Partnerships at the 2006 census: preliminary findings.

Heard, Genevieve

INTRODUCTION

Married and de facto partnerships in Australia are the site of continued dramatic change. In 2004 Birrell et al. reported a 'precipitous' decline in proportions partnered over consecutive census collections. In particular, proportions married were shown to have 'plummeted' between 1986 and 2001. (1)

The explanation provided for these trends was primarily economic. Noting that the decline in partnering was particularly marked among men and women without post-school qualifications, Birrell et al. hypothesised that insecure employment at the lower end of the job market had left many without the resources to form families.

Data from the 2006 census now permit an update on these trends. The focus of this brief article is change in the relationship patterns of men and women aged 20 to 44 years over the course of the decade to 2006. The article examines these patterns for groups of men and women by educational attainment and by income. It argues that these groups are primarily distinguished by patterns of marriage, more so than de facto partnering.

PROPORTIONS PARTNERED, 1996 TO 2006

Table 1 reveals some stabilisation in proportions partnered between 2001 and 2006, albeit from low levels. Among men, over half aged 25 to 29 years, one third aged 30 to 34 years, and a quarter aged 35 to 39 and 40 to 44 years are unpartnered. Unpartnered women are a considerably smaller proportion of all women in the younger age groups, but nevertheless exceed one quarter by 35 to 39 years and 40 to 44 years.

The stabilisation evident in proportions partnered is due to increased proportions living in de facto relationships (as high as 21 per cent among men and women aged 25 to 29 years, up from 14 per cent in 1996). However, proportions married continued their sharp decline. The proportion of women aged 30 to 34 years who were married fell to 56 per cent in 2006, down nine percentage points from 1996. Meanwhile, the proportion of men in this age group who were married fell below half (49 per cent), down eight percentage points from 1996. These are substantial changes in the course of only one decade.

Marital dissolution cannot explain the declining proportions of men and women in married partnerships. On the contrary, Table 2 shows that proportions separated and divorced have declined among men and women of all ages. Rather, it seems men and women are increasingly reluctant to enter marriage in the first place. Proportions ever married have declined even more sharply than proportions married.

Table 1: Social marital status by age and sex, 1996, 2001 and 2006 (a)

Age Males (per cent)

 Married De facto Proportion of Total Total not
 (b) partnered who partnered partnered
 are de facto

20-24 years

 1996 7 10 58 17 83
 2001 5 10 67 16 84
 2006 4 12 74 16 84
25-29 years
 1996 34 14 29 49 51
 2001 28 18 38 46 54
 2006 25 21 46 46 54
30-34 years
 1996 57 11 16 68 32
 2001 51 14 22 66 34
 2006 49 18 26 67 33
35-39 years
 1996 68 8 11 76 24
 2001 63 10 14 73 27
 2006 60 13 18 73 27
40-44 years
 1996 72 6 8 79 21
 2001 68 8 11 76 24
 2006 65 10 14 75 25

Age Females (per cent)

 Married De facto Proportion of Total Total not
 (b) partnered who partnered partnered
 are de facto

20-24 years

 1996 16 14 46 30 70
 2001 12 16 57 28 72
 2006 10 18 65 27 73
25-29 years
 1996 48 14 22 61 39
 2001 41 18 30 58 42
 2006 36 21 37 57 43
30-34 years
 1996 65 9 13 74 26
 2001 59 12 17 71 29
 2006 56 15 21 71 29
35-39 years
 1996 70 7 9 77 23
 2001 66 9 12 75 25
 2006 63 11 15 74 26
40-44 years
 1996 72 5 7 77 23
 2001 68 7 9 75 25
 2006 65 9 12 74 26

Source: ABS cat. no. 2068.0- 2006 Census Tables; author's calculations
Notes: (a) Excludes persons who were temporarily absent on census night.
(b) Includes same-sex couples.

It is tempting to explain away these trends in terms of the continuing de-institutionalisation of marriage, by which is meant the weakening of the social expectation that partnerships will take the form of traditional marriage. (2) This phenomenon is common to many western countries: the UN states that 'formal marriage is receding everywhere' (3) Yet further disaggregation of these trend data by educational attainment and by income suggest that socio-economic factors are at play, complicating the broad-brush picture of marriage 'receding everwhere'.

