Overview

Reports on the results of research, commissioned by the Australian Institute of Family Studies (AIFS) and carried out in the Department of Demography, Australian National University, into marriage patterns in Australia from 1920 to 1985. Data analysed are drawn from three main sources: the 1981 Australian Census, the national vital registration system which tabulates marriages on an annual basis, and the Family Formation Survey of 2544 Australians aged 18- 34 years conducted in 1981-82 by the AIFS. Author states that an understanding of patterns of entry into first marriages - the ages at which people marry, the personal attributes which distinguish those who marry at different ages, the extent to which birth cohorts ever marry, and the characteristics which set those who marry apart from those who never marry - together with an understanding of changes in these patterns is central to an appreciation of the changing process of family formation in Australia.

Jointly published by Department of Demography, Australian National University and Australian Institute of Family Studies

Table of contents

• FOREWORD
• PREFACE
• AUTHOR'S PREFACE
• TABLE OF CONTENTS
• LIST OF TABLES
• LIST OF FIGURES
• CHAPTER 1 INTRODUCTION
  • Organisation of the study
  • Historical backdrop: Marriage from the convict era to the early twentieth century
• CHAPTER 2 AGE AT FIRST MARRIAGE AND PROPORTIONS MARRYING 1920-1970
  The period and cohort approaches to analysing first marriage activity
  • Period trends
    • The intensity of first marriage activity
    • The balance of the sexes in the marriage market
    • The timing of marriage
  • Cohort trends
    • Cumulative first marriage rates
    • The timing of marriage
  • Causes of the marriage boom
    • The historical dimension
    • The 1940s and 1950s
    • The 1960s and early 1970s
• CHAPTER 3 AGE AT FIRST MARRIAGE AND PROPORTIONS MARRYING 970-1985 AND BEYOND
  • Period trends
  • Cohort trends
  • Explanations for contemporary first marriage trends
    • The timing of change
    • Induced abortion: The trigger?
    • Living together
    • Economic forces
    • Ideological change
  • Likely future trends
• CHAPTER 4 DETERMINANTS OF AGE AT FIRST MARRIAGE: A REVIEW OF THE LITERATURE
  • Social origins
    • Parental socioeconomic status
    • Cultural heritage
    • Family structure
  • Socialisation and significant others
  • Personality
  • Contingencies
  • Educational, Occupational and Marital plans
  • Some concluding remarks
• CHAPTER 5 DETERMINANTS OF AGE AT FIRST MARRIAGE: INSIGHTS INFORM CENSUS AND VITAL REGISTRATION DATA
  • Age at first marriage and bridal pregnancy
  • Age at first marriage and education
  • Age at first marriage and occupation
  • Age at first marriage and birthplace
  • Age at first marriage and place of residence
  • Age at first marriage and race
• CHAPTER 6 DETERMINANTS OF AGE AT FIRST MARRIAGE: PRELIMINARY INSIGHTS FROM SURVEY DATA
  • The survival analysis technique
  • Findings from survival analysis
    • Father's occupation
    • Parent's education
    • Religion, Religiosity and Religious Upbringing
    • Metropolitan or Non-Metropolitan Upbringing
    • Parental Loss
    • Family Size and Birth Order
    • Home Life Whilst Growing Up
    • Quality of Parents' Relationship
    • Adolescent Dating and Sexual Behaviour
    • Attitudes to Marriage, Motherhood, Being Single and Living Together
    • Mate Selection Criteria
    • Number of Years of Secondary Schooling
    • Type of Schooling
    • Involvement in Tertiary Study
    • Parental Encouragement of Education
• CHAPTER 7 DETERMINANTS OF AGE AT FIRST MARRIAGE: A MULTIVARIATE APPROACH
  • General approach and selection of independent variables
  • Marriage timing among males and females aged 30-34: Preliminary analysis
  • Marriage timing among males and females aged 30-34: Multivariate analysis
    • Selection of variables for multivariate analysis
    • Outline of a general multivariate model of the marriage timing process
    • A preliminary test of the model
    • Refinement of the model: A path analytic approach
    • The refined model: Empirical results
  • Changes in determinants of female ages at first marriage
• CHAPTER 8 THOSE WHO MARRY AND THOSE WHO NEVER MARRY
  • Reasons for marrying
  • Factors associated with being never married at ages 35-49 years
    • Education
    • Occupation
    • Birthplace
    • Religion
  • Marriage intentions of never married 18-34 year olds
• CHAPTER 9 SUMMARY AND CONCLUSION
  • Trends in age at first marriage and proportions marrying
  • Determinants of age at first marriage
  • Reasons for marrying and factors associated with failure to marry
  • Final thoughts
• APPENDIX 1 A LAYMAN'S GUIDE TO STATISTICAL TECHNIQUES USED IN CHAPTER 7
  • One-way analysis of variance
  • Multiple regression analysis
• APPENDIX TABLES
• REFERENCES

FOREWORD

In 1970 the Department of Demography, Australian National University established the Australian Family Fonnation Project. This has since fonned the framework within which the Department's research on the family and fertility in Australia has taken place.

Since the beginning the Project has produced a series of significant books in the Australian Family Fonnation Project Monograph Series. The majority of these works have drawn upon survey and other field work of the Department.

The publication of this monograph by Dr. Gordon Cannichael follows the same pattern as did No. 9 by Dr. Christabel Young, in that it is a joint publication of both the Department of Demography and the Australian Institute of Family Studies. This time the research project was carried out by the· AIFS. Nevertheless, joint publication is appropriate as much of the analysis and writing was carried out at the ANU when Gordon was a Visiting Fellow in the Department of Demography.

It might be noted that Gordon was associated with the ANU Department of Demography before joining the AIFS and that he has now returned to the Department.

Once again, we are delighted that the two major institutions undertaking research on the Australian family have been able to cooperate in this way.

John C. Caldwell,
General Editor
Australian Family Fonnation Project Monograph Series

PREFACE

The study on which this report is based was the first major project of the then new Institute of Family Studies. Designed by the Director and research staff team, the project aimed at investigating what young people thought about marriage and family life, and what explained their actual behaviour in forming new families of their own.

Preferably, the study had to be longitudinal, so that we could follow through the national sample of young Australians to see, flrst, whether they did what they said they would do, and second, whether our theories about changes in family life would hold up. The first round of interviews took place in late 1981 and early 1982. Lack of adequate funding and pressure of other work meant delays to the follow-up study, but I am pleased to say that this will happen before the end of 1988. We still cannot afford face-to-face interviews, so a mailed questionnaire will have to sufflce. But the delay may have its advantages.

The decade of the 1980s has seen a consolidation of family trends emerging in the 1970s and our sample, now aged between 24 and 40, are the flag bearers of what could be called 'the new marriage'. On the whole, they left the parental home earlier, seeking independence rather than marriage; they often returned, then left again; they were more likely to live with a partner, marry later, and have children later than earlier generations; and larger numbers of young women were to retain their jobs and careers as a hedge against both the cost of housing and the poverty that might result from divorce. Reliance on a partner is less likely in the new marriage of autonomy and equal sharing, but the strains and inner conflicts generated by changing social standards and economic demands are evident in many marriages.

