Thread by @MForstater

UCU's bar chart is a crime-against-data-visualisation, which buries what its own survey data shows:

80% of respondents to their survey who said they had direct experience of sexual violence at work were women.

What sex were the perpetrators?

UCU didn't bother asking.

https://twitter.com/ucu/status/1473606163610624003

Bar charts like these with percentages and fixed overall lengths are usually read as 100% bar charts - each coloured slab is a proportion of the whole.

But in UCU's the percentages don't add up, because each one is a percentage of something else.

Its misdirection

It is a self selected survey (likely to attract people concerned about sexual violence and more likely to have been victims).

66% of respondents were women, 32% men, 2% trans/NB

And out of those who said they had experienced sexual violence 80% women, 16% men, 4% trans/NB

UCU has about 120,000 members & just under 4,000 filled in the survey.

UCU's headline about "risk" is nonsensical. Its a self selected survey - it can't tell you about the people who didn't fill it in.

I don't know what the P values at the bottom are supposed to mean 🤨

14 individuals identifying as trans/NB said they had experienced sexual violence

UCU's No. 1 recommendation is about terminology "gender based violence"

Nothing about the fact that women are the predominant victims

Nothing about the predominant sex of the perpetrators at all