A boy for you and a girl for me: do men want sons and women want daughters?

Gray, Edith ; Kippen, Rebecca ; Evans, Ann 等

Survey data show that more than 60 per cent of Australian couples yet to have a first child say that the child's sex does not matter. Of the minority who do express a preference most would prefer a son, though men more so than women. Where couples have one child and want another, most still say that the sex of the next child does not matter, though parents who already have a daughter are more likely to want a son than are those who already have a son to want a daughter. Where couples have two children and want a third, parents who have two children of the same sex are more likely to want a child of the opposite sex than are those who already have one son and one daughter.

INTRODUCTION

Since the late twentieth century, there has been a plethora of research showing that parents in Western industrialised countries want at least one child of each sex. This preference is expressed in the increased propensity of parents to have another child if their existing children are all boys or all girls, rather than a mix. For example, Australian mothers with two sons or two daughters are around 25 per cent more likely to have a third child than are mothers with a son and a daughter. (1)

One problem with this research is the weak theoretical basis for the preference for 'one of each' in developed countries. (2) Explanations of parental gender preference in developing countries focus on economic utility to explain son preference, and women's dominance in care duties to explain the importance of having at least one daughter. However, these are not convincing arguments in developed countries where children are an undoubted net financial impost on parents, and where it is less likely that parents now expect daughters to take on primary caring duties.

Explanations for the parental preference for 'one of each' have changed little since the 1950s, despite the extraordinary shifts in gender roles and social institutions from that time. In terms of understanding the value that sons and daughters provide, some research suggests that sons may provide particular benefits to fathers, and daughters to mothers. In aggregate this could explain the manifest preference for one child of each sex, since in heterosexual couples, men would prefer sons and women would want daughters. (3) However, in assessing research on sex-specific parental preferences for sons and daughters, Hank notes that: 'although it is frequently proposed that fathers have a pronounced desire for sons, the empirical evidence of sex-specific parental preferences is not always clear'. (4)

The current paper adds to this research by investigating whether Australian mothers and fathers, and prospective mothers and fathers, express preferences for sons or daughters. Using a nationally representative survey, we compare responses of men and women to determine whether they differ in their stated desire for boys or girls. We also compare mothers' and fathers' stated preference for the sex of their next child by the sex of their existing children. This is a better method than examining behavioural data for determining whether men and women want children of a particular sex. This is because in observing whether parents progress to another birth based on the sex of their existing children, it is difficult to disentangle men's and women's preferences since children are generally born to a couple rather than to individuals.

The remainder of the paper is structured as follows. The next section discusses previous literature on the desire for sons and daughters and outlines four questions to be addressed in this research. The data and method are then described and the three questions posed are then addressed.

STATED DESIRE FOR SONS AND DAUGHTERS

Most research investigating gender preference has been conducted using parity progression analysis, that is, the propensity of parents to have another child given the sex composition of their existing children. Arnold and others (5) assert that parents 'revise up' their family size in order to achieve families with a particular sex composition. Countries in which parents are more likely to have another child if all their existing children are female--rather than male or of both sexes--are said to exhibit a son preference. This is true of many traditional societies. In many developed countries, a preference for a mixed-sex composition is inferred because parents are equally likely to progress to a third birth if their first two children are both boys or both girls, but have a greater propensity to stop at two if they have a son and a daughter. For example, in the 1970s Young (6) found that Australian parents with same-sex children had a greater probability of having an additional child and also expected to have a larger family as compared to parents who had a mixed-sex composition. Hank and Kohler (7) compared seventeen European countries and found that preference for mixed-sex compositions dominated although some countries did show evidence of an emerging daughter preference. Research from the United States of America also finds a parental preference for a mixed-sex family. (8)

A study published in 1960 argues that sons and daughters provide different social connections; fathers do particular things with sons, such as various sporting activities, while mothers can share feminine interests with their daughters. (9) More recent research also suggests that sons are important for fathers. Many studies find that in intact families fathers spend more time with sons than they spend with daughters. (10) Other studies report a decreased likelihood of divorce for parents with sons, (11) which is attributed to greater involvement and attachment from fathers. However these conclusions have not been consistently reached in other research. (12)

There are few studies that directly examine stated sex preference. Of studies which do examine stated sex preferences, many focus on the desired sex of a firstborn child. Typically, these studies find a significant preference for sons as first children, and this preference is stronger for men than for women. (13) However, a weakness of many of these studies is that they are based on student samples; most of the participants would not be in the childbearing stage of their life course. Studies examining stated sex preferences of women pregnant with their first child find that daughters are preferred over sons, but most women state no preference. (14)

In Germany, Dahl et al. conducted one of the few representative surveys which asks questions about preferred sex of firstborn child as well as the preferred sex composition of a complete family. (15) They, like Statham et al., (16) find that most people (58 per cent) state that they do not care about the sex of their children, but more people prefer a son as a firstborn than a daughter. (17)

Aside from the research of Dahl et al., we have found little evidence of representative-survey research which asks people directly what they would like the sex composition of their family to be. Nor can we find much evidence of research which asks men and women the preferred sex of their next child. Yet this type of data would allow analysis of three questions about parental sex preference, outlined below.