PARTNERSHIPS BY EDUCATIONAL ATTAINMENT

With the increased take-up of education and paid work by women in the post-war era, some observers have considered it self-evident that increased female economic independence serves to lower the incentive to union formation. (4) Investment in education has brought a greater capacity for women to exercise choices regarding marriage, motherhood and work. The greater the investment, the greater the financial incentive to favour work. Commentators often assume that more highly educated women acquire a less traditional orientation, place less emphasis on family, and as a result are less likely to form partnerships. (5) This expectation particularly applies to marriage, as the most binding of commitments to family.

Table 3 shows that, in Australia, the status of this assumption has changed from true to false within the space of a decade. In 1996 women with post-school qualifications were indeed less likely to be partnered than those without, in most age groups. By 2006, this was no longer the case. The gap had closed among women in their twenties, while the reverse was true among women aged 30 to 34 years or more: that is, women with post-school qualifications were more likely to be partnered than those without.

Table 2: Registered marital status by age and sex, 1996, 2001 and 2006

Sex/Age Number Ever married (a) Separated or divorced (a)
(years) ('000) (per cent) (per cent)

 1996 2001 2006 1996 2001 2006 1996 2001 2006

Males

20-24 666.7 619.3 681.7 8 6 5 1 1 0
25-29 671.2 644.3 635.8 38 32 26 4 3 2
30-34 687.0 678.9 685.3 65 59 54 9 8 6
35-39 695.8 695.0 715.2 79 74 69 12 12 10
40-44 651.2 698.0 722.1 87 83 78 15 15 14

Females

20-24 655.7 604.4 665.7 18 13 10 2 1 1
25-29 682.1 657.9 641.1 55 46 40 7 6 4
30-34 703.1 710.8 714.2 77 71 65 12 12 9
35-39 716.2 723.9 751.0 87 82 78 16 16 14
40-44 664.1 726.8 749.6 91 88 84 18 19 19

Source: ABS, cat. no. 2068.0--2006 Census Tables; author's calculations
Note: (a) 'Separated or divorced' is a subcategory of'ever married'.
Therefore, separated and divorced persons are counted twice in the
percentages shown in this table. Percentages do not add to 100 per cent
as there are other categories of 'registered marital status' not shown
here.

Table 3: Women, Relationship in household by age and educational
attainment, 1996, 2001 and 2006

 Age Post-school Women
(years) qualifications ('000)

 1996 2001 2006

20-24 Bachelor and 85.8 91.1 111.6
 above

 Advanced 78.5 102.1 130.5
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 491.7 419.2 423.6

 Total 656.0 612.4 665.7

25-29 Bachelor and 121.3 164.3 201.9
 above

 Advanced 90.0 117.0 130.2
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 470.8 384.3 309.1

 Total 682.1 665.6 641.1

30-34 Bachelor and 109.2 151.8 206.4
 above

 Advanced 85.9 115.6 144.7
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 508.1 449.8 363.0

 Total 703.2 717.2 714.2

35-39 Bachelor and 111.5 132.7 179.0
 above

 Advanced 89.6 109.9 146.9
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 515.4 485.7 425.1

 Total 716.5 728.3 751.0

40-44 Bachelor and 99.3 133.2 151.3
 above

 Advanced 82.4 111.2 141.6
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 482.5 486.8 456.6

 Total 664.2 731.1 749.6

 Age Post-school Wife in
(years) qualifications registered
 marriage
 (per cent) (a)

 1996 2001 2006

20-24 Bachelor and 13 10 9
 above

 Advanced 16 13 11
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 16 12 9

 Total 16 12 10

25-29 Bachelor and 41 38 35
 above

 Advanced 50 42 38
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 49 41 35

 Total 48 41 36

30-34 Bachelor and 64 60 61
 above

 Advanced 67 60 56
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 65 58 53

 Total 65 59 56

35-39 Bachelor and 70 69 68
 above

 Advanced 73 66 63
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 70 65 60

 Total 70 66 63

40-44 Bachelor and 70 69 68
 above

 Advanced 73 67 64
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 72 67 64

 Total 72 68 65

 Age Post-school De facto
(years) qualifications (per cent)
 (a), (b)