During the past six years we have seen major shifts in women's labour force participation, the birth rate, the marriage rate, and in areas such as family law and social security. Australia has experienced the shock of 'Banana Republic' predictions, tightened flscal restraints, and a new conservatism in relation to government spending and the nature of the welfare state. It is to be expected that this changed economic and policy climate will be reflected in the attitudes and behaviour of our cohort as they enter the third century of white occupation of this land. The delay in our follow-up thus gives it added potential to reflect changes in family formation patterns, so that the second wave study should provide a vivid picture of the sorts of new families that are being formed. Our interest is not only in marital timing or fertility, but in what 'the family' looks like, in its many diverse forms, as we enter the 1990s.

Gordon Carmichael's report is thus timely, despite the fact that we collected the data on which it is based some six years ago. For in it, he gives a most authoritative summing up of the place of marriage in Australian society throughout this century, noting the trends and the reasons for change, and setting the' Institute's Family Fonnation Study data fIrmly within that wider historical context. His careful analysis draws out the major patterns in the timing and thinking about marriage, while ensuring that the cohort and other group differences are also clarified.

This is the second volume based on the Family Fonnation Study data and published with the help of the Department of Demography, Australian National University. The fIrst, written by Christabel Young and titled 'Young People Leaving Home in Australia', was published in 1987. Gordon Carmichael has been on secondment from the Australian Institute of Family Studies for the last three years, working on the ANUs major family project. We are grateful to the Department of Demography for these important avenues of co-operation and for their fInancial assistance in publishing both monographs. The members of the AIFS survey team which designed and carried out the study - Don Edgar, Helen Glezer, Margaret Harrison, Christine Kilmartin, Gay Ochiltree, Andrew Prolisko, Don Stewart, Des Storer and Terry Tremayne - are all very pleased that the data have been so well analysed and reported in this way. Other papers based on the study are available from the AIFS.

Don Edgar,
September, 1988

Author's preface

AUTHOR'S PREFACE

First marriage patterns in Australia have been extremely volatile in recent decades. They have responded to the upheaval of the Second World War, immigration trends, changes in the acceptability of marital contraception and in contraceptive technology, the evolution of a more permissive sexual morality, liberalization of abortion laws, economic trends, and changes in a wide variety of attitudes. This monograph brings together data from several sources in an attempt to tell this story and enhance understanding of contemporary patterns of entry into marriage and failure to marry. Hopefully, it will assist readers to ponder the future of marriage in Australia on a more informed basis.

The study was commissioned by the Australian Institute of Family Studies, but much of the research was carried out in the Department of Demography at the Australian National University. Its publication has been jointly sponsored by the two institutions. In recognition of this joint sponsorship I have endeavoured in my writing to achieve a balance between the rigour demanded by an academic readership and the readability demanded by the intelligent lay readership which is a major target of the AIFS. Data analysed are drawn from three main sources - the 1981 Australian census, the national vital registration system which tabulates marriages on an annual basis, and the Australian Family Formation Survey, a survey of 2544 Australians aged 18-34 conducted in 1981-82 by the AIFS. These data sources have their limitations, but collectively provide interesting insights into ftrst marriage trends and patterns in Australia. The interpretation of ftndings is at times necessarily speculative, effectively advancing hypotheses which invite further debate and rigorous testing.

I am grateful to the Australian Institute of Family Studies for its ftnancial support of the project and to the Department of Demography, Australian National University for facilities made available whilst working there as a Visiting Fellow. Useful comments on different sections of the study were made by colleagues at both the AIFS and the ANU. All were appreciated, but particular mention should be made of the contributions of Paul Amato, Michael Bracher, Peter McDonald, Andrew Prolisko, Gigi San tow and Frank Zimmerman. Teresa Tucker did a lot of the computer work for me and Daphne Broers-Freeman handled the ftnal editing and publication with her usual thoroughness; my thanks to them. Naturally, I accept sole responsibility for whatever flaws my work contains.

Gordon Carmichael

CHAPTER 1 INTRODUCTION

Virtually all Australians are born into families and grow up in families, albeit that membership of these families periodically may change, sometimes quite drastically. Eventually most form families of their own through the processes of partnering and parenting. These processes may occur alone in an individual's life cycle or each may precede the other, but historically partnering followed by parenting has been the standard and approved pattern. Historically, too, the taking of a partner usually has been marked by a formal marriage ceremony. Certainly informal cohabitation was common among the early white Australians (McDonald, 1974; Sweeney, 1981), and through the 1970s living together as a prelude if not an alternative to formal marriage became increasingly popular. But that marriage and partnering, and hence marriage and family formation have been closely associated throughout most of the period of European settlement in Australia is indisputable.

It follows that an understanding of patterns of entry into fIrst marriages - the ages at which people marry, the personal attributes which distinguish those who marry at different ages, the extent to which birth cohorts (groups of people born in the same year) ever marry, and the characteristics which set those who marry apart from those who never marry - together with an understanding of changes in these patterns is central to an appreciation of the changing process of family formation in Australia. This monograph pursues these themes. It concentrates on the period since about 1940, and within that period devotes particular attention to more recent trends, relationships and differentials. These naturally are the most interesting ones, and the basis on which any assessment of the future of formal marriage in Australia should be made.

ORGANIZATION OF THE STUDY

The study is in two parts. It begins with an analysis of trends in age at fIrst marriage and proportions of the population ever marrying. Comprehensive marriage registration data are available from 1920; hence Chapter 2 focuses on the period 1920-70 during which Australia experienced a marked, but temporary slump in fIrst marriage activity associated with the 1930s depression, and then a marriage 'boom' which saw the percentages of young men and women ever marrying rise to unprecedented levels and the ages at which they married rapidly fall. Since the early 1970s these two trends have reversed sharply. Chapter 3 describes this phenomenon, discusses likely reasons for it, and speculates on what may lie ahead.

Having examined changes in broad age patterns of first marriage and in the propensity to marry at all, attention is switched to the fact that some members of any birth cohort marry earlier than do others and to ascertaining some of the characteristics of those individuals who never marry. The monograph at this point abandons its temporal focus and adopts a strictly contemporary, cross-sectional one. What distinguishes early marriers (i.e. persons who are relatively young when they marry) from later marriers? Numerous previous studies, few of them Australian, have addressed this question and these are reviewed. Three chapters then examine what light the available data shed on the determinants of marriage timing in Australia. Official sources (census and vital registration data) permit the influence of premarital pregnancy, education, occupation, birthplace and, more tenuously, religion and place of residence on marriage timing to be assessed. The unreliability of census marriage data for Aboriginals, however, precludes a satisfactory analysis of the relationship between race and age at flrst marriage.

