned from these conversations, but generally analysis only goes so far to determine that a topic is trending, yet neglects to tell us what is really being said. Thus there is a gap in the market for issue-mapping in sociocultural online conversations – a gap which is screaming my name.

I read an article during the summer holidays on issue mapping and social media controversies, which focused on #GamerGate as a social case study. I was completely inspired by the research methodology; Burgess went beyond the trending topics algorithm presented on Twitter (which is limited to identifying what is being spoken about – without detail – and the rate at which it spreads between nodes) and used various issue mapping tools to generate an understanding of what was actually being said about GamerGate on a global scale, which allowed the researchers to gain a clear and vast understanding of the players in the debate.

Imagine this; you’re sitting at a table in a busy cafe. The woman sitting at the table behind you has said “the feminists” in the midst of her conversation. As you walk out to pay the bill, you hear the word being spoken over and over, throughout the different tables, by all different people. Curiosity becomes you; who are the feminists? What are people saying about them? Why are they talking about feminism now?

Internet environments are structured similarly; we know that a topic is trending, but we don’t know what is being said, where the influence comes from or who the instigators or leaders are. Burgess’s article dives into the gaps of information analysis and opened up a new world of opportunity for me. As a result of this I intend to look at the conversation as a whole on Twitter, instead of designing a survey or organising a focus group of a limited sample.

To scope this project, I will focus on the #heforshe campaign, and analyse the conversation surrounding it on twitter. I will map various responses in relation to different stimuli and media events (such as Emma Watson‘s UN speech in 2014), and over time, using open source programming, sentiment analysis and tools such as keyhole.

sentiment1
(source) Sentiment Analysis measures the positive/negative words in a tweet to determine direction of conversation

As an aspiring social media analyst, this project will hopefully be the first of many, and provide me with an insight into the various tools and methodologies of social media data analysis to determine the effectiveness of a digital campaign.

-Claire

 

you don’t put your hand on a hotplate twice (unless you’re me)

cur.jpg
source

The first way we experience and discover the world is through our senses. As infants we do not have the luxury of learning in a multi-tiered lecture theatre; we “absorb information through (our) senses”. (Doorley, 2014). We are inherently curious about the world surrounding us, and at the most basic level we use our sight, smell, taste, hearing and touch to investigate unknown materials.

The most classic example (for me) of wanting to know something, and using sensory knowledge – as a form of research – to discover the answer, is the hotplate. Is it hot enough to put food on ? There’s no faster way of finding out than (DISCLAIMER: don’t try this at home) putting your hand on it. I’m not talking one dainty finger, either, I’m talking a mad palm-slap to the centre of the hot plate. Obviously that’s a shitty idea, but I when I’m curious about something I don’t necessarily acknowledge why I am curious, or look for background knowledge or sit back and think of how best to handle that particular situation, I just explore it. I don’t even have time for “I’m going to touch this hotplate to see if it is hot” to process through my mind, I just react – yes, I’ve done it several times.

The point of curiosity in this instance is to learn; the average human would learn to associate the pain of a burnt hand with touching a stove, and thus stop doing it (I’ll get there one day). Curiosity is how we learn to either refine or take a question further, focus on a new question, or simply learn no.

Curiosity is a fundamental aspect of research. It’s how we learn. For me, it’s an inherent, strong desire to discover connections, find answers and build on the sum total of human knowledge in an area. The same chemical reaction in my brain behind my urge to discover the temperature of my stove is the same as what causes me to wonder about the changing scope of film distribution across platforms, for example. We’re human, we want to know things; this is a distinctive hunger which creates the foundation for passionate and informed research – but hopefully I’ll find a safer and more fulfilling research topic than hot stoves in BCM212 this semester.

giphy

 

source

If the hot plate didn’t kill this cat, maybe curiosity will 😉

-Claire