Ignorance in Bliss: Modeling Knowledge of Rights in Marriage and Cohabitation

Abstract

We live our lives against an extensive backdrop of legal rights and responsibilities, yet a growing number of studies indicates low levels of public legal literacy. In the context of opposite-sex cohabitation and marriage law, this study employs new survey data from the United Kingdom to explore, in detail, how many and which people are ignorant of the law, and what are the nature and origins of erroneous beliefs. We find that people's beliefs about both cohabitation and marriage law are frequently wrong. They are also strikingly similar, and reflect the divergence of social attitudes from the law. Our findings are consistent with the notion that legal literacy links to salience of issue. They are also consistent with recent public legal education initiatives that affected public understanding of cohabitation law, but we argue that social attitudes and the intransigence of erroneous beliefs generally present significant challenges to such initiatives.

We live in a “law-thick” world (Hadfield 2010: 133), in which our lives are played out against an increasingly complex and extensive backdrop of legal rights and responsibilities. Yet, as Williams (2009: 734) has commented, it appears that “in some contexts, the vast majority of the population systematically mispredicts … the content of the law.”

Over the past half century, a steadily growing number of studies of the public's understanding of law has pointed to a substantial knowledge deficit (e.g., Cortese 1966; Williams & Hall 1972; Saunders 1975; Baker & Emery 1993; Darley et al. 1996, 2001; Kim 1999; Barlow et al. 2005; Militello et al. 2009; Parle 2009). The deficit appears greater in some areas of law than in others, and this is, at least in part, a function of salience (Saunders 1981). There is, after all, little reason for individuals to possess knowledge that has no bearing on their lives. So, for example, Casebourne et al.'s (2006) study of employee awareness of employment rights in the United Kingdom demonstrates that those with dependent children were “understandably” more likely than others to know a lot or a fair amount about the detail of the parental right to request “flexible working” (40 percent versus 27 percent). There is also less reason for people to have a grasp of the intricacies of the law than to understand broader legal principles, and again this is reflected in recent findings (Casebourne et al. 2006). Nevertheless, it remains the case that individuals are frequently ignorant of the law as it relates to their daily lives and futures.

Particularly in those areas of life in which the law can play a central role in people's well-being and prospects, it is therefore important to understand how many and which people are ignorant of the law, what are the nature and origins of erroneous beliefs, and what are the implications of public ignorance of law.

In recent years in the United Kingdom there has been particular interest in, and concern about, public ignorance of the law pertaining to opposite-sex cohabitation. This is an area of law of great significance for many, yet it is commonly misunderstood.

Cohabitation in the United Kingdom

Recent decades have seen rapid change in family forms, with opposite-sex cohabitation “skyrocketing” (Bowman 2010: 93) across the Western world, largely at the expense of marriage (e.g., Cherlin 2010; Kiernan 2004).¹ In the United Kingdom, opposite-sex cohabitants (hereafter “cohabitants”) now make up more than 15 percent of all families, a rise from 12.5 percent in 2001 (Beaumont 2011), itself a substantial increase on earlier decades (Shaw & Haskey 1999). More than 13 percent of dependent children now live within cohabiting couple families (Beaumont 2011).

In the United Kingdom such cohabitation has been described as a “classless phenomenon” (Law Commission 2006: 33), with “the social class differences between cohabiting and married adults … not large” (Wasoff et al. 2010: 10). However, cohabitants are distinctly younger than those who are married and, in part as a consequence of this, less affluent and more likely to live in disadvantaged areas (Law Commission 2006; Miles et al. 2009; Wasoff, Miles, & Mordaunt 2010).²

Nevertheless, when looking only at families with dependent children, data from the Millennium Cohort Study indicate that cohabitant parents are more likely than their married counterparts to be white, to have low incomes, to receive welfare benefits, to live in social housing, and to have few educational qualifications (Kiernan et al. 2011; Kiernan & Smith 2003). These differences, though, are less marked than those observed in the United States, and, in the case of ethnicity, they have a different character (Kiernan et al. 2011).³

Several studies have also suggested that United Kingdom cohabitants' relationships are somewhat more likely to dissolve than those of married persons (Law Commission 2006). Kiernan (2006), for example, found that while 95 percent of mothers married to a child's father at birth were still living with the father three years later, the figure was 83 percent in the case of cohabiting mothers. There is some evidence, though, that the difference reduces when data are age standardized (Miles, Pleasence, & Balmer 2009). In any event, the presence of children appears to promote relationship stability, as the relationships of cohabitants and spouses who do not have dependent children are more likely to dissolve (Miles, Pleasence, & Balmer 2009).

Again, differences in relationship stability between cohabitants and spouses are less marked in the United Kingdom than in the United States (Cherlin 2010; Kiernan et al. 2011). This reflects evidence that levels of commitment are generally high among cohabitants in the United Kingdom (e.g., Barlow et al. 2005), with relationships seemingly more often “marriage like” than those in the United States (Kiernan et al. 2011: 11), where there may be more “slide” or “drift” (Manning & Smock 2005: 1000) into and out of cohabitation.⁴

The Common Law Marriage Myth

Despite the rapid growth in the numbers of people cohabiting, there is evidence of substantial public confusion about the legal rights and responsibilities of cohabitants (Barlow & Duncan 2000; Barlow et al. 2005, 2008; Hibbs 2001; Panades et al. 2007). Drawing on data from the most recent British Social Attitudes Survey (BSAS) to address the issue, Barlow et al. (2008) have shown that a majority of the English and Welsh public believes, erroneously, that unmarried cohabiting couples accrue, over time, similar legal rights and responsibilities to those enjoyed by married couples. This belief has come to be known as the “common law marriage myth.”⁵

If cohabitants were to be immune to the common law marriage myth, then perhaps it would not have become such a cause for policy concern (Law Commission 2006, 2007a). However, Barlow et al.'s findings (2008) also suggest that the myth is subscribed to by a majority of those who cohabit, as well as a majority of those who do not. The 2006 BSAS indicated that “the same proportion of cohabitants as married people still falsely believe that common law marriage exists (53 percent in both cases)” (Barlow et al. 2008: 42).⁶

Although (as we discuss in more detail below) the BSAS questions are ill suited to determining the accuracy of people's beliefs about the substance of cohabitants' rights—as there is a presumption that people have good knowledge of the spousal rights to which cohabitants' rights are being compared—Barlow et al.'s (2005, 2008) findings nevertheless give cause for concern that many people, including many cohabitants, wrongly believe that cohabitants somehow acquire extensive and distinct relationship-based rights over time.

Of course, people who broadly subscribe to the common law marriage myth do not all subscribe to a single set of erroneous beliefs. Barlow et al. (2005) identify two distinct types of beliefs to which people may subscribe, in varying degrees and independently of one another. One concerns rights between cohabitants; the other concerns rights and responsibilities involving children. One is tied to relationship duration, the other to biological parenthood.