 1996 2001 2006

20-24 Bachelor and 10 13 15
 above

 Advanced 14 18 23
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 14 15 17

 Total 14 15 18

25-29 Bachelor and 12 17 20
 above

 Advanced 12 17 23
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 14 18 21

 Total 13 17 21

30-34 Bachelor and 8 11 14
 above

 Advanced 8 11 15
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 9 12 16

 Total 9 12 15

35-39 Bachelor and 6 8 10
 above

 Advanced 6 8 11
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 7 9 12

 Total 7 8 11

40-44 Bachelor and 5 6 9
 above

 Advanced 5 6 9
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 5 7 9

 Total 5 7 9

 Age Post-school Total
(years) qualifications Partnered
 (percent)

 1996 2001 2006

20-24 Bachelor and 22 23 25
 above

 Advanced 30 31 33
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 31 27 26

 Total 30 27 27

25-29 Bachelor and 53 55 56
 above

 Advanced 62 59 60
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 63 59 57

 Total 61 58 57

30-34 Bachelor and 72 71 74
 above

 Advanced 75 71 72
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 74 70 69

 Total 74 71 71

35-39 Bachelor and 76 76 78
 above

 Advanced 79 74 74
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 76 74 72

 Total 77 74 74

40-44 Bachelor and 75 75 77
 above

 Advanced 78 74 73
 diploma,
 diploma, Cert
 IV, Cert III

 Other/None 77 74 73

 Total 77 74 74

Source: ABS, customised census datasets
Notes: percntages exclude women temporarily absent on census night.
Percentages for each group in each year do not add to 100 per cent as
there are other 'relationship in household' categories not shown here.
(b) In 1996 and 2001 , women in same sex relationships were included
with 'other related or unrelated individuals'. In 2006 they
were instead counted as de facto partners. The inclusion
of same sex partners with women in de facto relationships in 2006
boosts the proportion of women in this group and in the 'partnered'
group by approximately 0.3 percentage points overall and by no more
than 0.5 percentage points in any given age group.

Again, it is primarily marriage that distinguishes the educational groups, since the proportion of women in de facto relationships has increased at a similar rate across all categories of educational attainment. In the younger age groups, proportions married have declined across all educational categories. However, among women aged 30 to 34 years or more, proportions married have stabilised among degree-qualified women whilst continuing to decline rapidly among women with a lesser qualification and women with no post-school qualification.

The corollary of the assumption that educated women would reject marriage was that women without post-school qualifications would focus on family roles. Until recently, the relatively high rates of marriage among these women was consistent with this expectation. The reverse is now the case. Among women aged 30 to 34 years in 2006, 61 per cent of those with degrees were married, compared to 53 per cent of those without post-school qualifications.

Are these patterns repeated among men? Men with better economic prospects have traditionally made more attractive partners as they are better able to fulfil the breadwinner role. This leads to the expectation that male educational attainment will be positively related to partnering outcomes.

Among men in their twenties, those with Certificate III (trade) qualifications appear to enjoy a distinct advantage in partnering (see Table 4). For the other age groups shown, a positive gradient emerges as expected, such that men with degrees are most likely to be partnered and men without post-school qualifications least likely. While this was the case in each of the census years shown, the gradient has grown steeper in the ten years to 2006. At age 40 to 44 years in 2006, only 16 per cent of degree-qualified men were not in any kind of partnership, compared to 32 per cent of men without qualifications.

