I ran textometrica on 500 tweets a week after this round of tweet analysis, and got similar (and some interesting) results. You can’t really compare the two network maps because I added some categories. This is, after all, a leaning process at the moment.
Last time I omitted the category for links, but this time I coded for it. There was a strong correlation to advocacy (sharing links to shelters, websites, etc), to types of abuse, but also to legal action – a category similar to advocacy, but is delimited to actions in the legal system. I also coded for words that showed support or blame for the violence. I really need to run a sentiment analysis on the data, though. Anyway, what I did find was a similar high number of tweets showing advocacy for victims (sending hotline numbers, telling where shelters are, etc), as well as a large discussion about what constitutes abuse (i.e., is emotional abuse a type of domestic violence). I also found that women were mentioned twice as much as men (no surprise there, considering statistics of domestic violence). Three categories were also interesting: the high number of tweets both blaming the victim, but also blaming men in general; how often social media was mentioned (often in relation to advocacy), and of course related to links, and how prevalent the discussion about abusers entering the legal system.
Next time I run the analysis I will use a different tweet miner. I was using TwitterSave, but it only lets you download max 500 tweets at a time. Next time I will try out TAGS, which lets you download more tweets by mining hashtags. Then, of course, running them through textometrica again. We will see where it leads.