Sitting at the heart of the “time eligibility” (Barlow et al. 2005: 41) form of the common law marriage myth are beliefs that cohabitants can acquire, once they have lived together for some period of time, equivalent rights to spouses to their partners' assets and to financial support in the event of relationship breakdown or death. Mistaken beliefs along these lines are less common among the public than are mistaken beliefs concerning children, but they are pervasive nonetheless. For example, the 2000 BSAS found that 38 percent of respondents incorrectly thought that, if a cohabiting couple were to split up after 10 years of cohabitation, rights to “financial support” would be the same as those for divorcing spouses (Barlow et al. 2005). Similarly, 37 percent wrongly supposed that, upon the death of a partner after 10 years of living together, a cohabitant would have the same rights as a spouse to remain in a family home bought in his or her partner's name (Barlow et al. 2005). In follow-up interviews, it was evident that people holding these beliefs were far less confident as to exactly when these rights crystalized than they were as to the fact of the rights' existence, but the small number of estimates given ranged up to six years.

Far less dependent on the duration of relationships are beliefs that regarding natural children, unmarried fathers have rights and responsibilities equivalent to those of their married counterparts (Barlow et al. 2005). Half of all 2000 BSAS respondents thought wrongly that a cohabiting father had the same rights as a married father to make decisions about a child's medical treatment, at a time when unmarried fathers could obtain “parental responsibility” only through a formal parental responsibility agreement or court order.⁷ In follow-up interviews it was apparent that those making this incorrect assumption “assumed that legal rights and responsibilities towards children would depend not on marital status, but rather on the fact of parentage” (Barlow et al. 2005: 34).

The Origins of the Common Law Marriage Myth

Until recently it was widely thought that the common law marriage myth was of old, stemming from law and practice before Lord Hardwicke's Marriage Act of 1753, which codified marriage law.⁸ However, Probert's (2008, 2011) analysis of the historical record suggests that the term common law marriage is relatively new, and that the myth's origins can be “traced back no further than the 1970s” (Probert 2008: 21). Probert observes that it was only from this time—as legal rights and responsibilities began to be bestowed upon cohabiting couples—that the media began to use the term common law marriage regularly and nonpejoratively. Infrequent uses of the term in earlier years were “too negative to convey the impression that such living arrangements attracted legal rights” (21–22).

Since the 1970s, there has been increasing acceptance of cohabitation as “a valid partnering life-style choice” (Barlow et al. 2008: 34); increasing support for marriage-like rights for cohabitants (Barlow et al. 2005, 2008), with differences in public attitudes toward marriage and cohabitation now “not large” (Barlow et al. 2008; Dey & Wasoff 2007: 165); and increasing social conflation of cohabitation and marriage. But against this backdrop, as Barlow and James have noted, the “piecemeal” development of cohabitation law has left it “in a state of confusion, uncertainty and complexity” (2004: 143). The law treats cohabitants as everything from indistinguishable from spouses to “complete strangers” (Barlow et al. 2005: 50). So, in a context of legal uncertainty and complexity, the term common law marriage has had space to become “part of mainstream discourse” (Barlow et al. 2005; Lewis et al. 2009; Panades et al. 2007; Probert 2008: 19), with common usage of the term contributing further to public confusion (Probert 2008: 22). This is particularly likely to be so in the case of the (still) frequent use of common law marriage as a formal “definitional category” (Panades et al. 2007: 7) in organizations' dealings with the public. For example, with notable irony, “common law” remains a marital status category used by the brokerage Web site www.confused.com!⁹

More generally, Barlow et al. (2005: 45) have suggested that people's beliefs about cohabitation law, whether accurate or not, are based less upon acquired knowledge and more upon “notions of social logic, fairness and morality,” which (as we detail below) are similar with respect to cohabitation and marriage (Barlow et al. 2008). This reflects the finding of the 2004 Scottish Social Attitudes Survey (SSAS) that those who thought cohabitants should be able to make financial claims against each other upon separation were more often mistaken in their corresponding beliefs about cohabitation law (Wasoff & Martin 2005). A similar finding also emerged from the 2006 BSAS. In secondary analysis we undertook for the purposes of this study, we found that 55 percent of 2006 BSAS respondents who thought cohabitants of ten years standing should have the same rights as equivalent spouses also believed that “unmarried couples who live together for some time have a ‘common law marriage’ which gives them the same legal rights as married couples.” However, the figure was just 44 percent for those who did not think cohabitants should have such rights. These findings are in line with those of Kim's (1999) study of workers' knowledge of employment law in California, Missouri, and New York. She reports that workers' beliefs were “systematically erroneous” yet “remarkably similar” between states, “despite wide variations in the states' laws” (447 and 452). She concludes that respondents “assumed that the requirements of the law coincide with their beliefs as to how employers should behave and, therefore, answer the legal questions according to their own notions of fairness” (490). Similarly, Darley et al.'s (2001) small-scale study of people's understanding of three of four aspects of criminal law in Wisconsin, Texas, South Dakota, and North Dakota points to little variation between states in citizens' beliefs about the law, despite substantial differences in the actual law. Drawing on the “false consensus effect” (Ross et al. 1977), the conclude that people “assume that the state, in its moral wisdom, shares their personal views” (168). Even in the case of Texans' relatively good understanding of Texas's irregular law concerning the use of deadly force to protect property, differences in beliefs about the state of the law “disappeared completely when the relevant attitudes of the citizens [were] covaried out” (178).

The broad notions of fairness lying at the heart of the common law marriage myth are reflected in other areas of law. For example, the accumulation of legal rights with time is a feature of housing law (e.g., Law of Property Act, s.75, and Land Registration Act 2002, Sch.6, s.1), employment law (e.g., Employment Rights Act 1996, s.55 and s.108), and immigration law (e.g., British Nationality Act 1981, Sch.1, s.1).¹⁰ In the context of constructive trusts and proprietary estoppel, notions of fairness also underpin the role of detrimental reliance that frequently comes into play in the context of cohabitation, but do not extend far enough to support marriage-like rights.