Census data have the advantage that, because they derive from a full enumeration of the population, they allow one to focus on minority groups that often are not sufficiently well represented in surveys for meaningful conclusions about them to be drawn (minority birthplace and religious groups, for example). Relevant survey data, on the other hand, normally provide a richer variety of variables to work with, enabling a phenomenon to be investigated in greater depth. Thus it is that Chapters 6 and 7 explore marriage timing differentials using data from a national survey of 2544 young people which the Australian Institute of Family Studies (AIFS) conducted in 1981-82. The information from this survey is not always as complete or as precise as might be wished; the survey was not designed with a study of marriage timing differences in mind. Speciflc shortcomings, however, are not so severe as to prevent the data from providing some useful insights into the forces determining the ages at which Australians marry.

Two broad approaches are taken to analysis of the survey data. First, survival analysis is used to construct graphs, 'flrst marriage curves', which trace the probabilities of different subgroups of the AIFS sample remaining unmarried at successive birthdays. Comparisons of these graphs, drawn separately for males and females, permit identiflcation of subgroups which have relatively early and relatively late age patterns of flrst marriage. The focus in this exercise is on individual variables which might influence marriage timing; that is, the marriage patterns of different education subgroups are compared, those of religious subgroups are compared, and so on. Subsequently a multivariate approach is adopted. Having ascertained variables seeming individually to bear some relation to the ages at which survey respondents had married, the questions of which variables are the more important ones and how they flt together and influence one another in the marriage timing process are addressed.

The flnal substantive chapter, Chapter 8, concentrates on differentiating between persons who marry and those who never marry. It flrst discusses the reasons given for having married by ever married respondents to the AIFS Survey and then analyses the limited data available from the census concerning subgroups of the population that are particularly likely never to marry. It concludes by examining the marriage intentions of never married respondents to the AIFS Survey. These unmarried 18-34 year-olds are divided into those who envisaged marrying at some stage and those who thought they would never marry. Comparisons of the background characteristics, dating and sexual behaviour during adolescence, and attitudes of these two groups then are made in an effort to understand more about those young Australians who currently seem to be consciously avoiding formal marriage.

HISTORICAL BACKDROP: MARRIAGE FROM THE CONVICT ERA TO THE EARLY TWENTIETH CENTURY

It is appropriate at this point to briefly outline what is known of fIrst marriage patterns among European Australians prior to 1920. In the early years after the First Fleet landed in 1788 a measure of order was maintained over matrimony (McDonald, 1974). Governor PhiIlip, appreciating the potential of marriage for promoting stability, actively encouraged it and supported the one clergyman available to conduct marriage ceremonies, the Reverend Richard Johnson. Naturally, however, transported convicts often were of dubious character. They were .overwhelmingly drawn from the poorer classes, among whom it was not uncommon to dispense with the fonnalities of marriage. In addition the English Marriage Act of 1753 which was assumed to apply in the Australian colonies was of little significance to Scots and Irish convicts. Thus, as the population increased, control over marriage was soon lost.

In the first decade of the nineteenth century infonnal cohabitation was widespread. Governor King in 1806 reported that a mere 28 per cent of adult women in the Colony of New South Wales were married, and that most of the remainder were living in de facto unions. Subsequent. governors from Macquarie onward saw the regularization of marriage as one of their highest priorities, Macquarie registering his disapproval of 'the scandalous and pernicious custom so generally and shamelessly adopted throughout the territory of persons of different sexes cohabiting and living together un sanctified by the legalities of matrimony' (cited in Sarantakos, 1984: 14). By 1860 this objective had been substantially achieved. Transportation had ceased and the nature of the immigrant inflow had changed. Domestic marriage laws had been enacted in all colonies, the supply of marriage celebrants had improved, and both civil marriages and marriages by churches other than the Church of England had been sanctioned.¹

Early marriage patterns in Australia were the product of several forces, but none was more important than the substantial excess of males over females in the population. The sex ratio among convict arrivals varied, but overall more than six males were transported for every female (Robson, 1965). Sex ratios for the combined free and convict adult populations were lower, but there were still some 3-4 males per female in New South Wales throughout the period 1800-36. Under those circumstances it was inevitable that a significantly higher proportion of females than of males married, or indeed lived in any sort of marriage-like union.

The adult sex imbalance lessened during 1832-51 as free settlement began in earnest. Almost three-quarters of the 93 thousand free settlers aged 14 and over arriving in New South Wales during this period came on assisted passages, and although males were marginally more numerous in the total 'adult' migrant intake, females predominated among the assisted arrivals. Things continued to improve in all colonies during the 1850s, except in Victoria where the discovery of gold caused the sex imbalance to worsen early in the decade. By mid-decade there was close to one single adult male for every single adult female in Sydney, Adelaide, and Hobart, although single males continued to greatly outnumber single females in the rural areas of all colonies (McDonald, 1974: Table 11). This greater scarcity of marriageable females in the bush had developed with the rapid settlement of the interior of Eastern Australia during the 1830s and 1840s.

Discussing the period 1861-91 McDonald (1974) identifies three dis.tinct patterns of marriage in Australia. The first was one dominated by the level of the balance of the sexes. It was the early experience of all colonies except South Australia, and was characterized initially by the sex differential in proportions marrying already noted. Then, as the excess of marriageable males diminished, the proportion of males marrying tended to remain stable whilst the proportion of females marrying rapidly declined. This process was completed first by Victoria during the 1860s and 1870s as net out-migration of males and in-migration of females rapidly reduced the sex ratio. In New South Wales and Queensland it was still in progress by 1891, these colonies having continued to receive significant male-dominated net immigration.

The other two patterns McDonald identifies are the 'South Australian' and the 'later age at marriage' patterns. The form,er reflected the relatively even sex balance which was a feature of the family-oriented settlement of South Australia from the outset. Because of this the proportion of males married was much higher than in other colonies and females tended to marry quite early. The later age at marriage pattern was marked by a distinct deferment of marriage by both sexes, but especially by males. It prevailed in Victoria in both 1881 and 1891, and had also appeared in South Australia by the latter date. In McDonald's view it emerged for two broad reasons. First, the ethic of a 'proper time to marry' that developed among the English middle and upper classes during the 1850s and 60s was transferred to Australia by migrants attracted by the growth of the government, commercial, and fmancial sectors. To some extent it was even imposed by employers who forbade their employees marrying until specified career milestones had been reached. Second, prosperity and confidence in the future were such in Victoria in the 1880s that even working class men were influenced to delay marriage by the real prospect of being able to save sufficient to become home owners. Caldwell and Ruzicka (1978: 84) are adamant that 'Contemporary references, in literature, in the press, and in reminiscences, make it quite clear that marriage was deferred [in the third quarter of the nineteenth century] because of the costs inherent in marriage itself'.

Whereas before 1891 colonial marriage rates had changed only slowly, during 1891- 1921 there were some substantial movements over short periods. This 30 year span saw the proportions of females and males never married at ages 45-49 respectively rise from 6.5 to 16.6 per cent and fall from 23.6 to 19.7 per cent (Spencer, 1969b). The trend for males is confounded by improvements in the availability of marriage partners, but that for females points to the severe consequences which the depression of the 1890s had for the marriage prospects of birth cohorts then of prime marrying age. In all States except Western Australia (where gold had been discovered) marriage rates declined rapidly for both sexes between 1891 and 1901, partially recovered during 1901-11, and then rose again between 1911 and 1921.