1. Are respondents who are pregnant or whose partner is pregnant for the first time more likely to state a desired sex of their next child than those who are not pregnant, and, if so, are they more likely to state that they would like their child to be a daughter than those not pregnant?

2. Does the propensity to nominate a preferred sex of child vary by parity (number of children) and by sex of existing children?

3. Are men more likely to state a desire for sons than for daughters as compared with women (and vice versa)?

DATA AND METHOD

To answer these questions about stated preferences for sex of children, we use data from the Wave 5 collection of the Household, Income and Labour Dynamics in Australia (HILDA) survey (information on the Wave 5 data collection conducted in 2005 is available from MIAESR). (18) In 2005, questions on fertility intentions and desired sex of children were asked for the first time as part of a fertility module. The fertility module was asked of all females aged 18 to 44 years and partnered males aged 18 to 54 years if their female partner was less than 45 years. Respondents (both females and males) aged less than 18 years who lived independently of their parents were also asked this module of questions. Respondents who were sterilized or where there were health difficulties in achieving pregnancy were excluded from the fertility module. This results in a weighted sample of 5,238 persons (2,874 males and 2,364 females). Of this sample, 3,354 respondents indicated that they intend to have a(nother) child in the future. The question about preferred sex of children in HILDA takes the following form:

'Would you prefer your [first/next] child to be a boy or a girl?'
Response options: 1 'Boy', 2 'Girl', 3 'Doesn't matter'

We take the following analytical approach to investigate our questions on parents' stated sex preference of children. For those respondents who state that they want a[nother] child, we examine stated sex preference to investigate statements of women and men firstly by pregnancy status, and then by parity and sex composition of existing children.

VARIATION IN STATED SEX PREFERENCES OF CHILDLESS MEN AND WOMEN BY PREGNANCY STATUS

Previous research has found that women who are pregnant with their first child are more likely than non-pregnant women to state that they would prefer a girl. (19) We use HILDA to examine this proposition for Australia. The number of women who were pregnant with their first child and the number of men with no children whose partner was pregnant at the time of survey totals 105 people (Table 1).

Table 1 shows no significant difference in the pattern of reporting of preferred sex of first child by whether the respondent or respondent's partner is pregnant or not. For both groups just over 60 per cent of respondents state that it 'doesn't matter' what the sex of their first child will be. Men were more likely to state that they would prefer to have a boy as a first child (35 per cent) than to have a girl (five per cent). Women were also more likely to state that they wanted a boy (23 per cent) than a girl (16 per cent), but there is a statistically significant difference in the reporting patterns of men and women. Men are much more likely to state a desire for a son than a daughter as compared to women.

THE EFFECT OF THE NUMBER AND SEX OF CHILDREN ON STATED SEX PREFERENCES

The previous results indicate that most parents-to-be do not mind what the sex of their first child is, although a sizeable proportion (over one in three men and one fifth of women) indicate that they would like their first child to be a boy. We now examine stated sex preference for parents with one or two children who intend to have more children in the future. Figure 1 shows the pattern of response of males and females who have one child and state that they want another child (n=209) by the sex of that child. The pattern shows differential responses of males and females. For males, around 65 per cent say that sex of the next child 'doesn't matter', but this response differs by whether their first child is a son or a daughter. For men who have a son, 74 per cent say that the sex of their next child does not matter, while 18 per cent state that they would like a daughter next. In comparison, 55 per cent of men who have a daughter say that the sex of their next child does not matter, with 37 per cent stating that they would prefer their next child to be a boy. These results can be interpreted as a desire by many fathers for sons.

The opposite pattern of response is found for females, but the difference between stated desire for a boy or girl is not quite as strong as it was found to be for males. Women are almost equally likely to state that the sex of their second child does not matter whether they have a son (65 per cent) or a daughter (71 per cent). Around 30 per cent of women who have a son state that they would like their next child to be a girl and almost 20 per cent of women who have a daughter state that they would like their next child to be a son. This suggests that women are more likely to state a preference for a daughter, but the difference between women's stated sex preferences are not as great as men's.