The Implications of Public Misunderstanding of Cohabitation Law

The need for cohabitants to understand the implications of cohabitation law might seem self-evident, particularly given the continuing increase in the popularity of cohabitation. As much as possible, people should make important life decisions with regard to their implications. The rights of cohabitants as regards their children, and the rights and responsibilities of cohabitants upon relationship breakdown and death, can be markedly different from those of spouses. For example, maintenance (alimony) is not available to cohabitants who are financially dependent on their partners, a situation compounded by the fact that ordinary property law governs asset division between cohabitants.¹¹ In addition, cohabitants are not recognized by intestacy laws and are therefore at a disadvantage to spouses in this regard,¹² and parental responsibility is still not automatic for cohabiting fathers, although it has followed from a father's details being registered on a child's birth certificate since 2003, and the Welfare Reform Act 2009 made provision for promoting and facilitating such registration.¹³

Recent empirical studies have shown that the actual outcome of separation also differs greatly for cohabitants and spouses (Arthur et al. 2002; Douglas et al. 2007; Lewis, Tennant, & Taylor 2009). Douglas, Pearce, and Woodward (2007: 137), for example, found that many separated cohabitants “would have fared better if they had been married, especially the women with children who had no interest in property, and those (of either gender) claiming compensation for contributions to property owned by their partner.” Similarly, Lewis et al. (2009: 179) found that whereas divorce acts as a “redistributive” process, “financial division on the breakdown from cohabitation sustains the financial power dynamics in the relationship.” Thus, the president of the family division of the high court commented recently that “women cohabitees, in particular, are severely disadvantaged by being unable to claim maintenance and having their property rights determined by the conventional laws of trusts.”¹⁴

An understanding of rights is also important because it forms the basis of expectations as to the outcome of family breakup. If expectations are distorted, this may lead financially vulnerable cohabitants to “underinsure against the financial cost” of relationship breakdown (Williams 2009: 759),¹⁵ “militate against amicable resolution of disputes” (Dowding 2009: 207), and create “needless emotional harm” (Williams 2009: 735).

Moreover, more generally, it has been argued that legal literacy fosters confidence in individual and community problem solving, enhances perceptions of social justice and equality, promotes citizenship and community cohesion, contributes to effective government, and purges negative social myths (e.g., PLEAS Task Force 2007).

Thus, in the wake of Barlow et al.'s (2005) exposé of the extent of belief in common law marriage, and as part of a broader movement to increase legal literacy through public legal education (Advice Services Alliance, Citizenship Foundation and Legal Action Group 2004, 2005, PLEAS Task Force 2007), substantial efforts have been made in England and Wales to inform prospective cohabitants about cohabitation law and to provide guidance on how cohabitants' interests can be protected. Most notably, the government funded the Living Together campaign, launched in July 2004.

An early study of users of the Living Together Web site¹⁶ found that it was regarded as informative and often led users to identify a “need to take some form of legal action” to protect their positions, though even “legally aware” cohabitants rarely took appropriate legal steps” (Barlow et al. 2007: 166). Signs from the 2006 BSAS also suggest that the Living Together campaign made some progress in debunking the common law marriage myth. Thirty-nine percent of cohabitant respondents to the 2006 survey indicated they did not believe in common law marriage, compared to 35 percent in 2001. Furthermore, the percentage of respondents agreeing with the proposition that “marriage gives couples more financial security than living together” increased sharply between 2001 and 2006, from 48 percent to 61 percent (Barlow et al. 2008).¹⁷ Elsewhere, despite demonstrating that Texans' (relatively) good understanding of the laws regarding deadly force might be explained by reference to attitude, Darley, Carlsmith, and Robinson (2001) also pointed to higher levels of local media discussion of the issue, as compared to the other criminal law issues they studied, where beliefs were constant between states notwithstanding substantial differences in the actual law.

Nevertheless, a majority of respondents to the BSAS—including a majority of cohabitants—still believed aspects of the common law marriage myth.

Impediments to Legal Rationalism

It may be, though, that people would still choose to cohabit rather than to marry, and would take no or few steps to protect their interests, even if there were no constraints to marriage and their misconceptions about the law were rectified. Although it appears less the case that people slide or drift into and out of cohabitation in the United Kingdom as compared to the United States, the motivations to cohabit are evidently different from the motivations to marry (e.g., Lewis 2001). Moreover, various studies suggest that being better informed only rarely leads cohabitants to take action to better protect themselves (Barlow et al. 2005, 2007, 2008; Panades et al. 2007). However, Lewis (2001) reports that married couples sometimes mentioned “security” and entitlement to pensions as reasons for their having married, and Barlow et al. (2008) reports that cohabitants in longer-term relationships were more likely than others to take action to protect their legal positions.

Before relationships get into trouble, there are various inhibitors to “legally rational” (Barlow et al. 2005: 55) contingency planning, not least of which is that people are busy (Panades et al. 2007). Also, in the case of wills, death may seem a long way off. As one respondent to Panades et al.'s (2007: 32) study of the impact of providing legal information to unmarried parents commented, “I don't plan on dying just yet!” Furthermore, as various studies have shown, cohabitants (along with other people in relationships) may actively avoid raising issues that seem “negative or unromantic” (Panades et al. 2007: 33)—the raising of which may even be seen to signal uncertainty about a relationship¹⁸ —or asking questions whose answers they might not like (Douglas et al. 2009: 144). As Stake (Stake and Grossberg 1998: 427) puts it, “[N]o one wants to start the conversation.”¹⁹

More generally, optimism bias (e.g., Weinstein 1980; Williams 2009) acts to suppress accurate risk appreciation for those embarking on relationships. As Panades et al. (2007: 32) report, cohabitants “were generally optimistic about their relationship and, as they did not anticipate any problems, they did not, in general, see the point in preparing for them.”

In the United States, Baker and Emery (1993) have famously exposed the optimism bias of those in the early stages of intimate relationships. They found that, despite the accurate knowledge of people who were recently married or about to get married that around half of all American marriages end in divorce, more than half of respondents believed there was a 0 percent chance of their own marriage ending in divorce. People also believed that, were the worst to happen, the law would treat them more favorably than it would treat other people in general. More recently, Mahar (2003) again found the median perceived risk of becoming divorced to be 0 percent, with a mean risk of only 10 percent. Furthermore, Baker and Emery (1993) have demonstrated that this unrealistic optimism is resistant to correction. Completing a family law course made no dent on the confidence of respondents about the stability of their relationships!²⁰

Public Understanding of Marriage Law

As suggested above, it is often assumed that while the law relating to cohabitants may be frequently misunderstood, the public (presumably through shared experience and extensive media coverage) has a good grasp of marriage law. Indeed, Williams, while commenting that “in some contexts the vast majority of the population systematically mispredicts … the content of the law,” goes on to declare, without reference to any empirical evidence, that “most people, including new spouses, tend to be adequately informed about the basic rules of divorce” (758). Barlow et al. (2005: 54) have also suggested, again without reference to evidence, that while “it is not the case that ordinary members of the public are experts on laws about … divorce … they are aware of the general direction and message; certainly they are not so spectacularly and pervasively wrong as with common law marriage.” However, Baker and Emery's (1993), in their exploration of people's beliefs about divorce law, a subject they note to “have scarcely been studied,” despite suggestions that spouses find the law to be “unexpected” (439, citing Weitzman 1985), report that people had “relatively poor knowledge” of Virginia divorce statute law: they averaged only 60 percent accuracy when asked whether a series of statements about the law were true or false.

This Study

In this study, we draw on data from the 2010 English and Welsh Civil and Social Justice Panel Survey (CSJPS) to bridge and to build upon earlier findings concerning public misunderstanding of law.