The 1890s depression lasted longer than that of the 1930s and resulted to a much greater extent in permanent, as compared to temporary, postponement of marriage. It had an impact in both urban and rural areas and, perhaps because of the stark contrast it provided with the buoyancy of the 1880s, resulted in an especially high level of permanent postponement in Victoria. There, in 1921,21.1 per cent of females aged 45-49 had never married compared to figures for other states ranging from 8.4 per cent in Western Australia to 17.3 per cent in South Australia (McDonald, 1974: Table 40). In 1891 a mere 7.4 per cent of Victorian women aged 45-49 had never married. This downturn in first marriage activity was a major factor in the decline of the birth rate which caused considerable consternation around the turn of the century (Ruzicka and Caldwell, 1977;· Hicks, 1978). A further recession occurring during 1902-04 may help explain why marriage rates initially recovered only slowly. McDonald also speculates (1974: 149) that 'the long period of low marriage rates may have caused some kind of breakdown of the existing facilities for courtship and marriage', citing the 'extraordinary success' of commercial matrimonial agencies that emerged in Sydney and Melbourne in the mid-1890s.

During 1901-21 percentages never married at ages 25-29 declined for both sexes in all states except Western Australia, where female percentages already were low due to the heavy excess of adult males. Declines for males were especially sharp, amounting to 7.4 percentage points in Queensland and ranging between 12.1 and 14.8 percentage points in the other five states. They were also especially interesting because previously this index had remained surprisingly stable in the face of considerable improvement in the balance of the sexes. The main mechanism responsible for these declines appears to have been an increase in marriage rates at all ages, with decline in the ages at which people married of distinctly secondary "importance. In the absence of comprehensive marriage registration data temporary setbacks to marriage, most notably during World War I, tend to obscure any long-term trend in age at first marriage in the early twentieth century. However, having risen considerably during the 1890s, ages at marriage probably began falling slowly for both sexes from some time around 1910.

In general, early Australian marriage patterns appear to have derived substantially from those prevailing in England and Wales (Jones, 1971; McDonald, 1974, 1982). This statement holds particularly with respect to age at marriage. With respect to proportions marrying Australia's population displayed a relative conservatism. Even after the sex ratio evened out as the first large generations of native-born white Australians came of age in the 1870s and 1880s, significant numbers of Australian men never married. As a corollary of this the proportion of lifetime spinsters increased. There are good reasons why this should have been so. Males still heavily outnumbered females in rural areas, where insecurity of employment and the failure of most schemes aimed at promoting family settlement did nothing to enhance marriage prospects. It is possible, too, that a certain cult of bachelorhood developed during the years when marriage partners were hard to find. Another factor may have been the sizeable proportion of Australia's European population who were Irish, the Irish being renowned for being diffident towards marriage. Then there was depression in the 1890s. This not only affected the economic feasibility of marriage but coincided with and hastened the movement of substantial numbers of single men to Western Australia's goldfields.

CHAPTER 9 SUMMARY AND CONCLUSION

The subject of marriage is one that readily captures people's imaginations. In recent years, as indeed throughout Australia's 200 years of European settlement, there has been much that has been highly emotive spoken and written about what is happening to the family and to the institution of marriage which underpins it. This monograph has sought to enhance understanding of one aspect of marriage, entry into the married state, through objective analysis of relevant vital registration, census and survey data. A knowledge of patterns of entry into marriage and how they have changed is essential to understanding the c\langing process of family formation in Australia. It also is essential to achieving a balanced appreciation of recent divorce trends. The significance to the rapid rise in Australia's divorce rate during the 1970s of the unprecedentedly young ages at which marriages took place in the 1960s, the carefree way in which decisions to marry were made at that time, the types of individuals who were marrying young, and "the fact that marriage had become well nigh universal often is overlooked.

TRENDS IN AGE AT FIRST MARRIAGE AND PROPORTIONS MARRYING

After informal cohabitation had been widespread in the early convict years and massive shortages of marriageable females had dominated marriage patterns through much of the nineteenth century, Australians by the early decades of the present century were marrying at ages dictated largely by considerations of economic propriety and preparedness, with significant minorities of both sexes failing to marry at all. The marriage boom extended from !lpproximately the outbreak of World War IT through until the early 1970s. During this period ages at first marriage declined appreciably, their range was compressed, and marriage became a much more universal experience. A good deal of groundwork for these trends already had been laid by the time World War IT began. Parent-child and husband-wife relations had become gradually more egalitarian, generational identity among the young had strengthened, and marital contraception had become more widely known about, approved of, and practised. The War itself set the trends decisively in motion. With the impending departure of troops for overseas many young couples hastened to marry, anxious not to be deprived of the experience and able to disregard many of its usual economic implications.

The pace of change following World War II differed by sex. Heavy immigration of young single persons during the late 1940s and the 1950s was dominated by males so that men faced a severe shortage of potential marriage partners. Then, with postwar baby boom children entering the marriage market in the mid-1960s, it was the turn of females to be 'squeezed' as they competed for husbands from smaller wartime birth cohorts. These circumstances affected the comparative speed and timing of further declines in age at marriage for the two sexes, and prevented marriage from ever becoming quite as universal among men as it became among women.

Essentially, though, the marriage boom continued. Wartime experiences coming on top of childhoods affected by the Depression were decisive forces for generational independence and for questioning notions of a predictable life course within which there was a proper time to marry. During the 1950s earlier marriage became a popular vehicle for expressing independence. The right to practise contraception within marriage was by now taken even more for granted, familistic ideals were widely promoted and professed, and economic circumstances were favourable. Increased freedom of association with the opposite sex also led to a marked increase in early marriages involving premarital pregnancies.

The marriage boom might have petered out but for the advent in the early 1960s of oral contraception at a time when young Australians faced increasing peer pressure to be sexually active. The pill was to all intents and purposes a failsafe method of preventing pregnancy, and some young people seized the opportunity it offered for marrying sooner without committing oneself to immediate parenthood. To some, faced with conflicting pressures to be sexually active yet premaritally chaste, it was a godsend. Anyway, through the late 1960s and early 1970s there was a surge in marriages in~olving non-pregnant brides at younger ages.

At the same time as marriage ages in Australia were falling to their lowest levels ever, some of the forces that were to emphatically reverse the trend were gathering momentum. Among other things, marriage was becoming less the catalyst to a range of major life cycle transitions - leaving home, becoming regularly sexually active and becoming a parent After 1971, fIrst marriages at younger ages in particular declined rapidly. Ages at marriage quickly rose, and by the mid-1980s it seemed possible that a fifth to a quarter of young people born in the 1960s might never marry. The new trends were triggered, though not caused, by the freeing up of access to induced abortion; early on it was youthful marriages involving pregnant brides that declined sharply. Changes to State abortion laws or their interpretation were, however, brought about by feminist and other forces which, in the wake of rising levels of unmarried motherhood and 'shotgun' marriage, demanded greater control for the young over family formation. These were among the real underlying causes of the change in fIrst marriage behavior.