Table 4: Men, Relationship in household by age and educational
attainment, 1996, 2001 and 2006

 Age Post-school qualification Males
(year) ('000)

 1996 2001 2006
20-24 Bachelor and above 54.5 57.3 72.4
 Advanced diploma, diploma, Cert IV 29.8 33.1 47.4
 Cert III 102.4 94.8 101.5
 Other/ None 479.9 443.3 460.3
 Total 666.6 628.5 681.7
25-29 Bachelor and above 97.5 117.4 142.4
 Advanced diploma, diploma, Cert IV 38.4 47.1 52.7
 Cert III 150.2 144.5 126.3
 Other/ None 385.4 344.6 314.4
 Total 671.5 653.6 635.8
30-34 Bachelor and above 100.4 121.0 152.4
 Advanced diploma, diploma. Cert IV 41.6 50.4 63.3
 Cert III 168.3 163.4 151.5
 Other/ None 377.2 352.6 318.0
 Total 687.4 687.4 685.3
35-39 Bachelor and above 111.0 117.5 145.0
 Advanced diploma, diploma. Cert IV 46.0 51.8 65.6
 Cert III 168.9 177.0 166.2
 Other/ None 370.1 356.5 338.4
 Total 695.9 702.9 715.2
40-44 Bachelor and above 107.4 122.2 133.0
 Advanced diploma, diploma. Cert IV 46.9 54.6 65.5
 Cert III 145.2 175.9 176.2
 Other/ None 351.2 352.8 347.3
 Total 650.7 705.5 722.1

 Age Husband in registered marriage De facto
(year) (per cent) (a) (per cent) (b)

 1996 2001 2006 1996

20-24 6 5 4 6
 7 5 4 7
 12 8 6 14
 6 4 3 9
 7 5 4 10
25-29 31 28 25 11
 37 31 28 11
 42 36 31 15
 32 25 21 14
 34 28 25 14
30-34 61 56 56 8
 63 57 55 8
 64 58 55 11
 52 46 41 11
 57 51 49 11
35-39 75 71 70 6
 74 70 67 6
 72 68 64 8
 62 56 52 8
 68 63 60 8
40-44 79 77 76 5
 78 74 72 5
 76 71 68 6
 68 62 57 7
 72 68 65 6

 Age Total partnered (a),(b)
(year)
 2001 2006 1996 2001 2006
20-24 8 9 12 13 14
 8 10 14 14 15
 16 20 26 24 26
 9 10 15 14 14
 10 12 17 15 16
25-29 15 18 42 42 43
 15 20 47 46 48
 20 27 57 56 57
 17 20 46 42 42
 17 21 48 46 46
30-34 12 15 69 68 72
 11 16 71 69 71
 14 19 75 72 74
 14 18 64 60 59
 14 18 68 65 67
35-39 7 10 80 79 81
 8 12 80 78 79
 10 14 80 78 78
 11 14 71 67 66
 10 13 75 73 73
40-44 6 8 84 82 84
 7 9 83 80 81
 8 11 82 79 79
 8 11 74 70 68
 8 10 78 75 75

Source: ABS, customised census dataset
Notes: (a) Percentages exclude persons temporarily absent on census
night. Percentages for each group in each year do not add to 100
per cent as there are other 'relationship in household' categories not
shown here.(b) In 1996 and 2001, men in same sex relationships
were included with 'other related or unrelated individuals'. In 2006
they were instead counted as de facto partners. The inclusion of same
sex partners with men in de facto relationships in 2006 boosts the
proportion of men in this group and in the 'partnered' group by
approximately 0.3 percentage points overall and by no more than 0.6
percentage points in any given age group.

Again, this change is due to patterns of marriage rather than cohabitation, which increased among all men. Among men with post-school qualifications, the proportions married and therefore in total partnerships stabilised in the five years to 2006. By contrast, marriage continued its steep decline among men without post-school qualifications. By age 40 to 44 years, the marriage gap is enormous: 76 per cent of degree qualified men were married in 2006, compared to just 57 per cent of men with no post-school qualifications.

PARTNERSHIPS BY INCOME

Table 5 introduces male income to the analysis and reinforces this picture even more clearly (female income is less useful because so many women with young children indicate no income or reduced earnings). Across all age groups, the consistency of the positive relationship between male income and proportions partnered is striking. And again, the gulf between the partnering outcomes of men at the lower and higher ends of the income spectrum is substantial. In terms of total proportions partnered, up to 36 percentage points (among men aged 25 to 29 years and 30 to 34 years) separate those earning less than $400 per week from those earning $1,600 or more.

