MNHQ here - got any data queries (and take a look at the work we’ve done with the Open Data Institute)

[JustineMumsnet]

Hello,

As some of you may have spotted, we recently worked with the Open Data Institute on a report about children’s lives during the pandemic - you can read the report here and some coverage of it in the press here.

The ODI is a non-profit organisation, whose purpose is to explore the value of data, and to advocate for the innovative use of data to affect positive change across the globe, working with companies and governments to build an open, trustworthy data ecosystem. I joined the ODI as a non-executive director in 2018.

The report uses aggregated, non-personal data from Mumsnet as well as from Barnardo’s and a few other sources to draw a picture of how the pandemic has affected children and families. The ODI also identified some ‘data gaps’ - areas where more data would help policy makers to make better decisions for children’s benefit.

Mumsnet’s data was used to understand patterns in the words and phrases MNers use to talk about how we are all dealing with the impact of schools closing and reopening, and how it’s affecting our mental health as well as our children’s mental health and wellbeing.

This was aggregated data that did not identify individuals. The ODI was given no access to users’ individual records, nor did they analyse posts at an individual level. You can see more information about how the ODI thought through and addressed ethical data issues on this project here.

The ODI analysed raw data from all of the forums across Mumsnet, searching by date, forum and search term. You can see some examples of the Mumsnet data accessed by the ODI here.

If you have any questions about this project or other internet data-related issues please post them here and either we or the Jeni Tennison from the ODI will be happy to respond.

Fire away!

Thanks

I would say the MN view on schools is very biased though, there is a level of hysteria online that doesn't reflect what the majority feel on a day to day basis. How do you prevent skewed results?

That's a good question. In general, data is always biased by how it's collected and who from, so in any analysis you have to recognise that and take it into account.

In the report, we highlighted the fact that the users of MN are not representative of all parents/carers, used other sources of data to cross-validate what the MN data was telling us, and were careful about the conclusions we drew from it.

We didn't highlight that what people post isn't necessarily what they feel day to day, but we should have: that's a really good point about the limitations of using data from social media platforms.