We explore, uniquely and in great detail, the extent to which people hold inaccurate beliefs about both cohabitation and marriage law, the nature of inaccurate beliefs, and the relationship between inaccurate beliefs and social attitudes. Following from Barlow et al. (2005) and Baker and Emery (1993), we hypothesize that (1) there is substantial public misconception of both cohabitation and marriage law. Following from Kim (1999), Darley, Carlsmith, and Robinson (2001), Barlow et al. (2005), and Wasoff and Martin (2005), and with reference to findings from the BSAS, we hypothesize that (2) misconceptions reflect social attitudes (where they depart from the law). In particular, we hypothesize that (2a) the public holds similar beliefs about cohabitation and marriage law, (2b) people increasingly believe that both cohabitants and spouses to have rights to their partners' assets and to financial support in the event of relationship breakdown or death as relationship duration increases, and (2c) people's beliefs about cohabitant and spousal rights regarding biological children are independent of relationship duration.

In exploring the nature of inaccurate beliefs, we develop the first quantitative models of the temporal dimensions of the common law marriage myth.

We also investigate whether beliefs differ by marital status and/or by other demographic traits that have been associated with levels of legal literacy. Following from Saunders (1981) and Casebourne et al. (2006), we hypothesize that (3) cohabitants and spouses each have more accurate views than the other regarding their own legal position.

Finally, we ask what our findings suggest for the future development of the law as it relates to cohabitation and marriage.

Methods

Data

Data used for this study came from the 2010 CSJPS, a nationally representative survey of 3,806 individuals aged 16 or over, living in 2,318 households across 194 sample points in England and Wales. The survey was conducted face-to-face in respondents' homes, with all members of households interviewed separately.²¹ The data structure is hierarchical, with respondents nested in households. Interviews for the survey averaged 37 minutes, with a household response rate of 88 percent and a cumulative eligible adult response rate of 54 percent.

The 2010 CSJPS included a series of randomized experiments designed to explore public legal understanding. One experiment was focused on family law and involved presenting respondents with a hypothetical scenario (or vignette), describing the life circumstances of a couple and a sequence of life events, and respondents were asked to identify their beliefs about the rights and responsibilities of the protagonists in each circumstance and life event. Each respondent was presented with just one version of the scenario, and the relationship status of the couple (cohabiting or married) was allocated randomly at the outset. To fully explore the role of relationship duration on people's understanding, the duration of the relationship was also randomized across 16 discrete durations ranging from one month to 20 years. The scenario was presented to 3,453 respondents in total, meaning that each of the 32 versions of the scenario was presented to 108 people, on average.

The basic details of the scenario were that while the couple, John and Sarah, lived together, John earned a good salary and had sizable savings, while “Sarah looked after the home and had not worked” (for all references to the 2010 CSJPS and our analysis of it, see Pleasence et al. 2011).

The first life event to befall John and Sarah was that their relationship broke up at a point when they had no children. At this point respondents were asked whether Sarah would have “a good legal claim to financial support from John.” The second life event to befall the couple was that, instead of breaking up, John died without leaving a will. Respondents were then asked whether Sarah would “automatically inherit” any of John's savings or belongings.

John was then resuscitated, and a one-year-old child (of the couple) was introduced into the scenario. It was also explained that, “through an oversight, John's name was not put on the birth certificate,” and John had never signed any other forms or documents concerning the child.

The third life event was another breakup, but this time there was a small child in the family. Respondents were now asked whether, if the child went to live with Sarah, John would have “a legal obligation to provide financial support for the child” (italics in original) and, separately, for Sarah.

Finally, the couple's relationship was again put back together, but, rather than living happily ever after, John and Sarah's child was involved in an accident, and they were faced with a choice between medical treatments associated with very different risks. Respondents were asked, first, whether John would “have the legal right to decide upon the treatment, in the absence of Sarah” and, second, whether Sarah would have the legal right to decide upon the treatment, in the absence of John.

Analytical Methods

First, we summarized the responses to each of the relationship-scenario questions on the basis of whether John and Sarah were married or cohabiting.

Second, we attempted to fit asymptotic regression models (e.g., Seber and Wild 2003; Stevens 1951) to the data for each question in the relationship scenario, comparing positive responses to negative and “don't know” responses combined. This was to assess how responses changed with the experimental manipulation of relationship duration. Asymptotic regression models are useful in cases where y (the proportion of respondents answering a question positively) tends to a limit as x (relationship duration) tends toward infinity. The model equation is

$y=\alpha +\beta {\rho }^x,$

where α is the asymptotic value of y, with y tending toward this value as x tends toward infinity, β is the change in y as x goes from zero to infinity, and ρ is the factor by which deviation of y from its asymptotic value is reduced for each unit step along the x-axis.²² Models were fitted with nonlinear regression using STATA version 11. We attempted to fit 12 models in all—two for each scenario question, with one for when John and Sarah were married and one for when they were cohabiting.

Third, we fitted cross-classified logit models for each of the six scenario questions. These models were used to assess whether there were differences in responses (again comparing positive responses to negative and “don't know” responses combined) based on whether or not respondents, and John and Sarah, were married or cohabiting. The models controlled for the experimental variations in John and Sarah's relationship, to examine the impact of relationship status independently of the impact of relationship duration. The models also included respondent gender, age group, personal income, and academic qualifications; whether or not respondents had children living in their homes; and the interaction of these factors with John and Sarah's relationship status. This strategy was meant to ensure that any differences in respondents' views by relationship status were not simply consequences of other factors such as age or socioeconomic status. In terms of the data structure, responses could be “classed” by the duration of John and Sarah's relationship and by household. This type of data structure can be described as cross classified (see Goldstein 2010 for an introduction) and can be conveniently modeled using Markov Chain Monte Carlo methods (Browne 2009) within MLwiN (Rasbash et al. 2009a). This accounted for clustering in responses within relationship durations and within households, by creating random terms for each. In addition, since the impact of relationship duration on responses may have varied by whether John and Sarah were married or cohabiting, the relationship status term was also allowed to vary by duration of relationship. More generally, there are several consequences of failure to correctly account for this type of data structure, including underestimation of standard errors associated with regression coefficients (Rasbash et al. 2009b).

Findings

Summary of Responses

Table 1 summarizes responses to the relationship scenario questions on the basis of whether John and Sarah were married or cohabiting.

Table 1. Responses to the Six Relationship Scenario Questions on the Basis of Whether John and Sarah Were Married or Cohabiting

Table 1 oversimplifies respondents' general understanding, since the length of John and Sarah's relationship was randomized into 16 durations, with relationship duration potentially influencing responses.

Asymptotic Regression Models and Time Dependency

Table 2 presents output from the asymptotic regression models and thus examines the extent to which respondents believed that John or Sarah had rights (or responsibilities) in each of the scenarios as relationship duration varied. Separate output is provided for marriage and cohabitation scenarios.