Living together became more acceptable and more widely practised, eroding further the status of marriage as a pivotal event in the life cycle. A trend in this direction was inevitable once the pill had in the interval from marriage to first birth widen considerably through the 1960s, and a tide of liberal philosophy which flowed across many issues either side of 1970 helped to establish it. Economic circumstances during the 1970s ':Vere also conducive to delaying if not bypassing marriage. Youth and young adult unemployment rates rose alarmingly, and it became much more diffIcult for young couples to acquire their own homes. Getting educated now took longer, and this, too, placed upward pressure on ages at marriage. Underpinning everything, however, were ideological changes that occurred.

A new individualism pervaded and, for many, marriage was not to be considered until experiences available to single young adults had been sampled fully. Courtship became a more cautious, even tentative process, and the religious basis of norms traditionally governing it was widely rejected. In particular, the adverse consequences for young people of the sexual revolution led to both a reassessment of the principles which should govern decisions to marry and the psychological separation of marriage from sexual expression. Young men perhaps became more loathe to commit themselves to married life, and a rapidly rising divorce rate provided additional incentive to be circumspect in the marriage market. But above all, young women's attitudes and priorities changed. Prompted by a surge of feminism founded on rising levels of married female employment and advances in contraceptive technology, they became more career conscious, more determined to achieve academically, more concerned with sharing and equality within marriage, and less inclined to see motherhood as their ultimate calling in life. As a result they abandoned the carefree approach to marriage of their sisters in the 1960s, looked more to alternative sources of satisfaction, and exercised much greater care when selecting their husbands.

DETERMINANTS OF AGE AT FIRST MARRIAGE

Vital registration data showed that despite a sharp decline in the incidence of bridal pregnancy through the 1970s, by the mid-1980s very youthful marriages remained especially likely to involve premarital pregnancy. Census data then indicated that the earlier a young person had left school the earlier he or she was likely to have married, with really late leaving females seeming especially prone to not marrying at all. Men with educational qualifications and women without them were the more likely to have married in age groups 20-24, 25-29, and 30-34, but among the qualified the university educated stood out as very late marriers while males with trade qualifications seemed to be relatively early marriers.

Analysis of occupational differences in marriage timing was an exercise full of potential pitfalls. Male professional and clerical workers, whose jobs often required that they be well educated, were relatively late marriers; members of the armed forces, small businessmen (especially, perhaps, skilled tradesmen), miners, and transport workers had married at younger ages. Among women analytical difficulties arose out of, among other things, a tendency to enter or avoid some occupations as a consequence of marriage. Again, however, professional workers stood out as very slow to marry. More detailed investigation of male occupational groups showed clergymen and artists, entertainers, etc. to be slower to marry than other professionally employed people. Among salesmen, commercial travellers were early marriers, although this occupation might have special appeal to married men just as others, such as well drilling, clearly do not. Aircraft pilots married very late compared to other transport workers, and it was speculated that occupational differences in marriage timing among 'Tradesmen, production process workers and labourers' might have owed something to personality differences. The broad 'Service, sport and recreation workers' grouping included occupations with both the highest (policemen and firemen) and lowest (waiters and bartenders) percentages of men ever married in their twenties and early thirties. Generally speaking the forces responsible for occupational differences in marriage timing seemed complex; certainly they could not uniformly be attributed to characteristics of occupations directly influencing marriage timing decisions.

Some marked variations in marriage timing by country of birth were established. Topping the list of female early marriers were the heavily Muslim Turks and Lebanese, followed by Cypriots and the various Southern European groups. Filipino females aged 20-24, too, were especially likely to have married, many having come to Australia for that express purpose. Among males, Turks again, Portuguese, and then Dutch, Maltese, Lebanese, and Cypriots were the earliest marriers. Identifying groups who married relatively late compared to native Australians was more problematic. Groups with comparatively low percentages ever married in their twenties frequently included numerous temporary residents whose reasons for being in Australia predisposed them to being single

Because of their numerical significance, Southern' Europeans were accorded particular attention. Two earlier studies had shown that Australian Southern Europeans had married considerably earlier than was nonnal in their countries of birth. Among explanations offered for this were selectivity in the migration process, the capacity of impending migration to bring about early marriage, sex imbalances in ethnic marriage markets, and attempts to circumvent the Greek dowry system. Evidence indicated that marriage ages had risen lately among Southern European females, but that they still married appreciably earlier than native-born women. Advocacy of traditional family values and the interest parents take in courtship provide strong support for this differential. Another support, although one that has weakened dramatically, has been lower levels of education among Southern European immigrants. Southern Europeans raised in Australia seem to have patronized the education system very enthusiastically. This is almost certainly a major reason why ages at marriage of Southern European women have risen, and doubtless also helps explain why second generation Southern European females appear to be marrying later than the first generation and more in conformity with native Australians.

Census data are not ideal for studying religious differences in marriage timing. They show 'Christians' to have married more readily than persons claiming no religion, but earlier marriage among members of some smaller sects than among members of larger denominations may reflect lower levels of nominal adherence. Jehovah's Witnesses are extremely early marriers, with members of the Brethren Assemblies and Salvation Anny, Pentecostalists, Seventh Day Adventists, and Orthodox females similarly inclined. The data were too crude to determine whether Catholics marry later than Protestants in Australia. Muslims, male Jehovah's Witnesses apart, were the earliest marrying of all religious groups whilst Jews, probably due to a strong commitment to tertiary education, married especially late.

Census data seemed to indicate later marriage in capital cities than elsewhere in Australia and comparatively early marriage in Tasmania. It was impossible to tell, however, how far these findings were products of upbringing and how far they were artifacts of migration processes. Deficiencies in data pertaining to Aboriginal marital statuses and widely different cultural perspectives on marriage precluded any meaningful comparison of marriage timing among Aboriginal and non-Aboriginal Australians.

Data from the AIFS's 1981-82 Australian Family Fonnation Survey offered th~ prospect of delving more deeply into the detenninants of marriage timing in Australia. Survival analysis showed sons of professional and managerial workers to be later marriers than sons of each of three other groups of workers, with sons of skilled tradesmen also marrying later than those of unskilled or semi-skilled workers. Daughters, too, of professional and managerial fathers had married slowly, although not that much more slowly than daughters of lower level white collar workers. Both sexes, but especially females, had married sooner if their fathers had left school before age 16; both had also married earlier if their mothers had left school early. Persons whose fathers were university educated were relatively late marriers, as were sons of university educated mothers. Amongst females, however, having a mother with any sort of post-school qualification was associated with later marriage.

Slightly later marriage was found among Catholic than among non-Catholic Christian females, although this difference disappeared when only regular churchgoers were compared. Much more striking was the comparative reluctance to marry of persons who claimed to have no religion. Non-churchgoers seemed to have married more freely at younger ages than regular churchgoers, but then to have married m6re slowly through their middle and late twenties. It was speculated that non-churchgoers had, over the roughly 15 year period encompassed by the analysis, been both more prone to marry early in response to premarital pregnancy and more disposed to defer marriage by cohabiting informally. Some evidence was found, more particularly for females, that having been raised in a non-religious home was associated with earlier marriage, while females whose mothers only were· regular chiJrchgoers also had married sooner than daughters of two regular churchgoers.

