Personal relationships / Marriage matters?

Marriage matters?

Discussion of the social significance of marriage rarely leaves the headlines. This is particularly true when the topic concerns children, as shown by the long-running debates about whether or not parents' choice to cohabit rather than marry has a negative impact on their children's social and developmental outcomes (Goodman and Greaves, 2010). Most recently, the passage of the Marriage (Same-Sex Couples) Bill through the House of Commons and subsequently through the House of Lords in 2013 attracted ferocious debates among both Conservative MPs and the party faithful more generally. Many opponents couched their opposition to the Bill by reference to the 'sanctity' of heterosexual marriage, the union between a man and woman that has long been the social, legal and religious norm - and for many the ideal - when it comes to sex and parenthood. 

undefinedDespite these debates, the last 30 years have seen huge changes in Britain's marital behaviour, with an increasing proportion of people either delaying getting married or not marrying at all. Between 1983 and 2010 the marriage rate in England and Wales (which is the number of marriages among every 1000 unmarried men and women aged 16 and over) more than halved, from 52 to 22 (among men) and 42 to 20 (among women) (Office for National Statistics, 2012a). This partly reflects an increasing tendency for couples to cohabit, either as a precursor to, or instead of, marriage. When the British Social Attitudes survey began in 1983, the majority of couples did not live together before tying the knot; this applied to only a minority, around three in ten. Now, it is those who get married without living together first who are unusual. In 2004-2007 around eight in ten first-time married couples lived together first (Beaujouan and Ni Bhrolchain, 2011). As a result, cohabitation rates have increased considerably; between 1996 and 2012 the number of cohabiting heterosexual couples increased from 1.5 to 2.9 million, and the number of dependent children living in these households doubled, from 0.9 to 1.8 million (Office for National Statistics, 2012b).

It would be surprising if such major societal shifts were not accompanied by fundamental changes in the way that we think about marriage and its role. In this section we explore this by focusing on two issues: the acceptability of sex outside marriage, and views about whether marriage and parenthood should go hand in hand. We begin with attitudes to premarital sex. To assess this we ask the following question:

If a man and woman have sexual relations before marriage, what would your general opinion be?

In 1983, 28 per cent thought such premarital sex was "always" or "mostly wrong"; since then, the figure has now more than halved, to 11 per cent. In the meantime, the proportion who think sex before marriage is "not wrong at all" has increased markedly, from 42 to 65 per cent (a further 10 per cent think it is "rarely wrong"). 


We also ask participants their views about the acceptability of extramarital sex, defined as "a married person having sexual relations with someone other than his or her partner". Here there has been very little change over the last 30 years; with the vast majority, consistently eight in ten or more (84 per cent in 2012), seeing this kind of behaviour as always or mostly wrong. So, while sex outside marriage is acceptable to the majority of the British public, there is a continued commitment to marital fidelity - that people who are married should be faithful to their partner.  

Traditionally one of marriage's key functions has been parenthood. Earlier we saw that the proportion of children being born outside marriage has increased sharply since the early 1980s and, to some extent, this behaviour change is reflected in trends in attitudes. One of the ways in which we assess this is by asking people whether they agree or disagree with the following statement:

People who want children ought to get married

As Table 1.3 shows, when we first asked the question in 1989, seven in ten (70 per cent) people agreed with this view (with 25 per cent agreeing strongly) while less than two in ten (17 per cent) disagreed. Over the last few decades opinions have shifted considerably; now 42 per cent agree (nine per cent strongly) while around a third (34 per cent) disagree. Notably, the proportion of people who opt for the middle ground of neither agree nor disagree has increased too, from one in ten in 1989 to more than two in ten now. 


undefinedThis change marks a shift in a more liberal direction, but it is clear that opinion is more evenly divided than was the case for premarital sex. Indeed, the most prevalent view remains that marriage should precede parenthood. This more nuanced view about marriage once children enter the equation has been noted before (Duncan and Phillips, 2008) and is evident in responses to other questions included on the survey. For instance, between 1994 and 2006 we asked people to respond to the statement "one parent can bring up a child as well as two parents together". In 1994, just over a third (35 per cent) agreed with this view, while 46 per cent disagreed; by 2006 there had been a slight shift, with agreement going up to 39 per cent and disagreement down to 40 per cent; nevertheless, opinion remained divided on the issue.  

So far we have seen that, although marriage continues to be seen as a prerequisite for sex only by a small minority, more doubts remain about moving away from the traditional family when it comes to bringing up children. We turn now to consider what best accounts for these changes over time, trying where possible to identify the relative importance of generational differences, religion and education. 