Table 2. Estimated Coefficients (and Standard Errors) for Asymptotic Regression Models of Respondents' Responses to the Six Relationship Scenario Questions (Statistically Significant Terms in Bold)

Cross-Classified Models and the Impact of Respondent Relationship Status, John and Sarah's Relationship Status, and Their Interaction

Table 3, which presents statistical output for the six cross-classified logit models, shows the tendency for respondents to positively proclaim that John or Sarah had rights or responsibilities in each situation (as compared to negative or “don't know” responses) on the basis of respondent relationship status (married or cohabiting), John and Sarah's relationship status, and their interaction (as well as a range of other sociodemographic predictors and their interaction with John and Sarah's relationship status). In each model, random terms were included to account for the data structure (see Analytical Methods above). Figure 1 (a to f) shows the proportion of respondents who responded positively to each scenario question, by both respondents' and John and Sarah's relationship status, derived by simulation from the models in Table 3.

Figure 1. a–f. Proportion of respondents agreeing with each of the relationship scenario questions, derived from models A to F in Table 4 (which includes a key below the table), on the basis of John and Sarah's relationship status and whether respondents were married (white bars) or cohabiting (gray bars). Note: scale varies by figure.

Table 3. Estimated Coefficients (and Standard Errors) for Cross-Classified Logit Models of Respondents' Responses to the Six Relationship Scenario Questions (Statistically Significant Terms in Bold)

† Model A—If the couple break-up … will Sarah have a good legal claim to financial support from John?

Model B—If, instead of breaking up, John dies without leaving a will, will Sarah automatically inherit any of his savings or belongings?

Model C—If the couple break-up … and the child goes to live with SARAH, will JOHN have a legal obligation to provide financial support for the child?

Model D—Will SARAH have a good legal claim for herself to financial support from JOHN?

Model E—Would JOHN have had the legal right to decide upon the treatment (of their child), in the absence of SARAH?

Model F—Would SARAH have had the legal right to decide upon the treatment (of their child), in the absence of JOHN?

Table 4. Estimated Coefficients (and Standard Errors) for Asymptotic Regression Models of Respondents' Responses to “Whether Sarah Would Have a Good Legal Claim to Financial Support from John Following Break-up?” on the Basis of Their, and John and Sarah's, Relationship Status (Statistically Significant Terms in Bold)

Financial Support Following a Breakup with No Children Involved

In the case of the first question (concerning whether Sarah would have a good legal claim to financial support from John following separation), the asymptotic regression models fit well, with the three model coefficients statistically significant, whether John and Sarah were married or cohabiting (testing the model terms; John and Sarah married—α, t = 28.16, p < 0.001; β, t = −8.99, p < 0.001; ρ, t = 8.49, p < 0.001; John and Sarah cohabiting—α, t = 17.05, p < 0.001; β, t = −13.09, p < 0.001; ρ, t = 19.94, p < 0.001). As shown in Figure 2, which presents the fitted models and observed values, the proportion of respondents believing that Sarah would have a good legal claim to financial support upon separation increased over time, whether Sarah and John were married or cohabiting.

Figure 2. Observed and fitted asymptotic regression model values of whether respondents felt “Sarah would have a good legal claim to financial support from John following break-up?” on the basis of whether John and Sarah were married or cohabiting.

However, there were differences in coefficients between the models. When John and Sarah were married, the asymptotic value was higher. In all, 64.8 percent of respondents believed Sarah would have a good legal claim when married, as the duration of their relationship tended toward infinity, compared to 54.5 percent when Sarah and John were cohabiting. Also, the overall change over time can be seen to have been lower in the case of marriage (31.0 percent versus 43.7 percent). The smaller ρ-value in the case of marriage also indicated that opinion approached the asymptotic value more quickly than it did in the case of John and Sarah's cohabitation.

In regard to the observed data, while 34.2 percent of respondents thought that Sarah would have a good legal claim to financial support after one month of marriage, the figure was 11.4 percent after one month of cohabiting. As relationship duration increased, so did the percentage of respondents believing that a spouse or cohabitant would have a good legal claim, with the gap between the respective relationship statuses narrowing considerably for relationships of greater duration—after 20 years, the figures were 64.6 percent and 51.6 percent, respectively.

As there was clear evidence of time dependency for both models for the first scenario question, a further four asymptotic regression models were fitted. These considered the respondent's relationship status (married or cohabiting) in addition to that of John and Sarah. Table 4 shows statistical output for the additional models, with the fitted models and observed data shown in Figure 3.

Figure 3. Observed and fitted asymptotic regression model values of whether respondents felt “Sarah would have a good legal claim to financial support from John following break-up?” on the basis of whether John and Sarah and the survey respondent were married or cohabiting.

All model terms were statistically significant, with the exception of the ρ term for the “John and Sarah married/respondent cohabiting” model, suggesting some uncertainty over the rate of increase in earlier years compared to other models. However, significance for the total increase over time (β, t = −3.09, p = 0.002) and maximum (α, t = 13.00, p < 0.001) suggest greater confidence in their values.

In the case of the two models where John and Sarah were married, a far lower percentage of cohabiting than married respondents believed that Sarah would have a good claim to financial support after a very short relationship. Also, despite a greater increase over time (54.7 percent compared to 30.9 percent), the maximum level of belief (as relationship duration increased) remained lower for cohabiting respondents than for married respondents (59.7 percent compared to 70.8 percent).

As can be seen in Table 4 and Figure 3, in the models where John and Sarah were cohabiting, total change, shape of change, and the maximum percentage agreeing that Sarah had a good legal claim were fairly similar. Cohabitant respondents were moderately, and more accurately, less positive than were their married counterparts.

The cross-classified models, which controlled for relationship duration and household effects, show that if John and Sarah were cohabiting, rather than married, there was a significant reduction in the probability of respondents' believing that Sarah would have a good claim to financial support upon separation (see Figure 1a²³). Nevertheless, a sizable minority of respondents believed that Sarah had a good legal claim for support in the cohabitation scenario (see Figure 1a). There was no evidence of an interaction between John and Sarah's and the respondent's relationship status.

Elsewhere, there was some evidence of a significant increase in positive responses for male respondents, regardless of John and Sarah's circumstances (χ² ₁ = 6.36, p = 0.011).

Inheritance

In the case of the second question (concerning whether, if John died without leaving a will, Sarah would automatically inherit any of his savings or belongings), the asymptotic regression model fit relatively well in the case of the cohabitation scenario, with all three terms statistically significant (α, t = 3.68, p < 0.001; β, t = −2.92, p = 0.004; ρ, t = 14.39, p < 0.001). The percentage of respondents believing Sarah would automatically inherit went from 7.1 percent for the shortest relationship duration to an asymptote of 26.7 percent, with the increase fairly gradual over time. When John and Sarah were married the model reached a maximum of 44.4 percent, although there was uncertainty around the size and shape of change over time.²⁴

In terms of the observed data, while 51.3 percent of respondents thought that Sarah would automatically inherit some of John's savings or belongings after one year of marriage, the figure was still 50.5 percent after 10 years. With Sarah and John as cohabitants, the figures were 2.9 percent and 17.9 percent, respectively.