Table 5: Men, Relationship in household and registered marital status
by age and individual income, 2006

Age (year) Individual weekly income Males ('000)

20-24 Negative income, nil, $1 $399 253.9
 $400-$799 247.0
 $800-$1, 599 100.7
 $1, 600 or more 8.3
 Total 681.7
25-29 Negative income, nil, $1 $399 114.7
 $400-$799 247.0
 $800-$1, 599 201.2
 $1, 600 or more 36.2
 Total 681.8
30-34 Negative income, nil, $1 $399 96.3
 $400-$799 201.2
 $800-$1,599 218.3
 $1,600 or more 85.3
 Total 685.3
35-39 Negative income, nil, $1 $399 102.2
 $400-$799 171.0
 $800-$1,599 267.3
 $1,600 or more 113.1
 Total 715.2
40-44 Negative income, nil, $1 $399 111.9
 $400-$799 171.2
 $800-$1,599 260.9
 $1,600 or more 118.4
 Total 722.1

Age (year) Realationship in household (per cent) (a)

 Husband in a Partner in de Total partnered
 registered facto marriage
 marriage

20-24 2 5 8
 4 14 19
 8 22 30
 10 29 38
 4 12 16
25-29 14 12 27
 4 14 19
 22 21 44
 37 26 63
 25 21 46
30-34 31 14 45
 22 21 44
 31 25 56
 64 17 81
 49 18 67
35-39 40 12 53
 53 14 68
 67 14 80
 76 11 87
 60 13 73
40-44 47 10 57
 60 11 71
 71 11 81
 80 9 88
 65 10 75

Age (year) Registered marital status (per cent) (b)

 Ever married Separated or
 Divorced

20-24 3 0
 5 0
 8 0
 10 1
 5 0
25-29 17 2
 5 0
 25 2
 38 2
 26 2
30-34 37 6
 25 2
 33 2
 68 4
 54 6
35-39 52 12
 65 12
 75 9
 82 7
 69 10
40-44 63 17
 75 16
 83 13
 88 9
 78 14

Source: ABS, customised census dataset
Notes: (a) Percentages exclude persons temporarily absent on census
night. Percentages do not add to 100 per cent as there are other
'relationship in household' categories not shown here.
(b) Separated or divorced' is a subcategory of 'ever married'.
Therefore, separated and divorced persons are counted twice in the
percentages shown in this table. Percentages do not add to 100
per cent as there are other categories of 'registered marital status'
not shown here.
(c) Totals include males whose income was not stated.

Again, this effect is due to differential patterns of marriage. Men on low incomes are considerably less likely to have ever been married than their higher-earning counterparts, but are also more likely to be separated or divorced. As a consequence, in the snapshot provided by the 2006 census, there are fewer low income than high income men in registered marriages, and fewer low income men than high income men partnered.

Consider men in what may be considered the prime partnering age group of 30 to 34 years. It was previously noted that the proportion of men in this age group who were married has fallen below half (49 per cent). Table 5 shows that this situation is attributable to the behaviour of men earning less than $400 per week (31 percent of whom were married in 2006) and men earning $400-$799 per week (43 percent married). By contrast, the proportion of high income men ($1,600 or more per week)who were married is closer to two thirds (64 percent).

DISCUSSION AND CONCLUSION

There is some stabilisation of total partnering in Australia, due to increasing proportions of men and women in de facto relationships. However, proportions married continue to decline steeply.