Overall, most parents of one child say that the sex of their next child does not matter. However, there is evidence that many parents want to 'balance' their families by adding a child of the opposite sex. Further, there are distinct differences in the responses of men and women, with men more likely to want a son when they have a daughter than to want a daughter when they have a son, and vice-versa for women. For parents who have two children and intend to have a third (n=513) we again see evidence of a substantial percentage of parents wanting to balance their family (Figure 2). In fact, many more parents state that they desire a child of a particular sex than is evident at the smaller family sizes, and sex of existing children is an important factor in stated sex preferences. Figure 2 shows that parents who have a son and a daughter, and who want more children, are most likely to state that the sex of the next child 'doesn't matter'. Over 90 per cent of men and 70 per cent of women who have a son and a daughter state that the sex of the next child does not matter.

[FIGURE 1 OMITTED]

However, in families with two children of the same sex who want more children, the percentage of parents who state a preferred sex is larger than for respondents with no children (Table 1) and parents with one child (Figure 1). Almost 70 per cent of women and over 50 per cent of men who have two sons say that they would like a daughter for their next child. A smaller, but substantial percentage of parents with two daughters say that they would like a son next (over 40 per cent of both mothers and fathers express a desire for a son). These findings for parents of two children support our previous research that parents with two children of one sex are more likely to have another child than are parents with a son and a daughter, in order to try for a child of the missing sex. (20) Thus, we would expect sex preferences for the next child to be much stronger for parents of two boys or two girls who intend to have another child.

DISCUSSION

In this paper we have illuminated three dimensions of parental sex preferences in Australia. Firstly, stated sex preference does not differ by whether a respondent, or a respondent's partner, is pregnant or not. While research from other developed countries has found that women who are pregnant are more likely to state that they would prefer a daughter than a son, we do not find this pattern of response. Respondents are equally likely to state a sex preference whether they are expecting or not, with around 40 per cent stating a preferred sex for their first child. (21)

Secondly, the number and sex of children that a person already has influences their propensity to state a preferred sex of their next child. A sizeable percentage (40 per cent) of people with no children state a preferred sex for their first child while fewer (35 per cent) state a preferred sex when they have one child and intend to have a second. A readiness to state a preferred sex of child is most likely to be demonstrated by people who have two children and intend to have a third. Parents who have a son and a daughter are least likely (over 80 per cent) to state a preferred sex of next child. It simply does not matter to most of these parents. However, parents who have two sons or two daughters are the most likely to state a preferred sex of their next child, with parents most likely to nominate a child of the opposite sex to the two children they already have.

[FIGURE 2 OMITTED]

Thirdly, while a strong preference for one child of each sex is exhibited, on the whole women are more likely to express a desire for a daughter and men are more likely to say that they want a son. For example, of those parents with one child who expect to have another, 31 per cent of mothers but only 18 per cent of fathers with a son want a daughter next, while 37 per cent of fathers but only 19 per cent of mothers with a daughter want a son next (Figure 2). A sizeable proportion of both male and female prospective parents express a desire for a first-born son, however this preference is significantly stronger for men than it is for women (Table 1). The overwhelming conclusion is that parents want a child of each sex in order to balance their family, (22) but that many also have a marked preference for at least one child of their own sex. This suggests that many men perceive benefits from having a son, and women from having a daughter, that are not present with a child of the opposite sex.

Acknowledgements

This paper uses unit record data from the Household, Income and Labour Dynamics in Australia (HILDA) Survey. The HILDA Project was initiated and is funded by the Australian Government Department of Families, Community Services and Indigenous Affairs (FaCSIA) and is managed by the Melbourne Institute of Applied Economic and Social Research (MIAESR). The findings and views reported in this paper, however, are those of the authors and should not be attributed to either FaCSIA or the MIAESR. This research forms part of a larger project 'Do Australian parents want both a son and a daughter?' funded by the Australian Research Council (DP0558818).

References

(1) R. Kippen, E. Gray and A. Evans, 'The impact on Australian fertility of wanting one of each', People and Place, vol. 13, no. 2, 2005, pp. 12-20

(2) K. Hank, 'Parental gender preferences and reproductive behaviour: a review of the recent literature', Journal of Biosocial Science, vol. 39, no. 5, 2007, pp. 759-767

(3) ibid.

(4) ibid., p. 761; K. Hank and H-P. Kohler, 'Gender preferences for children in Europe: empirical results from 17 FFS countries', Demographic Research, vol. 12, no. 1, 2000

(5) F. Arnold, 'Gender preferences for children', Demographic and Health Surveys Comparative Studies No. 23, Maryland: Macro International, 1997; F. Arnold, M. Choe and T. Roy, 'Son preference, the family-building process and child mortality in India', Population Studies, vol. 52, no. 3, 1998, pp. 301-315; A. Basu and M. Das Gupta, 'Family systems and the preferred sex of children', in N. Smelser and P. Baltes (Eds.), International Encyclopedia of the Social and Behavioral Sciences, vol. 8, pp. 5350-5357

(6) C. Young, 'Family building differences between same sex and mixed sex families in Australia', Australian Journal of Statistics, vol. 19, no. 2, 1977, pp. 83-95

(7) Hank and Kohler, 2000, op. cit.