Referring to the cross-classified model (Model B in Table 3 and Figure 1b), as with financial support following separation, when John and Sarah were cohabiting there was a significant reduction in the likelihood that respondents believed Sarah would inherit some of John's property (see Figure 1b).²⁵ The respondent's relationship status and the interaction between John and Sarah's and the respondent's relationship status had little or no impact on responses.

Of the other predictors in the model, respondents with a high level of academic qualifications were more likely to respond positively where John and Sarah were married (χ² ₁ = 7.87, p = 0.005) and less likely to do so where they were cohabiting (χ² ₁ = 13.34, p < 0.001).

Financial Support Following Separation When the Couple Has a Child

As shown in Table 2, in the case of the third question (concerning whether, following separation, John would have a legal obligation to provide financial support for his child), there was little difference between the marriage and cohabitation scenarios in terms of the maximum percentage of respondents answering positively (72.2 percent for the marriage scenario compared to 70.6 percent for the cohabitation scenario). Moreover, in both cases, the change in the percentage of respondents answering positively as relationship duration increased was relatively small and nonsignificant. The shape of any duration-related change was also unclear, as denoted by the nonsignificant ρ terms.

The question concerning child support was followed by the question of whether Sarah would have a good legal claim for financial support from John for herself, given the introduction of a child into the scenario. Whether or not John and Sarah were married or cohabiting, there was limited evidence of coherent time dependency of answers. While the model for the cohabitation scenario showed a 43.7 percent change in positive responses as duration increased, this was clearly nonsignificant, with the model fitting poorly. Likewise, when John and Sarah were married, the 10.3 percent change as relationship duration increased was clearly short of significance.

Turning to the cross-classified models (Models C and D in Table 3 and Figure 1c and 1d), in the case of the question concerning financial support for the child, none of John and Sarah's relationship status, the respondent's relationship status, or the interaction of the two had any significant impact on the percentage of respondents answering positively. As regards Sarah's claim for herself, John and Sarah cohabiting was related to a reduction in the likelihood of respondents' believing Sarah had a good claim (see Figure 1d²⁶). Again, the respondent's relationship status (χ² ₁ = 0.15, p = 0.70) and its interaction with John and Sarah's relationship status (χ² ₁ = 0.55, p = 0.46) had little impact on responses.

Of the other variables in the models, in Model C, age group had some impact. Compared to the youngest respondents, other age groups were more likely to respond positively where John and Sarah were married and less likely to do so where they were cohabiting (see Table 3). For Model D, male respondents were more likely to respond positively, particularly where John and Sarah were cohabiting.

Medical Treatment

In the case of the final two questions (concerning John and Sarah's right to decide upon medical treatment for their child in each other's absence), relationship duration made no difference to respondents' answers. As can be seen in Table 2, three of the four asymptotic regression models failed to converge, indicating that the model was not appropriate, while the other showed no evidence of any link between relationship duration and answers.

Referring to the cross-classified models (Models E and F in Table 3 and Figures 1e and 1f), there was a significant but modest reduction in the likelihood of respondents to believe John had the right to decide upon treatment alone when John and Sarah were cohabiting (see Figure 1e).²⁷ There was also a reduction in the case of cohabitant respondents, though this fell short of significance (χ² ₁ = 1.62, p = 0.20). There was no significant interaction between John and Sarah's and the respondent's relationship status. For Sarah's right to decide upon treatment in the absence of John, none of John and Sarah's relationship status, the respondent's relationship status, or their interaction had a significant impact on the very high percentage of respondents agreeing that Sarah did have the right.

Elsewhere in the models, for Model E, male respondents were more likely to respond positively, particularly where John and Sarah were married (χ² ₁ = 6.34, p = 0.011). For Model F, there was an increase in positive responses among those with high academic qualifications, while the respondents with children in their homes were more likely to respond positively where John and Sarah were cohabiting (χ² ₁ = 5.89, p = 0.015).

Discussion

Knowledge of the Law

In line with hypothesis 1, our findings confirm substantial and ongoing public misunderstanding of cohabitation law. For example, 52 percent of respondents erroneously believed that a financially dependent cohabitant—where one partner has worked while the other has looked after the family home—would have “a good legal claim to financial support” on separation after 10 years.

The 52 percent figure is higher than the corresponding 38 percent figure derived from the 2000 BSAS (Barlow et al. 2005). However, there are important differences between the surveys that make it difficult to compare the figures. First, the corresponding BSAS question did not set out any details of the division of labor within the relationship. Second, as we note above, the BSAS asked whether a cohabitant would have the same right as a spouse to claim financial support, rather than whether they would actually have the right. Thus, it is impossible to fully establish what proportion of BSAS respondents held erroneous beliefs about the underlying law, as we do not know what rights they believed spouses to possess. This becomes particularly problematic if there is significant misunderstanding of spousal rights, which our findings demonstrate that there is. A full 35 percent of respondents did not believe that a financially dependent spouse would have a good legal claim for financial support after 10 years of marriage, and 29 percent mistakenly stated that such a spouse would not have a claim. We also found evidence of misunderstanding of rights in the case of short marriages. For example, 34 percent of respondents overstated a spouse's legal claim for financial support after a marriage of just one month's duration.²⁸

Turning to rights of other kinds, 14 percent of people wrongly believed that if a cohabitant dies without a will, his or her partner would automatically inherit some of the deceased's property. Conversely, 48 percent of people wrongly believed that a spouse would not automatically inherit some of his or her intestate partner's property.

We also found that 21 percent of respondents wrongly believed that a cohabiting biological father would have no legal responsibility to provide financial support for a child on the breakup of a relationship, and 20 percent believed the same in the case of a married father.

Furthermore, 47 percent of people wrongly believed that a cohabiting father who has not met the formal requirements for parental responsibility would have the right to decide upon important medical treatment. Conversely, 36 percent of people wrongly believed that a married father would not have such a right.

