Active discussions

Please or to access all these features

Talk
Site stuff

Join our Innovation Panel to try new features early and help make Mumsnet better.

See all MNHQ comments on this thread

MNHQ

Heads up: some academic research taking place

664 replies

RowanMumsnet · 30/08/2018 16:53

Hello

Some researchers from a UK university are going to be collecting posts from some boards on Mumsnet over the next few weeks. They will be looking only at posts published from this point onwards. Selected posts will be anonymised (ie usernames will not be stored anywhere), and all processes will be GDPR-compliant.

EDITED BY MNHQ: to be clear, NO private or back-end data is being shared with the researchers and they will have NO access to any extra information about users. They will only be analysing published posts, ie things that are already in the public domain.

The researchers will analyse the text of the anonymised posts as part of their study. The full text of these anonymised posts will not be reproduced in the study - no pieces of text that make you or your situation identifiable will be published.

Once the study is complete, the database containing the anonymised posts will be destroyed.

Obviously we'd love to tell you what the study is about - but we can't because it will skew the results. However once the research has been completed we'll share it with you and you can read all about it. The academics running the study are people we've worked with before, and the topic is one we think lots of MN users will agree is interesting and useful.

IMPORTANT BIT: if you DO NOT want your posts to be considered for inclusion in the research, please shout here on this thread or email us on [email protected]

And of course let us know if you have any questions and we'll see if we can answer them.

Thanks

MNHQ

OP posts:
TheFallenMadonna · 31/08/2018 08:40

There are other, qualitative, ways of analysing discourse.

CoteDAzur · 31/08/2018 08:37

"You are assuming they are doing frequency analysis. Why?"

If you actually read Rowan's posts, you would see that I don't assume but know:

it will be something along the lines of 'Parents discussing child carseat choices tend to use this sort of language and these particular words figure prominently, whereas when they're discussing family meal choices the language they use has these differences.'

CoteDAzur · 31/08/2018 08:33

"Without usernames, what extra info could they possibly find useful? All those things (IPaddress, location) are linked to usernames. Which they are not collecting."

IP addresses - To rule out those of us who live outside the UK, for example. (Many are not native speakers so perhaps should not be included in language & word frequency analysis, others may be Brits who lived there for many years which would usually mean a difference in some daily vocabulary)

Name change history - As I explained in the post above, this would be useful if not essential in determining which words are actually used frequently and which are just the same person posting very often with the same words.

sansouci · 31/08/2018 08:33

Could it be... an MNHQ troll??

Cockapoomummy · 31/08/2018 08:32

I’m not. I’m analysing the statement we have been given.

TheFallenMadonna · 31/08/2018 08:31

You are assuming they are doing frequency analysis. Why?

Cockapoomummy · 31/08/2018 08:31

It says nothing that identifies us will be published.

It doesn’t say the researchers won’t have access to the full info held about us to do the research.

GONNAENo · 31/08/2018 08:30

Opting out, thanks. After the intern thing I'm not sure why I'm still here to be honest.

Cockapoomummy · 31/08/2018 08:29

It says part of. Not only. We don’t know what the rest of it is.

It says “collecting posts from some boards” - the researchers are doing this. So what way is this being done ? Will they collect all posts and then run them against an exclusion list? And exclude anyone who has opted,out?

Anonymisation is happening AFTER the posts have been selected by the researchers.

CoteDAzur · 31/08/2018 08:29

"OP says they are not collecting usernames, so lists of previous usernames would be pointless."

You don't understand.

Frequency analysis would only make sense if they know namechanges, so that they can normalise the effect of a very prolific poster who has posted under several names in the same topic, for example

TheCatFromOuterSpace · 31/08/2018 08:28

I'd like to opt out thanks

I would only give my consent to take part in research if I understood the aims of the research, and the potential risks and benefits involved.

TheFallenMadonna · 31/08/2018 08:26

Without usernames, what extra info could they possibly find useful? All those things (IPaddress, location) are linked to usernames. Which they are not collecting.

CoteDAzur · 31/08/2018 08:25

Yes, anybody can do a word frequency analysis on MN posts now, without asking for permission.

So what is the need for informed consent really about?

@MNHQ Are you providing these researchers with any information about MNers?

Cockapoomummy · 31/08/2018 08:24

The researchers will analyse the text of the anonymised posts as part of their study. The full text of these anonymised posts will not be reproduced in the study - no pieces of text that make you or your situation identifiable will be published.

Part of. Not all of. We don’t know what in the back office they are being given access to.

TheFallenMadonna · 31/08/2018 08:24

Why would that be useful if they are not collecting usernames?

TheFallenMadonna · 31/08/2018 08:23

The OP says they are not collecting usernames, so lists of previous usernames would be pointless.

Cockapoomummy · 31/08/2018 08:23

Absolutely anybody could, but we don’t know if they have access to any back office data such as age, general location, IP address, email address, nor do we know if that is a necessary part of the research and is what is being done.

TheFallenMadonna · 31/08/2018 08:22

They could do that now. Anybody could.

TheFallenMadonna · 31/08/2018 08:21

Which people have posted on an open, searchable forum, and which will be used with less information than they provide themselves on that open, searchable forum, and the published results will not even contain the full posts.

Cockapoomummy · 31/08/2018 08:19

What steps are in place to ensure that the researchers done crawl through all the posts by a particular user name and use software to jigsaw info together to find RL identity?

BadderWolf · 31/08/2018 08:19

Please do not include my posts in this research, thank you.

sansouci · 31/08/2018 08:16

Yeah well, whatever. I'm sleepy!

CoteDAzur · 31/08/2018 08:15

@RowanMumsnet Will you give the researchers a list of past usernames?

If not, they will have no way of knowing if what they see as "this sort of language and these high-frequency terms" are just the same 2 people posting under various names.

Cockapoomummy · 31/08/2018 08:14

Because personal data is not just usernames. According to the ICO it is any data that can be used to identify a living person.

TheFallenMadonna · 31/08/2018 08:13

If they are not even collecting usernames, how can it not be GDPR compliant?

Swipe left for the next trending thread