(8) M. Pollard and S. P. Morgan, 'Emerging parental gender indifference? Sex composition of children and the third birth', American Sociological Review, vol. 67, no. 4, 2002, pp. 600-613

(9) D. S. Freedman, R. Freedman and P. K. Whelpton, 'Size of family and preference for children of each sex', The American Journal of Sociology, vol. 66, no. 2, 1960, pp. 141-146

(10) Many of these studies find that fathers are involved in play-based activities, but also in discipline and educational activities. See W. Yeung, J. Sandberg, P. Davis-Kean and S. Hofferth, 'Children's time with fathers in intact families', Journal of Marriage and the Family, vol. 63, no. 1,2001, pp. 136-154; S. P. Morgan, D. Lye and G. Condran, 'Sons, daughters, and the risk of marital disruption', American Journal of Sociology, vol. 94, no. 1, 1988, pp. 110-129; S. Lundberg, S. McLanahan and E. Rose, 'Child gender and father involvement in fragile families', Demography, vol. 44, no. 1,2007, pp. 79-92. These studies find that fathers are more involved with sons in marital than non-marital relationships consistent with an argument about the importance of fathers for long-term development of sons.

(11) Morgan et al., 1988, op. cit.; A. Katzev, R. Warner and A. Acock, 'Girls or boys? Relationship of child gender to marital instability', Journal of Marriage and the Family, vol. 56, no. 1, 1994, pp. 89-100; G. Dahl and E. Moretti, 'The demand for sons: evidence from divorce, fertility, and shotgun marriage', National Bureau of Economic Research Working Paper No. 10281

(12) No relationship between sex of children and divorce was found in an Australian study by M. Bracher, G. Santow, S. P. Morgan and J. Trussell, 'Marriage dissolution in Australia: models and explanation', Population Studies, vol. 47, no. 3, 1993, pp. 403-425.

(13) For an overview see J. Marleau and J-F. Saucier, 'Preference for a first-born boy in Western societies', Journal of Biosocial Science, vol. 34, 2002, pp. 13-27; D. Swetkis, F. Gilroy and R. Steinbacher, 'Firstborn preference and attitudes toward using sex selection technology', Journal of Genetic Psychology, vol. 163, no. 2, 2002, pp. 228-238.

(14) Marleau and Saucier, 2002, op. cit.; H. Statham, J. Green, C. Snowdon and M. France-Dawson, 'Choice of baby's sex', Lancet, vol. 341, 1993, pp. 564-565

(15) E. Dahl, M. Beutel, B. Brosig and K-D. Hinsch, 'Preconception sex selection for non-medical reasons: a representative survey from Germany', Human Reproduction, vol. 18, no. 10, 2003, pp. 2231-2234

(16) Statham et al., 1993, op. cit.

(17) It is possible that there is a general reluctance to express a desire for a child of a particular sex resulting in the proportion of people that state that they 'do not care' about the sex of a child. However, it is impossible to disentangle whether people are actually refusing to state a desire for a child of a particular sex or whether they actually 'do not care'.

(18) Melbourne Institute of Applied Economic and Social Research (MIAESR), Hilda Survey Annual Report 2006, University of Melbourne, Melbourne, 2007

(19) Marleau and Saucier, 2002, op. cit.

(20) R. Kippen, A. Evans and E. Gray, 'Parental preference for sons and daughters in a Western industrial setting: evidence and implications', Journal of Biosocial Science, vol. 39, pp. 583-597

(21) Note that the percentage stating that sex of first child doesn't matter (60 per cent) is very similar to that found by Dahl et al., 2003, op. cit.

(22) This is consistent with arguments made by IVF specialists, for example Kylie de Boer talking on Lateline in 2004 <http://www.abc.net.au/lateline/content/2004/s1188430.htm>.

Table 1: Percentage distribution of preferred sex of first child by sex
of respondent and pregnancy status of female respondent or male
respondent's partner

 Sex and pregnancy status of respondent/respondent's
 partner
Preferred sex Male
of first child Partner pregnant Partner not pregnant Total*

Male 32.1 35.0 34.9
Female 7.1 5.4 5.5
Doesn't matter 60.7 59.6 59.7
Total 100.0 100.0 100.0
Number 56 1,355 1,411

 Sex and pregnancy status of respondent/respondent's
 partner
Preferred sex Female
of first child Pregnant Not pregnant Total*

Male 20.4 22.9 22.8
Female 14.3 15.9 15.9
Doesn't matter 65.3 61.2 61.4
Total 100.0 100.0 100.0
Number 49 1,036 1,085

Source: HILDA 2005
* p<0.0001. Statistically significant difference in the distribution of
stated preferred sex of first child for males and females.