Public Understanding of Law and Attitudes

In line with hypothesis 2, our findings demonstrate clearly that the many misconceptions held about both cohabitation and marriage law reflect the divergence of social attitudes from the law. There is symmetry of error in people's beliefs about marriage and cohabitation law, where beliefs about both cohabitation and marriage law err from their (often opposing) correct legal positions to rest more closely in line with social attitudes. This produces, in line with hypothesis 2a, substantial convergence of beliefs as to the law between different forms of union (with the exception of intestacy provision, discussed below), mirroring the convergence of social attitudes. In the absence of a clear understanding of the law, it would appear that people tend to assume it concurs with what they think it ought to be.²⁹

While cohabitation law and marriage law are framed very differently, people's attitudes toward how they should be framed are quite similar (Barlow et al. 2008). For example, when 2006 BSAS respondents were presented with a scenario of a childless couple having lived together for 10 years, one partner having worked unpaid to build up the other partner's business, with the partner running the business also owning the family home, 87 percent thought that there should be a “right to financial provision on separation” (45) if the couple were cohabitants, compared to 93 percent if they were married. When one partner had had a well-paid job requiring frequent moves and the other had worked where possible but not had a settled career, the figures dropped to 69 percent and 81 percent, respectively. When the couple had been together for just two years, both figures dropped substantially, to 38 percent and 61 percent, respectively.³⁰

So, for long-term relationships involving financial dependence, there was a high level of support for rights to financial provision upon separation, with only a slight favoring of spouses over cohabitants, but this support dropped substantially in the case of short-term relationships, with a much more pronounced decrease for cohabitants. Notwithstanding the modest, but significant, differences in beliefs concerning cohabitants' and spouses' rights in relation to financial provision demonstrated by the statistical models, in line with hypotheses 2a and 2b, this picture is very similar to that of 2010 CSJPS respondents' understanding of the law, where we also see a narrowing in the difference between beliefs concerning rights to financial support for cohabitants and spouses as relationship duration increases. For relationships of 10 years (as described above), 52 percent of respondents believed that a financially dependent cohabitant would have a good legal claim, with 65 percent believing the same in the case of a spouse. For relationships of just one month, the figures were 11 percent and 34 percent, respectively. The converse of the first of these findings is that 35 percent of respondents did not indicate that a financially dependent spouse would have a good legal claim for financial support after 10 years of marriage, and 29 percent mistakenly believed that such a spouse would not have a claim.

In quantitative terms, the models suggest that the aspect of the common law marriage myth that concerns financial support for cohabitants came to be believed by 50 percent of people once the relationship duration reached ten years and two months.

Curiously, where a biological child was present in a family, relationship duration became an ancillary factor influencing beliefs about the legal rights of dependent cohabitants to receive financial support from their partners. This points to conflation of the position of children and their primary caregivers, and qualifies our findings in respect of hypothesis 2b.

Again illustrating symmetry of error, in the case of intestacy, while 18 percent of respondents incorrectly believed a cohabitant would automatically inherit some of a deceased partner's property after a 10-year relationship, 44 percent incorrectly believed a spouse would not automatically inherit some of a deceased partner's property after 10 years. In line with social attitudes and hypothesis 2b, these beliefs were influenced by relationship duration. So, while 18 percent of respondents erroneously believed that a cohabiting partner would automatically inherit some of a deceased's property after 10 years, the figure was just 8 percent after one year. The temporal dimension of beliefs in relation to spouses was less clear, though. Also, as noted above, contrary to hypothesis 2a in this instance, there was a significant and considerable overall difference between beliefs pertaining to cohabitation and those pertaining to marriage law. However, this was very much the exception rather than the rule.

Symmetry of error was also evident in the case of a biological father's right to independently decide upon important medical treatment for his child. Only 52 percent of respondents correctly believed married fathers to have such a right—a very similar figure to the 47 percent that incorrectly ascribed such a right to a cohabiting father without formal parental responsibility. In line with social attitudes and hypothesis 2c, beliefs about rights and responsibilities concerning children were independent of relationship duration.

Obstacles to Accurate Public Understanding of Law

Overall, the level of consistency of belief about the content of family law across cohabitation and marriage scenarios is striking, and it reflects increasing consistency of attitude. Our findings build upon earlier studies (e.g., Barlow et al. 2005; Darley, Carlsmith, & Robinson 2001; Kim 1999; Saunders 1981) and suggest that, in general, public legal understanding may be substantially driven by attitudes. Consequently, our findings also highlight additional obstacles that public legal education initiatives must overcome in order to be successful. In the family context, as well as indifference to law resulting from optimism bias, erroneous beliefs based on attitudes are likely to be, in Kim's words, “resistant to change” (1999: 447). She reports that in the employment context, a fairness norm “overshadows the influence of most … experiential factors” (448). Likewise, Ellickson (1991: 115), in his study of the law's place in ranching disputes in Shasta County, California, reports that “the cattlemen resist absorbing information that is inconsistent with their folklore.” Thus, even repeated experiences of insurance companies and courts following different principles did not dislodge the belief that, in the event of road collisions in “open range,” “the motorist buys the cow.”

This challenges the notion that misunderstandings result simply from lack of information, and, as Ellickson (1991) notes in the context of ranching disputes, is in line with what we would expect from cognitive dissonance theory.³¹ Cognitive dissonance theory (Festinger 1957: 3) predicts that when a person's cognitions (which include knowledge, opinions, and beliefs) are at odds, or dissonant, this is “psychologically uncomfortable” and will motivate the person to “reduce the dissonance and achieve consonance.” Cognitive dissonance can therefore lead people to modify their beliefs so they are in accord, or to avoid “situations and information that would likely increase the dissonance” (3). In the case of law, as Louis and Taylor (2005: 107) have observed, “people's perceptions of their rights and duties are learned in a social context.” People's immersion in the social world and exposure to social attitudes therefore act as obstacles to being receptive to contradictory information about the law.

The Relationship between Legal Literacy and Demographics

While the cross-classified logit models did not indicate any statistically significant differences in the propensity of cohabitants and spouses to answer positively questions about the rights of cohabitants and spouses, those differences that were visible were all consistent with the hypothesis that cohabitants and spouses, as groups, each have more accurate views than the other regarding their own legal position. So, differences (albeit nonsignificant) between the beliefs of cohabitants and spouses in relation to financial support for a partner, intestacy, and paternal rights to decide on medical treatment (Figures 1a, 1b, 1d, and 1e)—all matters where cohabitation and marriage law diverge—were all in the direction required by hypothesis 3. On the other hand, virtually no differences were observed (and no consistent differences were observed) in relation to child support and maternal rights to decide on medical treatment (Figures 1c and 1f)—both matters where cohabitation and marriage law concur. Furthermore, as is shown by Figure 3, beliefs concerning how cohabitant and spousal rights to financial support vary with time are consistent with hypothesis 3.

Thus, there is some level of support for the general proposition that legal understanding increases along with the salience of the issues involved. This, though, sits somewhat at odds with Barlow et al.'s (2008) finding of no difference in levels of belief in the common law marriage myth between cohabitants and spouses.

The differences we observed between the beliefs of cohabitants and spouses are also consistent with a positive impact of public legal education initiatives, such as the Living Together campaign, though that campaign did not extend to spousal rights, where, along with cohabitants' rights, we see a different pattern of belief as between spouses and cohabitants. Also, or alternatively, the differences between the beliefs of cohabitants and spouses might reflect a greater emphasis on personal autonomy on the part of cohabitants (Lewis 2001). This is not to subscribe to the view that cohabitation is indicative of the “individualistic hell” that Lewis (2001) ascribes to social “pessimists” (123). Commitment clearly continues to be a central feature of intimate relationships in the United Kingdom, including relationships of cohabitation (e.g., Bowman 2010; Duncan et al. 2008; Lewis 2001), but it is a feature that is increasingly “articulated in terms of something that comes from within rather than being imposed from without” (Lewis 2001: 123). Of course, there is great diversity in the nature of cohabitation relationships, as there is in the nature of marriages (e.g., Bowman 2010; Douglas, Pearce, & Woodward 2009; Lewis 2001; Manting 1996). Lewis (2001) has described cohabitants of 25 years standing who are functionally equivalent to married couples, as well as young married couples that regard themselves as having “no obligations” to one another (144).

