Collecting sensitive data - a personal postscript

I have just published an article about collecting data - and some of the dangers of (even well-intentioned) data collection in creating unhelpful bias in statistics and the conclusions which can be drawn.

Today I experienced a small personal example of this, and I thought I would share it, to drive home the point.

[Also, it gives me a chance to post another baby photo. I do this because, based on analysis of (i) my academic posts and (ii) my "I've just become a father" post, readers of LinkedIn are about 140 times more interested in my daughter than what I have to say as a lawyer. But not all statistics are surprising.]

Earlier today I went with mum and baby to have a six week check up. When I arrived, we weighed and measured Olympia and chatted about nappies for about an hour. I'm delighted to say that everything was perfectly well. And then we came home.

Later that afternoon, my partner had a couple of missed phone calls from the Health Worker. She rang back to find out what might be wrong.

"There is a question that I need to ask you, that I couldn't this morning, because your husband was there."

"What's that?"

"Are you suffering any form of domestic violence at home?"

Now, I'm pleased to say that my partner is not suffering any form of domestic violence, from me or anyone else. And I can of course see why the Health Worker was mandated to leave that question out of the list of things to ask, since I was present at the check up.

But what is clear is that this was not an accidental element, for it was not the first time that my partner had been asked the question: she was asked on three occasions during the pregnancy. (Indeed, on one occasion I was asked to stand outside in the corridor for that part of the check up).

Let's just pause and think about data collection.

The first point is that apparently most pregnancy check ups are attended only by mothers - and there is a concern that, if fathers also turn up to these appointments, they are not there just to look after baby, or lend a hand (or still less that they are better with babies than the mum!) but rather that they are there to control the narrative and prevent the mother from being able to give a frank account of domestic life. That may of course be true. And if it is, it's a very sad thing indeed, which needs to be protected against.

But here is the other sad and dangerous thing: at no stage in the last 9 months did anyone ask me if I had experienced domestic violence or threat of harm at the hands of my partner.

Now, statistics (from the National Centre for Domestic Violence) indicate that about 75% of domestic violence in the UK is perpetrated by men against women. But that means that about 25% of domestic violence in the UK is committed by women against men. And since the Health Workers are mandated to ask mums about whether the male partner is being violent, but not the other way round, it means that domestic violence by women against men is not being captured in the resulting data at all.

Nor is there any implication, apparently, that a mum who takes a child to a six week check-up without the father being present, is somehow seeking to control the narrative about domestic life, by excluding his voice.

So we will continue to see articles written about the misery of domestic violence, based upon the available statistical data. But in the case of the data collected during pregnancy, at least, that data will be fatally biased by deliberately excluding half of the relevant population.

That is why we need to be so careful. There is not always safety in numbers.