Harry Benson is director of Bristol Community Family Trust and the author of Let's Stick Together - The relationship book for new parents.
A new study by the IFS out this morning claims that it’s not being married that influences whether couples stay together or not. It’s the kind of people who marry. The usual suspects have got very excited about this – the Guardian, the Family and Parenting Institute.
Their excitement is premature. The technical analysis in the IFS paper may be really good. But the conclusion is really wrong.
The new IFS study replicates two studies I did (1, 2) based on the Millennium Cohort dataset of 15,000 mums who had babies in the year 2000 and 2001. Because IFS use the same dataset, it’s not surprising that their main findings are identical to my own. In essence they found that if you compare two groups of 100 cohabiting and 100 married parents, 26 of the cohabiting couples split up before their child’s fifth birthday compared to 9 of the married couples. On the surface at least, the extra 17 couples represents a big and highly significant gap in outcomes.
By knocking out various background factors one by one in a sequence of models (page 10 of their paper), they have then shown how much each factor contributes to this gap along the way. The way they have done this is really neat. I found it especially interesting, for example, that having parents who split up or being a stepparent each influence the odds of staying together less than I might have expected. But the key thing is that by comparing groups with similar backgrounds – mother’s age, ethnic group, family background and education, and father’s income – they show that the gap narrows to 7 more couples in the cohabiting group. Up to this point, their analysis concurs with mine. Amongst all the major background factors, marital status has the single biggest unique effect. So far, so good.
It’s at this point that the IFS researchers go astray. The final two factors they eliminate are whether the baby was planned or not and any differences in relationship quality between the parents. That reduces the gap enough for the researchers to show that marriage and cohabitation no longer have any residual influence on stability. Hence the excited story in the papers.
The problem here is a circular argument. To reach their conclusion, the IFS researchers also have to start with their conclusion, that the act of marriage is simply a rubber stamp that makes no other difference. It’s almost as if they can’t really understand why people marry in the first place. It is inconceivable that getting married could have an effect on stability but somehow not on the way couples relate and make their decisions about the future. Marriage and relationships go together. Had they eliminated marriage in their model at this point, they would doubtless have found that relationship quality has no residual effect on stability. Would the Guardian have published a story that says relationships don’t matter?
The cause and selection argument over marriage can never be completely resolved because it is impossible to design a study that assigns couples to marriage or cohabitation and then see what happens. However the selectionists need to consider two issues that challenge their circular argument.
The first is new research showing that men’s decisions about the future appear strongly linked to their subsequent commitment. This is much less true for women whose commitment appears to be based more on attachment. The theory and research on this is both fascinating and robust (read about sliding/deciding/inertia in this paper or in my book). And it has the air of plausibility about it. If you encourage men to get married, the decision this requires about their future will almost inevitably influence the way they behave in other ways. It is equally plausible that if you discourage women from cohabiting, fewer women will find themselves prematurely entangled in a less than ideal relationship from which it is now harder to exit. That’s an aspect of cohabitation that is rarely considered.
The second is that the major background socio-economic factors all suggest family stability should have improved in recent years. Older, richer and better educated parents should be more likely to stay together, not less. Yet family breakdown has risen relentlessly. Since 1980, the whole of this rise coincides with a shift away from marriage towards cohabitation while divorce rates have remained unchanged. If marital status is not the issue here, the IFS researchers are effectively saying we are becoming worse at holding down relationships. Patronising stuff. A more plausible explanation is that we have moved away from principles that protect and support relationship stability.
This new research by IFS does not challenge the belief that marriage keeps couples together. If anything, the findings affirm marriage far more than question it.