We start by looking at generational differences. In the case of premarital sex, if we simply look at attitudes by generation, there is now less clear-cut variation by age than in the past, even though in earlier years younger people were notably more liberal than their elders. The closing of what was once a huge generation gap is illustrated in Figure 1.2. It shows that in 1983 there were considerable generation gaps in people's views about premarital sex, with differences of over 10 percentage points between the views of those born in each of the five decades between 1900 and 1950. This contrasts clearly with the generations born in the 1950s and 1960s, who had very similar views to one another. This tendency for each new generation to have similar views to its predecessor has continued since then, as illustrated by the fact that the lines for the generations born after 1950 are very close to one another. 

Figure 1.2 also allows us to follow a particular generation and look at how their views have changed. We can focus, for example, on those who were born in the 1950s and compare their views in 1983 (when they were in their late 20s and early 30s) with their subsequent views in 2012 (when they were aged between 53 and 62). This allows us to see whether a generation has become more liberal as it has aged - which in this case would be indicated by a downward line in the graph (as happened for instance for the 1930s and 1940s generations between the late 1990s and 2003). 

These trends mean that the gulf that existed in 1983 between the views of different generations has shrunk dramatically. In 1983, just six per cent of the generation born in the 1950s thought premarital sex was always or mostly wrong, compared with 59 per cent of the group born between 1900 and 1909, a huge gap of 53 percentage points. Now the gap between our oldest generation (born in the 1930s) and the youngest (born in the 1980s) stands at just nine percentage points (with 22 and 13 per cent respectively thinking that premarital sex is wrong). It is these sharp generational differences that account for the large shift in public opinion we have seen over the last 30 years, as older, less liberal, generations have died out and been gradually replaced by younger, more liberal, ones. 


undefinedA different pattern emerges when we look at attitudes to having children outside marriage. Here distinct generational differences have persisted over time; indeed, they are just as marked in 2012 as they were in 1983. Take, for instance, the results for each generation in 2012 in Figure 1.3. They show that 28 per cent of the 1980s generation think people should get married before having children, rising to 31 per cent among the 1970s generation, 34 per cent of the 1960s generation and so on, until we hit highs of 62 and 84 per cent respectively among those born in the 1940s and 1930s. Although looking at earlier years it is clear that most generations have become slightly more liberal on this issue over time (indicated by a downward slope on the graph), the gap between the views of old and young is now actually wider than it has ever been. Note too that, as with attitudes to premarital sex, the three youngest generations are far closer together in their views than previous generations are to one another. These findings suggest that we will continue to see attitudes in this area become more liberal over time, as older generations die out, perhaps slowing at the point when the 1960s generation start to become the elder statesmen and women among Britain's generations (that is, from the 2040s onwards). 


Earlier, we set out why we would expect to find a link between religious faith and people's attitudes to personal relationships. This indeed proves to be the case for attitudes to marriage (Table 1.4). When it comes to attitudes to premarital sex we find non-Christians at one end of the spectrum (although the small sample sizes involved mean these figures should be treated with caution). This group stands out as the most disapproving of premarital sex, with just over half in 2012 thinking it is always or mostly wrong. Those affiliating to other religions are more tolerant; around one in ten Anglicans and Catholics think that sex before marriage is wrong, a view shared with one in five of those belonging to other Christian religions. The most tolerant of all are the non-religious; out of the 500 odd people we interviewed who defined themselves in that category, two per cent (10 people) said they thought that premarital sex was wrong.

undefinedAll religious groups, with the exception of non-Christians, have become more accepting of premarital sex over the last 30 years. Among Anglicans, for instance, the proportion thinking premarital sex is wrong is now a third of what it was in 1983 (10 and 31 per cent respectively). Among non-Christians opinions are now less tolerant than they were, bearing in mind the caveats mentioned earlier with regard to small sample sizes.


When it comes to the acceptability of parenthood outside marriage there is far more similarity of view between those from different religious faiths (Table 1.5). The key difference here is between those who are religious and those who are not. For instance, over half of Anglicans (54 per cent) agree with the view that people should get married before having children, compared with just 30 per cent of the non-religious.  

This link between attitudes and religion offers at least a partial explanation for the generational changes we saw earlier, as older generations are far more likely than younger ones to be religious. However, Britain's increasing liberalism about premarital sex and parenthood outside marriage cannot be solely put down to the fact that religious adherence has fallen over time, as this does not account for the fact that most religious groups have themselves become more accepting over the last 30 years. 

undefinedEducation is no longer strongly linked to a person's views about premarital sex; although in the 1980s graduates were more liberal than other groups on this matter, the views of all groups defined by educational qualification have become steadily more liberal over time. However, the picture is less clear-cut when it comes to attitudes to parenthood and marriage. Here the most liberal views of all are held by those whose highest qualification is a school-based one (that is, A levels or GCSEs, or their equivalent). Among those for whom A levels are their highest qualification, 35 per cent think people should get married before having children, compared with 41 per cent of graduates (the next most tolerant group) and 55 per cent of those without any qualifications (the least tolerant). On this issue then, the growth in the number of graduates does not help us account for increasingly liberal views about sex and marriage. 


