Introduction:
The topic I have
chosen is Gender, focusing on women being stereotypically emotive and men more
dominant. The aspect of language that I have chosen to analyse is the context
and length, also how straight to the point both sets of tweets are. MY
hypothesis is that men are more dominant and straight to the point. Whereas
women appear to 'waffle' and be more emotive linking the Dominance and Deficit
Theory.
Methodology:
To make sure that
my tweets were unbiased, I chose a tweet after every certain amount. So every
third tweet was chosen so that it was systematic. Due to it being systematic, it
didn't have any obvious anomalies.
Analysis:
Overall, we have
found out that our data does not support our hypothesis. The primary reason is
because we chose extremely different people: a political leader and a reality
star. What's significant is how unexpected our results were. Ed Miliband was more
emotive than Amy Childs. The quantitive evidence to support my hypothesis is
that women use more emojis and hash tags and so could be considered emotive.
However there is also evidence to disprove it.
Conclusion:
My findings
disprove my hypothesis. The main problem was that our approach to the choice of
people was wrong. We should have chosen people that were a lot more similar;
instead of choosing a politician and a reality star, we could have chose two
reality stars (like Joey Essex and Amy Childs) or two politicians (Ed Miliband
and Theresa May).
Also the data we
chose for Ed Miliband was from when he was under pressure to get public votes
for the election. He had to be positive and supportive of the public.
For the Future:
The draw backs of
using Twitter is that people can out what they want on it. They don't have to hide
from anyone and may react differently than they naturally would.
In the future I
would need to choose more suitable people for the investigation. I would also have
to choose the tweets from a more suitable time when there wasn't any elections.
A similar age for
the people would also be a positive.
Good ideas. You have thought in depth about comparability. In terms of reliability, choosing every nth tweet is good to get a spread of data (you could think about how often they tweet before deciding if it is every third, seventh, tenth or even 20th to get a sense of what is typical of their language use, not just what they are like on an 'off day!). This will not avoid any anomalies altogether as any single tweet could by atypical of a subject's usual tweeting style but it will reduce the effect of a whole swathe of them e.g. if someone is grieving, cross about something, ill etc on a particular day/week. You also need to find out if they tweet themselves or have aides who do it (famous, busy people have others who pretend to be them to tweet for them) and evaluate the effect of that.
ReplyDelete