Mumsnet Corpus

[TokyoBouncyBall]

Not a TAAT, but a bit of googling as a result of a now deleted thread has led me to this:

https://fold.aston.ac.uk/handle/123456789/18

I note it says that the License is uncertain. Can you confirm that you have given permission for posts to be used in this way, or is there something that Aston might like to look into?

I note it says Users who wish to access this dataset must make a detailed application to FoLD and the researcher, as well as potentially gain additional agreement from an external organisation before they can be approved for access.

Given one of the uses it is being put to, I think it is a bit dubious to say the least.

Have they accessed our real names/emails/locations too?

It's unlikely that they can get that information from Mumsnet unless it was posted on a thread by the poster. See Justine's post.

Even on occasions where MN have given permission for a university to use the data we provide, it seems really inappropriate for the university to approve women's data to be used for a research project that demonstrates such a huge misogynist bias. It implies pretty much anyone at Aston could get access to the data if they want it.

(Imagine a university ethics board backing an openly racist PhD using data from Lipstick Alley, or a an openly homophobic PhD using data from gay Reddit forums).

Why aren't Aston encouraging their researchers to deal with their unconscious bias before having access to data created by a marginalised subset of society?

We need to know all the reports that have been created using our data and have access to them and they need to delete the damn database.

Can you assure us you're not going to seek to make a profit financial settlement from this but will endeavour to seek the best outcome for your site users?

Why aren't Aston encouraging their researchers to deal with their unconscious bias before having access to data created by a marginalised subset of society?

I'd like to know that too. I imagine we can complain to the ethics department of Aston?

CDA approach to language analysis is fundamentally problematic due to the significant impact of subjective decisions made by the researchers.

This isn't just about pro/anti/dontcare on trans issues, it applies to any model built where subjective researchers start with a strong opinion on the subject. Methodologically the risk is saying "I define X, I say X is bad, this is X, now use that pattern to find more bad, oh look I have found bad". The whole approach of CDA leans to be politicised and in this particular case there are clear and strong pre existing views - its hard to see the credibility. Using this approach to build language models to sell or licence for NLP use is lucrative but against most Responsible AI guidelines and in areas which are politically or socially controversial they can be particularly discriminatory. Its a real challenge in building models, even for sentiment analysis in this type of space.

The issues of Author Authenticity and potential for abuse I think is already covered by the thread.

Reposting:

you have people blockquoting other posters prior to the Quote button being introduced, so you can't assume that a given post contains only material written by that poster.

the "recentsy primer" (my spelling might be wrong) [this refers to something mentioned in the 2019 talk video] might well be an artifact of posters quoting PPs. It's the most plausible explanation I can think of to explain a user adopting a phrase for only one thread and never using it again.

Pro-tip, Aston researchers: when A quotes B, that's B's words, not A's, not even "A's for the duration of the thread".

My per diem for tearing holes in your research is £1000 to Women's Aid.

Did they scrape the MN disability boards?

I think they scraped the whole site, for their "sandbox" rather than particular boards. We know they scraped posts about fertility treatment and dieting.

Probably yes, but I would like @JustineMumsnet to ask so that he is aware.

The disability boartds are slightly protected, they don't appear in trending etc.

Given the obvious anger and uncertainty this has caused many MN members could I suggest you also take a MN user with you? I would propose @ArabellaScott or @Boiledbeetle. As a user I feel it is important that the positionality of this research warrants assurances that the topic area that this research has crashed into is appropriately scrutinised.

I mention above as they are the most well versed in this area. I'm sure there are others.

@JustineMumsnet ? What are your thoughts?

Also apologies to @Boiledbeetle and @ArabellaScott for throwing you the lit branch which you may not want?

I would really want to understand how the system works. What is being scraped, and how, who owns the database, how do they decide who will get to use it.

Ask them if you can access the database.

It does rather sound like they're hoping this will go away and they can carry on after the call....

Probably yes, but I would like @JustineMumsnet to ask so that he is aware.

Absolutely.

Yeah this.

Will await results of convo, obvs.

But VC wants "to explain the reasons for the research, how they manage ethical approval and protect privacy and data" sounds like, "We did something illegal, we're not sorry, and we're going to yap at you until you give up and go away".

But VC wants "to explain the reasons for the research, how they manage ethical approval and protect privacy and data" sounds like, "We did something illegal, we're not sorry, and we're going to yap at you until you give up and go away".

It does.

I'm getting flashbacks to dealing with the builders who fucked up my house.

Fucking A, Mumsnet!

I’d also be interested in Aston’s (no doubt highly comprehensive) training for researchers dealing with data…..

I'm not sure we are going to go away, whatever the results of this conversation. The research exemption to explicit consent in the DPA 2018 hinges on the research being ethical and not liable to cause distress.

https://ico.org.uk/media/for-organisations/documents/1061/anonymisation-code.pdf

Chapter 9 and the appendices deal with research.

Yep. The VC has probably been advised by their lawyers not to admit to anything in writing.

If there's one thing Aston has learnt, it's not to piss off the vipers found on the feminism board.

I'm really looking forward to someone's PhD thesis on the subject.

• correction of technical points

Separately to the issue of whether Aston inadvertently re-identify previously pseudonymised data,

I would just like to add that in order to process special category data under UKGDPR any institution needs to have a lawful basis AND conditions for processing.

I expect Aston will say their lawful basis is “public task” Article 6(e) condition for processing is under Article 9(2)(e)- “made public by the data subject”.

However, it is worth noting that Mumsnet or any of the subjects can still challenge this on the basis that it is fair.

I would argue it is not fair because the research title makes the assumption that the discourse is transphobic.

Additionally, they need to consider the subject’s reasonable expectations for further use of the data.

They like MN language, so I hope Justine offers them the classic specimen of "No", as a complete sentence...

All their data, once published, will be identifiable if the post is extremely unique. All people have to do is paste the text into Google and they've found the poster.

Hasn't Eden Palmer already explained the reason for the research? To study "transphobic hate crimes on Mumsnet" as she said in her now deleted LinkedIn post. Not just transphobia, "transphobic hate crimes".

I would still like them to confirm it. The only definitive way they can test their theories is by having our real data. Since they scraped the site illegally, and they are computer bods and people that didn’t like working at mumsnet have been quoted in this thread, it seems a reasonable question to ask.