The critical distinction for marriage timing so far as residential location during childhood was concerned was whether childhood had been spent in a large city. A metropolitan upbringing was associated with later marriage, especially among females. No evidence was found to link parental separation during childhood with marriage timing, but there was a suggestion, deserving of further study, that women who have a parent die during childhood sometimes postpone marrying. Also among females, having grown up in a large family (5 or more siblings) seemed to make for earlier marriage, perhaps reflecting limited educational opportunities and a desire to escape chaotic if not oppressive domestic environments. No support was found for the proposition that fIrst-born children are apt to marry early because parents are particularly demanding of them.

Whether or not females considered it had been 'pleasant' living with their parents was unrelated to marriage timing. Males, though, who did not consider life at home was pleasant had both married quite early in some cases and been rather reluctant to marry in others. A crucial aspect of home environment affecting female ages at marriage appeared to be dissatisfaction over demands made for domestic assistance; women who felt imposed upon tended to have married earlier. So did women who felt their parents had hindered their social lives, the failure of a similar fInding to emerge for males perhaps saying something about the nature of parental interference in the activities of sons and daughters. Despite contrary fIndings in some overseas research, no impact of maternal employment on marriage timing could be detected.

Three dimensions of quality of parents' relationship were examined in relation to marriage timing. Assessed ability of parents to communicate was not important, but males who considered that their parents lacked affection for each other had in a few cases married very early but more commonly were rather tentative about marrying. Among females, whether parents were felt to have been too independent or insuffIciently independent was associated with earlier and later marriage respectively.

Data on adolescent dating and sexual behaviour were flawed. Members of both sexes, but especially females, who had dated seriously at ages 15-17 and 18 had married very young. This fInding is entirely plausible, but its strength may have been exaggerated because assessments of dating activity as 'serious' or 'casual' were made retrospectively. It was only possible to study the relationship between premarital sexual experience and age at marriage for males. Some evidence was found of later marriage among men who had reached age 19 without having premarital intercourse, but this same group also seemed more committed to marrying at some stage in their lives.

Commonsense says that attitudes to marriage, alternate lifestyles, and motherhood could affect marriage timing. These relationships should, however, be studied with longitudinal data so as to control for attitude change subsequent to marriage or in response to difficulties encountered in the marriage market Such an approach was not possible here, and consequently findings must be interpreted cautiously. Both sexes were much slower to marry, and much more likely to remain unmarried into their thirties, if agreeing that 'Marriage is not important to me but having a close personal relationship is'. Both also had married more slowly if they rejected the idea that motherhood is a woman's preeminent role. Appreciating the opportunities single status offers went with later age patterns of first marriage. Finally, acceptance of trial marriage and of the principle of testing sexual compatibility before marrying both also were predictive of later marriage. These relationships, probably weakened by some early marriers having subsequently adopted more liberal views, strongly suggest that actual experience of living together delays marriage.

Survival analyses based on how many of nine nominated qualities respondents felt a partner should have broadly confirmed for both sexes that the more choosy, or fussy, a person was the later marriage tended to have occurred. Further investigation indicated that the 'not very choosy' were inclined to emphasize nebulous qualities having to do with interpersonal relations and values at the expense of more measureable qualities related to social origins and socioeconomic status.

Males who had had 5-6 years of secondary schooling had married later than those with 'up to 2' or 3-4 years of secondary education, while among females the tendency for more schooling to mean later marriage extended across all three categories. Splitting of the '5-6 years' category according to whether or not full-time study was engaged in within two years of leaving school indicated that undertaking tertiary education led to the greatest deferment of marriage, although those who did not study further still married later than persons with 3-4 years of secondary schooling. Closely linked with these findings were findings based on the type of secondary education received. Both sexes had married later if educated at a non-Catholic private school than if educated at a Government school. The age pattern of first marriage for Catholic educated females tended toward that for Government school attenders, while that for Catholic educated males was an amalgam of the other two patterns. Predictably, having attended a single sex school was associated with later marriage, although more strongly for females than for males. Finally, females, but not males, had married later if they perceived their parents as having encouraged them to acquire education.

Having dealt individually with variables which might affect marriage timing, the obvious final step was to focus on their relative importance and how they fitted together. Preliminary analyses of variance reduced the number of 'important' variables to 16 and helped refine some variables. A general model of the marriage timing process then was constructed and initial tests of it were undertaken using multiple regression analysis. These produced quite a simple model for males which emphasized later marriage the higher the level of education attained as overwhelmingly the strongest determinant of age at first marriage. Other statistically significant variables associated later marriage with a favourable disposition toward trial marriage (and presumably with actual experience of premarital cohabitation) and with being choosy in mate selection, and earlier marriage with premarital pregnancy having influenced the decision to marry.

Three of these four variables also were part of the more complex model derived for females, the odd one out being 'Choosiness in selecting a partner'. Education was not, however, the dominant force in the female model. That status was assumed by a variable indicating earlier marriage where regular dating began before age 18, although the education effect was barely less important. Three other 'new' variables featured in the female model. Ranking about equal third in importance with the trial marriage variable was 'Parents' independence', pointing to later marriage where parents lacked independence and earlier marriage where they were felt to have been too independent. Here the former element might have been ideological to a point, while the latter could have captured defective father-daughter relations as well as sterile father-mother ones. Other components of the female model represented later marriage where single life was especially valued and earlier marriage where there had been frequent conflict with parents over domestic jobs.

Refinement of these regression models using path analysis preserved them, but allowed attention to be paid to variables antecedent to those directly affecting marriage timing. For males, level of education mediated the strongest antecedent effects as well as having the strongest direct effect on age at marriage. Status of father's occupation was an important determinant of level of education attained, both directly and via parental encouragement of education. Much the same could be said for females, although here attendance of a non-Catholic private school was also prominent in the complex of variables antecedent to level of education. Parental encouragement of education owed something to having grown up in a smaller family in the case of males, and to having had a metropolitan upbringing in the case of females.

The male refined model indicated that both acceptance of trial marriage and choosiness in partner selection depended partly on parents' affection for each other being· assessed as below par. It was speculated that the nature of the parental relationship might have caused marriage to be approached cautiously by generating doubt over one's ability to give affection as well as over the likelihood of receiving it. Among females acceptance of trial marriage showed a prior dependence on a metropolitan upbringing which perhaps reflected conservatism in non-metropolitan Australia. Above all else, however, the female model showed early dating activity occupying a crucial role in the marriage timing process, whereas in the male model it was totally unimportant. Women were significantly more likely to have dated regularly before age 18 the lower the status of their fathers' occupations. Early dating then had predisposed them to marrying early by placing them in the marriage market early, by exposing them to the risk of premarital pregnancy, and possibly by reducing their ability to concentrate on schooling. Males, by contrast, had treated early dating much more as recreation, and none of these other linkages was evident.