In summary, views about marriage have become more liberal over time. When it comes to sex outside marriage, there is a considerable unanimity of opinion, with only religion and, to some extent, generation still being clearly linked to differences in views now. In contrast, there remain marked differences between the views of different groups on marriage and parenthood, with age, religion and education remaining clear markers of a person's views on the subject. The shifts we have seen over time are mainly accounted for by generational change; although declining religious faith is in itself linked to generational differences, increasing secularisation is an insufficient explanation of the changes we have found. Education, although partly linked to a person's views on marriage and parenthood, is not an important factor in explaining the huge changes we have seen over time. Although it is hard to find clear evidence of the way in which period effects have influenced people's attitudes, the close correlation between marriage behaviour and attitudes suggests that, rather than having a simple causal relationship, the two are influenced by one another; so attitudes will help shape behaviour, and behaviour (or exposure to behaviour) will in turn shape attitudes. 

Of course attitudes to personal relationships may well be influenced by other factors beyond those we have considered here. Our Social class chapter examines attitudes to premarital sex in the context of its investigation into the changing role of social class in shaping people's attitudes. Multivariate analysis there shows that age, ethnicity and church attendance (in 1984 and 2012) and gender (in 2012) are all highly significant predictors of a person's views about premarital sex, even when a range of other measures are taken into account. The effect of age on attitudes appears to have weakened over time, while the effect of church attendance has strengthened. These trends support the findings we have described above, which show how the views of different religious groups have diverged over time, while the attitudes of different generations have become less distinct.   

undefinedWe finally examine the extent to which party affiliation is linked to people's views, to see what light this sheds on current political debates about marriage and the family (Table 1.6). To do this we use a series of questions that, together, measure the extent to which people readily identify or not with a particular party, their 'party identification' (see our Politics chapter for more details).[5] While there are no such differences associated with attitudes to premarital sex, there are clear differences between the views of supporters of the main political parties when it comes to marriage and parenthood. Conservative Party supporters are the most likely to think people who want to have children should get married; over six in ten think this, far higher than the rates found among Liberal Democrats and Labour supporters (three in ten and four in ten respectively). The most liberal of all are those who do not support any party, no doubt reflecting their younger than average age profile. 


Conservative supporters also stand out as having changed the least in their views, although the make-up of those supporting a particular party will of course have changed over the years. While the proportion who think that people should get married before having children has fallen by 15 percentage points between 1989 and 2012, this compares with a drop of 23 percentage points among Labour supporters. As a result, Conservative Party supporters have diverged from supporters of other political parties in their views on this matter over time. Indeed, they are now nearly three times more likely than those who do not support any party to think that parents should be married. David Cameron's decision to balance his social liberalism on issues such as gay rights with a continued emphasis on the importance of marriage is very much in line with the views of the Conservative Party faithful; it does however beg the question as to where new party support might come from, particularly given the trends towards increased liberalism on family issues that are likely to continue into the near future. 

Download chapter
  1. BBC news, 17th February 2001, available at:
  2. It is possible that the expansion of higher education will affect the relationship between degree-level education and social values, especially if the main mechanism by which education affects attitudes is socialisation (rather than cognitive development). So as a wider cross-section of young people attend university, the distinctive nature of their values is diluted.
  3. Bases for Table 1.4 are as follows:
  4. Bases for Table 1.5 are as follows: 
  5. In summary, there has been a considerable rise since 1983 in the proportion who identify with no political party whatsoever, up from eight per cent in 1983 to 21 per cent now. The proportion of Conservative identifiers has shrunk (from 39 to 27 per cent) and the proportion of Labour identifiers has remained broadly constant (33 and 36 per cent in 1983 and 2012 respectively). In 1983, 15 per cent of people identified with the Liberal/SDP Alliance, compared with six per cent in 2012 identifying with the Liberal Democrats. Further details can be found in the Politics chapter.  
  6. Bases for Table 1.6 are as follows:
  7. Speech by Margaret Thatcher to the Conservative Party Conference, 1987, 9th October, available at:
  8. See, for example, the ILGA Europe review of Lesbian, Gay, Bisexual, Trans and intersex people in Europe, available at:
  9. Bases for Table 1.8 are as follows:
  10. The Guardian Datablog, 7th October 2012, available at: