Dispatches from the front line of the internet

RSS

See if you can work out the major statistical flaw in this Telegraph piece on marriage

It's not that hard…

Photograph: Getty Images

It's bad stats Saturday, it seems. Before we start, see if you can work out for yourself what statistical stumble the Telegraph has made which renders this piece, headlined Almost no couples with children who stay unmarried stay together, study claims, entirely worthless. Go on, click through, we'll still be here.

Back? Hopefully you'll have been able to work out the main claim of the piece, which is that:

A study by the Marriage Foundation calculates that cohabiting couples who have children are more than twice as likely to split up as those who had tied the knot beforehand.

But of those who do not then go on to get married after having children, only a handful will still be together by the time the child is 16, it claims. [Emphasis mine]

It's a bold claim, and would certainly be a valuable fact for those who worry about the death of the "traditional" family to be able to point to. Unfortunately, it isn't true.

Further down the piece, we find the evidence behind the claim:

The report, which analyses figures from the Office for National Statistics, found that 93 per cent of couples whose relationships are still intact by the time their child is a teenager are married.

It calculated that out of a typical group of 100 16-year-olds, 45 of them would have experienced a family split, while 55 would still be living with both parents.

But only four of the 100 teenagers would have unmarried parents who are still together by the time they are 16. [Emphasis mine again]

In other words, of 100 sixteen-year-olds, only four have unmarried parents living together. That is, indeed, "only a handful" – but it is not a handful of "of those who do not go on to get married after having children". That figure is not given in the Telegraph report at all, and it's crucial. Without knowing what proportion of sixteen year olds were born to parents who were unmarried but living together, we can't know whether 4 per cent still living together is high, or low.

To find that out, we need to go to the original report, which claims that "out of the 47 per cent of children born to unmarried parents today… just 11 per cent will reach their 16th birthday with both parents intact and unmarried".

In other words, the survival rate of unmarried couples with children is over twice what the Telegraph implies; rather than 4 per cent, it is 11 per cent.

But there's something else as well. As the report says, "the rest will either marry or split up". In other words, a chunk of unmarried parents go on to marry before their child turns sixteen. That is also missed by the Telegraph's write up.

That is, where the paper writes that:

Only four of the 100 teenagers would have unmarried parents who are still together by the time they are 16.

It should actually read:

Only four of the 100 teenagers would have unmarried parents who are still together and still unmarried by the time they are 16.

And all of this stems from a report which is fundamentally based on assuming that patterns of marriage and cohabitation which were true for couples with children born in 1986 are still true for couples with children born in 2009. That's not something I'd stake money on; in case no-one noticed, we've had some pretty major changes to marriage recently.