Family size was a more prominent background variable for females. It impacted directly on education level independently of parental socioeconomic status, coming from a larger family tending to reduce the education level attained. It also was antecedent to the direct effect of conflict over household jobs on marriage timing and to a near-significant direct effect associating later marriage with having had a very heavy domestic workload at home. Both experiences had been more common in larger families.

Recognizing that 30-34 year-olds in the AlPS sample mainly had married before the recent downturn in first marriage activity in Australia, a comparative analysis of marriage timing among females aged 25-29 and 30-34 who had married by age 25 was undertaken. This aimed to check, as best one could, on major changes in the determinants of marriage timing during the 1970s. The main conclusion reached was that education level had supplanted early commencement of regular dating as the main force affecting marriage timing among females. Apparently early dating now less often represents entry into the marriage market, more often leads in the short term to informal cohabitation, and is treated more as recreation. Pregnancies resulting from early dating less often become catalysts to marriage. Maximizing educational achievement and ensuring that, once acquired, education is used to advantage in the labour market have, on the other hand, assumed a higher priority for young women, strengthening the relationship between education level and age at marriage.

REASONS FOR MARRYING AND FACTORS ASSOCIATED WITH FAILURE TO MARRY

Reasons for marrying offered by ever married respondents to the AIFS Survey emphasized 'love', but perhaps not as universally as might have been expected. Failure to mention this motive was most common where marriage had taken place at a very young age (where pregnancy frequently was implicated) or at ages 25-34. Some later marriers would have cohabited before marrying, and in consequence may have approached marriage less idealistically. Certainly they more often saw their marriages as the next steps in relationships and in terms of impending parenthood than did persons married at more normative ages. Males, especially those married at older ages, were most likely to have married because of boredom with single life and as a way of settling down; females married at younger ages were the most likely to have married to escape from home.

Census data showed no association for males between age on leaving school and the likelihood of still being single at ages 35-49. Among females, however, later schoolleavers were more likely not to have married. Unqualified males and qualified females were especially prone to being still unmarried at ages 35-49, the former probably because they competed less successfully in the marriage market and the latter because careers got in the way of marriage. Among qualified persons, females alone displayed wide variations in propensities to have never married by type of qualification. A fifth to a quarter of 35-49 year-olds with higher degrees had not married compared to just 2-3 per cent of those with trade certificates.

Despite marrying quite late, male professional and technical workers were not noticeably prone to never marrying. By contrast, very high percentages of female professional and technical workers aged 35-49 had not married. Male clerical workers had fared worse in the marriage stakes than any other broad occupational group, possibly because of personality factors and the suitability of clerical jobs for disabled persons. Remote work locations might have largely explained a slight tendency for men employed in primary production to more often be bachelors, although certain types of men could have been attracted to such jobs and not to marriage. At the other extreme, small businessmen, salesmen, and members of the armed forces had married almost universally, as had female sales workers.

Breaking down broad occupational categories, clergymen and members of religious orders stood out, with male artists, entertainers, etc. and nurses, most female professional groups and senior public servants, and policewomen also especially likely not to have married. Farm workers were more often still single at ages 35-49 than farmers and farm managers, while frequent absences from home seemed to make for a high level of bachelorhood among seamen. Several other occupations with high levels of bachelorhood were rather menial, seeming likely to have attracted disproportionately men whose personalities limited tQeir marriage chances. By contrast very low levels of singlehood were recorded for medical doctors and dentists, senior government officials, insurance and real estate salesmen, commercial travellers, aircraft pilots and officers, and policemen and firemen.

A willingness to intertilarry meant that British migrants (except the Irish) were less likely than native-born persons in their forties not to have married. Dutch migrants and German females also had rarely failed to marry. A particularly high percentage of Scandinavian males (including Finns) were still single, whilst with the notable exception of Spanish males all Southern Europeap birthplace groups recorded very low percentages never married beyond age 35. So did the largely ethnically Greek Cypriot community, the Turks, and the Lebanese.

Christians more often than non-Christians aged 35-54 were still unmarried in 1981, although persons professing no religion had been the most reluctant to marry. Catholic males stood out as including relatively high percentages never married at these ages, while among Christian females it was members of the Brethren Assemblies who, in stark contrast to Brethren males, most often had failed to marry. The celibacy demanded of Catholic clergy doubtless helps explain the former finding while the latter probably reflects very sheltered upbringings and a: restricted marriage market Outstanding for their very low levels of non-marriage were Iehovah's Witnesses, members of the Orthodox churches, and, among non-Christians, Muslims.

With a view to advancing understanding of the minority of young Australians who today are avoiding marrying, small groups' of never married male and female respondents to the AIFS Survey who saw themselves as unlikely to marry were compared with larger groups who thought that they would marry. Intending never to marry was unrelated to level of schooling completed, but was associated with coming from a poorer blue collar background, with coming from a broken home, with having had an unpleasant or unhappy home life, and with being not religious. Non-marriers subscribed less widely to traditional notions about marriage, were critical of its current state, more often saw it as threatening their autonomy, were less troubled by the prospect of remaining single long-term, and were more enthusiastic about informal cohabitation. Perhaps crucially they also displayed much less interest in parenthood, supplementing this with higher approval of unconventional settings for childrearing and a less traditional view of childrearing priorities. They were less committed to sexual exclusiveness, .more resigned to relationships breaking down, and more inclined than marriers to hold strong feminist views.

FINAL THOUGHTS

If one thing has stood out through this monograph it has been the theme of constant change in first marriage behaviQur in Australia. Against the evidence in that direction it is a brave person, perhaps, who dares to forecast the future. Nevertheless, the 1980s could prove to be a decade when the many changes set in motion during the 1970s will run their course to the point where a new equilibrium in Australian marriage patterns is established. There seems no basis for anticipating another marriage boom. Economic conditions might improve, but the ideological trends of the 1970s were so profound that the most plausible scenario is that considerably later marriage and a much higher level of permanent failure to marry will prevail for the forseeable future. In many respects Australia appears to be witnessing a return to marriage patterns of earlier this century, although the forces responsible for those patterns are quite different now from what they were then. To future demographic historians even more than to those of the present day it is the very early, almost universal marriage to which the marriage boom gave rise in the 1950s, 60s and early 70s that is likely to seem atypical. At present the recency of this pattern gives it an aura of historical normality, but ultimately it probably will be seen as a phase which marked the demise of one ideology concerning marriage preparatory to its replacement by a new one. For a demographer the interesting questions that remain concern just how late the normal marrying age eventually will be and how popular formal marriage will remain once present trends are fully complete.