As regards other factors linked to legal literacy, in keeping with earlier findings (e.g., Casebourne et al. 2006; Williams & Hall 1972), we found that respondents with more academic qualifications were more likely than others to attribute a right to financial support to spouses, and less likely to attribute such a right to cohabitants.³² They were also more likely to recognize that mothers always have the right to make decisions about her child's medical treatment. We also found that male respondents were more likely than their female counterparts to believe that a financially dependent (female) partner would have a good legal claim to financial support, irrespective of the relationship type and irrespective of whether or not the relationship produced children. Men were also more likely to believe that a father is legally able to decide on a child's medical treatment.

The Future of Marriage and Cohabitation Law

Recognition of the diversity of modern-day intimate relationships, along with continuing concern that opposite-sex cohabitants are often unaware of their potential vulnerability under the law (tied to fact that “cohabitants become economically interdependent during their unions” [Bowman 2010: 142] and can face a substantial decline in economic welfare on separation [Avellar & Smock 2005]), has led to various calls for law reform in England and Wales.

For example, the Law Commission of England and Wales (Law Commission 2007a) recently recommended that, after cohabitants have a child together or have lived together for a specified period of years (suggested to be between two and five), then, unless they have entered into an opt-out agreement, they should be able to apply for financial relief “to ensure that the pluses and minuses of the relationship [are] fairly shared between the couple” (Law Commission 2007b: 4). This approach, aimed at protecting the autonomy of cohabitants and preserving a clear distinction between those who do and do not enter into a formal public commitment, is similar to the earlier proposals of the Law Society (2002) and to s.28 of the Family Law (Scotland) Act 2006, and very different from the approach of marriage law. The concern is to reverse unequal economic advantage or disadvantage resulting from cohabitation, rather than to create an entitlement to equal division or to specifically address the needs of the parties. In the case of the death of a cohabitant, the Law Commission recommended that the Inheritance (Provision for Family and Dependants) Act 1975 be amended to remove the limitation of awards to what is reasonable for an applicant's maintenance, and also to require the court to “have regard to the award the applicant might reasonably have expected to receive in proceedings for financial relief on separation” (Law Commission 2007a: 127). No change to the law has resulted from the Law Commission's recommendations.

As Dey and Wasoff (2007) have detailed, there is a variety of competing and conflicting purposes of cohabitation law reform: to promote a particular form of relationship, to provide parity in the case of functionally equivalent relationships, and to protect those who are economically disadvantaged through relationships. Balancing these purposes is politically perilous, and relevant reforms that have been made around the world reflect very different approaches to the balancing act (Bowman 2010; Bulloch & Headrick 2005). In the United States, for example, concern about the impact of cohabitation on marriage is prominent. Thus, as Bowman (2010: 16) has reported, notwithstanding the questionable constitutional standing of these laws in light of Lawrence v. Texas, 539 U.S. 558 (2003), seven states still “had criminal laws against cohabitation” as of January 2009. However, the legal positions of opposite-sex cohabitees varies hugely from state to state, and in some, such as Washington, significant legal commitments attach to “meretricious” relationships (and, following Senate Bill 5688 of 2009, if one partner is at least 62 years of age, cohabitants can register a “domestic partnership” and be treated in the same way as spouses under state law). Moreover, in the case of rights concerning children, the U.S. Supreme Court “has struck down almost every distinction based on legitimacy of parentage as violating the Equal Protection Clause”³³ (Bowman 2010: 21).

In New Zealand, concerns to provide parity and to protect cohabitees from the unfair consequences of separation are more prominent.³⁴ As a result, New Zealand has “legislated extensively to recognize and provide for the legal consequences of cohabitation” (Kennedy 2004: 238). The Property (Relationships) Amendment Act 2001 sets out a principle of equal sharing, “with powers to depart from this to address significant economic disparity between the parties resulting from the division of functions between them in the course of a relationship” (Kennedy 2004: 242). Maintenance payments are a possibility. Alongside this presumptive provision, the Civil Union Act 2004 introduced a registration system (also open to same-sex couples) whose provisions “were based on [those] for marriage, but have been modernized to reflect current law, policy, and practice.”³⁵ However, emphasizing the contested nature of cohabitation policy, the passage of the Civil Union Act 2004 was controversial, with voting split on mainly party lines.³⁶

In England and Wales, cohabitation policy remains contested, with legal provision more akin to the United States model than to the New Zealand model. However, this contest is being played out against a backdrop of continuing profound social change, with a majority of the population now favoring an extension of rights to cohabitants (Barlow et al. 2005, 2008). Legal and social realities are increasingly distinct. Our findings demonstrate that legal reality and the public's perception of legality—in the case of both cohabitation and marriage—are also distinct. As we have argued above, this is likely a further reflection of the divergence of law and social norms.

Thus, building upon the work of Darley, Carlsmith, and Robinson (2001: 165) in the criminal sphere, our findings raise questions about the extent to which the form of the law can be expected to promote a particular form of relationship (the “ex ante function” of the law). Our findings may also anticipate the future shape of the law. As Roscoe Pound (1907; 615) famously argues, “[I]n all cases of divergence between the standard of the common law and the standard of the public, it goes without saying that the latter will prevail in the end.”

Whatever may come to pass, though, and whatever the merits of the various options for legal change, it is plain that mass public ignorance of the basic structure of the law underlying a major component of people's lives is far from satisfactory. It is important for both cohabitants and spouses to understand their legal positions. Decisions to live with or to marry another person are among the most momentous decisions in people's lives, and they involve profound consequences of many kinds, including legal.

Conclusion

In conclusion, our findings demonstrate substantial and ongoing public misconceptions about both cohabitation and marriage law, which are spheres of law that can profoundly influence the life course. They illustrate, uniquely, the connection between relationship duration and beliefs about rights to financial support, and confirm the independence of relationship duration and beliefs about rights and responsibilities concerning children. Our findings demonstrate, for the first time, a great similarity between beliefs concerning cohabitation and those concerning marriage law, with erroneous beliefs reflecting the divergence of social attitudes from the law. Our findings build upon and bridge previous findings (e.g., Barlow et al. 2005; Darley, Carlsmith, & Robinson 2001; Kim 1999; Saunders 1981) and suggest that, in general, public legal understanding may be substantially driven by attitudes. While both current cohabitants and spouses exhibited better understanding of the law relating to their own form of relationship than those of the other status group, misunderstandings were still common. These misunderstandings are likely to be difficult to dislodge. They may also signal the future state of the law.