Comment by aaron-lebo | Hacker News Reader

Comment by aaron-lebo | Hacker News Reader

Here is the paper:

http://politicalbots.org/wp-content/uploads/2016/11/Data-Mem...

I really dislike the basis of their research which is:

Twitter provides free access to a sample of the public tweets posted on the platform. The platform’s precise sampling method is not known, but the company itself reports that the data available through the Streaming API is at most one percent of the overall global public communication on Twitter any given time. [6]

So...we aren't going to mention a selection issue at all? Is that 1% of global Tweets a truly random sample? Can you confirm Twitter isn't manipulating the output? We have examples of them censoring stuff this cycle...

In order to get the most complete and relevant data set, the tweets were collected by following particular hashtags identified by the team as being actively used during the debate.

What's the methodology for selection? I'd hope research out of Oxford would be a little better argued. But of course since it's Oxford this research will get quoted in the New York Times by people who probably don't understand the paper they are reporting on and we'll have talking heads endlessly parroting why this justifies censorship.

All this research is showing evidence of is that the stream Twitter provides is pro-Trump. It's a massive leap of faith to assume this data truly is representative and non-biased.

Ignoring whether it's pro-Trump or pro-Clinton, don't you find the idea of automated parroting of political speech via bots a bit obnoxious?

And maybe the platforms - Twitter, etc - have a responsibility to not publish those posts?

> Ignoring whether it's pro-Trump or pro-Clinton, don't you find the idea of automated parroting of political speech via bots a bit obnoxious?

In a way, I'm terrified that fascism can now just "chatterbot" its way into the popular consensus. In another way, I'm incredibly relieved that a significant portion of the fascist speech I see on the web isn't real.

This has always been my concern about Twitter. If taken as an actual sample of the behavior of the population, anyone with a reasonable amount of money and time can portray the population as they choose. It's too easy to manufacture scandal, racial tension, phobias, inflammatory news and then treat a response as somehow reflective of the populace.

I've grown up and spent my entire life in the south. I've never had a parent, relative or friend even hint that I should view/treat any race differently than any other. You'd hear the occasional bad joke at school...but that was about it.

From the news cycle at this point, I feel like I'm supposed to have been mailed my own KKK gear as an expected housewarming gift.

Our best research indicates there's virtually no chance you actually view each race exactly the same.

His actual feelings aren't the point. The stereotypical venal and public racism the South is portrayed as being awash in, is nowhere in evidence.

I also live in the South and had different childhood experiences, but I'm older.

Did you just argue there is no evidence of racism in the South?

https://www.splcenter.org/hate-map

Looking at that map could explain my experience then. I'm in South Carolina and there seems to be significantly less here than surrounding areas. For what it's worth though, SPLC is a political organization that's generally know for trying to create issues.

Just as an example, it looks like they have a store that sell confederate flags in Abbeville listed as a hate group...which seems like more than a little bit of a stretch. Looks like there's a flag store in Charleston also listed as a white nationalist group?

What was the local reaction to e.g. Black Lives Matter, or did the subject never come up? There definitely are communities where politics is a taboo subject, but I think it's increasingly rare these days.

Ironically, that's a large part of the reason that I don't read social media often or read much on the daily news. I tend to read after the fact analysis that provides more perspective instead of breathless reporting.

That tends to help me more in my job and having an informed perspective as a citizen.

Regardless of whether they do, the press thinks it's great - so long as it's aimed at the right targets. Remember Arguetron, the bot intentionally designed to start arguments? http://www.theverge.com/2016/10/7/13202794/arguetron-twitter...

I think that the most that Twitter needs to do is, if it can tell that an account is used by a bot, let other users know it's a bot account.

Some bots might say useful things on Twitter. Asking Twitter to decide what's acceptable for a bot to say is asking for trouble.

Reply on news.ycombinator.com