In a very real sense, of course, those trends have been at least partly illusory. Couples have been entering cohabiting unions earlier and in larger numbers than statistics of formal marriages would indicate as the phenomenon of living together has gained in popularity and social acceptance. Considerably more research effort needs to be devoted to understanding this phenomenon and its relationship with formal marriage patterns, trends, and decision making. It is not uncommon for cohabiting unions to be referred to as 'de facto marriages', but this can be a misnomer. It is too simplistic to equate entry into such a union with 'partnering' in the sense that has meaning in the context of notions of 'family formation', and that is the context in which demographers traditionally have been interested in marriage. Marriage, it is assumed, presupposes a level of commitment that signals both an intent that a relationship should endure and, in the majority of cases, an intent sooner or later to procreate. That level of commitment mayor may not attach to informal cohabiting relationships. They may be substitutes for marriage, but equally they may be substitutes for the 'experimental dating', 'going steady' or 'engagement' phases of the traditional courtship process. Moreover, their statuses in these terms may change in the course of a relationship and may be perceived differently by the parties to a relationship at a given point in time. Assessing the impact of more widespread informal cohabitation on trends in formal marriage is not, then, simply a matter of . equating starting to live together with getting married.

Family formation is about partnering and parenting. What really is the role of the cohabiting union in this process? Is it changing, and if so, how? Is it likely that formal marriage will become tied more closely to decisions to have children? Is the perceived need to be married to have children diminishing, and if so how far is this trend likely to go? Given that, in terms of lifestyle, not a lot would seem to be gained from formalizing an informal union, just what does motivate couples to marry after having lived together? These are some of the intriguing questions to which, as yet, there are at best only partial answers.

Perhaps the sorts of statistics on which Chapters 2 and 3 of this volume were based are of limited use nowadays. Clearly the concepts of partnering and of marrying are no longer one and the same to the extent that they were 15-20 years ago, One of the beauties of marriage from the analyst's viewpoint is that, under the Western system of registered marriage, it has a clearly defined starting point. Once it is accepted that entry into informal unions or increases in levels of commitment to such unions as they develop may equally constitute partnering, problems of definition, motives for living together, perceptions of the permanence of relationships, and changes in motives and perceptions complicate matters considerably. Perhaps, however, it is time for those problems to be faced.

Cross-sectional variations in marriage timing have been investigated in this monograph in order to emphasize the variety of experience at the individual level which analyses of temporal change using measures of 'average' experience are inclined to obscure. Census data were particularly useful for pinpointing certain minorities within the Australian population whose marriage behaviour differed markedly from that of the population at large for reasons likely to have to do with culture and/or religious conviction. Not that more general forces influencing marriage timing were precluded from having helped to produce these differentials; there seemed little doubt that relatively early marriage among Southern European females, for instance, owed something to lower levels of educational attainment, with improved education among women largely raised in Australia and among the second generation likely to have been major factors in narrowing differentials over time. Commitment to the education process in fact stood out as the dominant force regulating marriage timing, overwhelmingly in the case of males and increasingly in the case of females.

In the comparative levels of importance of educational attainment in the marriage timing processes for males and females and in evidence that lately it has assumed increased importance in the case of females is encapsulated much of the ideological change that has contributed substantially to rising ages at marriage since the early 1970s. Models of marriage timing produced in Chapter 7 essentially are models relevant to the period when ages at marriage in Australia were at an all time low. Men, though freed from shackles that formerly had demanded that they defer marriage whilst establishing their capacity to support a wife, still displayed through differential ages at marriage an orientation towards occupational achievement as their major vehicle of status attainment. Women, while not disinclined to have delayed marrying in pursuit of education and indeed having sometimes foregone the experience altogether to achieve in the -labour force, nevertheless showed a marked tendency to have sought status through the marriage market. Many who had begun dating regularly at an early age hd also married early, and while sometimes the latter experience was the unintended consequence of unplanned pregnancy there is little doubt that much of the time it was an objective which had motivated their dating activity. Men, by contrast, who had dated regularly at an early age had for the most part done so with recreational motives to the fore.

The 1970s seem to have brought a significant strengthening in the orientation of young women to status attainment through educational and occupational achievement and a marked weakening of their orientation to status attainment through marriage. Level of education is now a much stronger determinant of marriage timing and early commencement of regular dating is a considerably weaker one. Undoubtedly the former development represents an increased determination to be independent, to achieve in one's own right and to enjoy the stimulus afforded by serious participation in the labour force, and a reduced tendency to treat education as an asset to be exploited in the marriage market rather than in the labour market. The latter development suggests that nowadays women who begin dating regularly in their midteens more often have the recreational motives attributed to males and more often avoid also marrying early through being better able to prevent or abort accidental pregnancies and better able to satisfy desires for regular intimacy and companionship through informal cohabitation. It is clear, though, that certain forces have in recent years been operating both to modify the determinants of marriage timing in Australia and to generally raise ages at fIrst marriage and foster the emergence of a signifIcant minority who are choosing not to marry. In other words they have affected marriage patterns both temporally and cross-sectionally.

In the recent past it seems that the few men who have failed to marry in Australia have tended to be men who simply did not compete well in the marriage market. With the exception of the odd ethnic group there is no evidence, for example, that migrants were the main victims of the very tight marriage market for men induced during the 1950s by heavy, male dominated post-war immigration. Rather those who missed out seem to have been more poorly educated Australian men with, perhaps, less than engaging personalities. Women, of course, were even less inclined to miss out on marriage through the early post-war decades. Where they did miss out they probably more often did so by choice, albeit that that choice sometimes may have been forced upon them by depletion of their marriage market while educational qualifications were being obtained and by a society that frowned upon women attempting to combine marriage with a serious career. Failure to marry was especially common among well educated women and women in professional or senior administrative employment.

Nowadays to not marry is not the personal disaster it was widely perceived to be only a few years ago; in fact in some quarters it is regarded as an objective to be pursued with some vigour. This change of heart has occurred, of course, in a climate which no longer equates failing to marry with living one's entire life alone or with one's parents. To not ~arry is not to forego the pleasures of cohabitation. Rather it is to retain flexibility in one's personal life, to dispense with meaningless ritual, to aff'rrm individuality and autonomy, and to recognize that these days personal relationships often do not endure. Those opting not to marry undoubtedly include both persons who either plan their lives that way or for some reason just expect them to turn out that way and persons who drift into remaining permanently single because of circumstances unforeseen. Evidence presented in this monograph suggests that members of the former group have some fairly distinctive characteristics. They are disproportionately from poorer blue collar backgrounds, nonreligious and likely to have come from broken or unhappy home environments. With parental marital disruption having become an increasingly common childhood experience among birth cohorts entering the Australian marriage market through the 1980s, there is reason here to anticipate that the trend to bypassing formal marriage will continue. There is also reason to think that both questioning of traditional notions as to the virtues of marriage and doubts as to its health in Australia may increase, further undermining the formal institution. Whether parenthood will decline further in popularity, unconventional settings for parenting will acquire increased respectability, or individualistic or feminist sentiment will intensify remains to be seen, but any of these developments, too, could cause more Australians not to bother with marriage. This is not, however, to signal its imminent demise. Marriage continues to be overwhelmingly popular and is likely to remain so for the forseeable future. The question is merely one of how large the minority opting not to marry will become, and of course failing to marry will for some be no more than an omission of legal detail.