When these trends are disaggregated by educational attainment and income, it becomes apparent that the notion of a 'deinstitutionalisation of marriage' is simplistic. Rather, we are witnessing the redistribution of marriage. Increasingly, married Australians are concentrated among those with higher earning potential. This pattern is relatively new for women, and although it is consistent with expectations regarding men, it is becoming more entrenched

The pattern points to a disturbing phenomenon--already evident in the US--whereby access to the means of family formation is increasingly dependent on economic resources for both sexes. The divergence of trends in union formation and dissolution has prompted a growing number of commentators to express alarm at the polarisation of partnering patterns by socioeconomic status.(sup.6) Goldstein and Kenney warn that 'marriage is increasingly becoming a province of the most educated'.(sup.7)

In Australia, there is little academic interest in this phenomenon. Consequently, explanations for the divergence in partnering outcomes are few. However, Birrell et al. hypothesised that economic restructuring had undermined job security in sectors traditionally dominated by poorly qualified men (for example, manufacturing), leaving many without the resources to attract a partner and establish a home and family. (8) It may be argued that the stabilisation in total proportions partnered between 2001 and 2006 fits such an economic explanation, given the prosperity enjoyed by Australians in this period. Perhaps the situation faced by poorly qualified men has improved under these conditions.

However, the 2006 data presented in this article create new theoretical challenges, given the continued decline in proportions married. It is sometimes argued that marriage requires a stronger economic underpinning than does cohabitation. (9) It is possible that recent conditions have been amenable to the formation but not the formalisation of partnerships. After all, the level of prosperity required to counter the hypothesised barriers to marriage caused by workforce deregulation and casualisation is unknown.

An alternative but compatible explanation starts at the other end of the socio-economic spectrum. In the US, Oppenheimer has pointed to the growing importance of two incomes for men and women who aspire to a middle-class lifestyle. "Collaborative relationships'--whereby both partners contribute to the household income--are not only advantageous but essential.(10)

In Australia, the high house prices of recent years render such concerns equally valid. Putting aside all the non-economic attractions of intimate relationships, educated men and women may be formalising and maintaining partnerships because the realisation of their middle class aspirations depends on it. Meanwhile, such aspirations may be out of the reach of many men and women without post-school qualifications.

It remains to be seen whether and how these theories can incorporate the empirical developments to which this article refers. The data presented are sufficient to warrant further investigation into the circumstances facing men and women across the socio-economic spectrum, and the impact of these circumstances on the formation and formalisation of partnerships in Australia.

References

(1) R. Birrell, V. Rapson and C. Hourigan, Men and Women Apart: Partnering in Australia, Centre for Population and Urban Research, Monash University and the Australian Family Association, Melbourne, 2004

(2) A. J. Chcrlin, 'The deinstitutionalization of American marriage', Journal of Marriage and Family, vol. 66, no, 4, 2004

(3) United Nations, Partnership and Reproductive Behaviour in Low-Fertility Countries, Population Division, Department of Economic and Social Affairs, New York, 5 (May) 2003 <http://www.un.org/esa/population/publications/reprobchavior/partrepro.pdf> accessed 27 March 2007

(4) G. S. Becker, A Treatise on the Family, Harvard University Press, Cambridge, Mass., 1981

(5) Z. Qian, 'Changes in assortativc mating: the impact of age and education, 1970-1990', Demography, vol. 35, no. 3, 1998; G. Santow and M. Bracher, 'Change and continuity in the formation of first marital unions in Australia', Population Studies, vol. 48, 1994

(6) A. J. Cherlin, 'American marriage in the early twenty-first century', Future of Children, vol. 15, no. 2, 2005

(7) J. R. Goldstein and C. T. Kenney, 'Marriage delayed or marriage forgone? New cohort forecasts of first marriage for US women', American Sociological Review, vol. 66, no. 4, 2001

(8) Birrell et al.,2004,op.cit.

(9) T. L. Huston and H. Melz, 'The case for (promoting) marriage: the devil is in the details', Journal of Marriage and Family, vol. 66,2004; [O.sub.+] Kravdal, 'Does marriage require a stronger economic underpinning than informal cohabitation?', Population Studies: a Journal of Demography, vol. 53, no. 1, 1999

(10) V. K. Oppenheimer, 'Women's rising employment and the future of the family in industrial societies', Population and Development Review, vol. 20, no. 2,1994

Data from the 2006 census provide an update on partnering trends (including both registered marriages and de facto partnerships). While total partnering had declined sharply from 1986 to 2001 it has now stabilised. This is due to increased proportions in de facto relationships; proportions married continue to decline, largely because of the trend away from registered marriage among men and women without post-school qualifications.