Reasons for gender preference Data and method
3 Empirical findings from Western countries

In the past 25 years there has been only limited empirical research on gender preferences in modern Western societies, mainly conducted in North America (see Marleau and Maheu [19] for an overview). In addition, Carr-Hill, Sampier and Sauve [6] investigate sex preferences of Aberdeen families, Gray, Duckworth and Nakajima [10] are interested in the case of Japan, Jacobsen, M°ller and Engholm [16] discuss Danish fertility rates in relation to the sexes of preceding children in the family, Schullstr÷m [25] studies Swedish cohorts born 1946-1975, and Young [29] analyzes data from Australia.

The methodological approaches used in these studies are quite different. While some researchers ask directly for the respondent’s gender preferences [17], others use various indirect statistical measures (a critical review of such methods is, for example, given by Haughton and Haughton [12], and McClelland [20]). Despite this methodological heterogeneity, the results are very similar:

There seems to be a consistent tendency for having at least one child of each sex, which supports the above hypothesis of a preference for a gender mix. However, when people are asked for the preferred sex of their first child, or if they have chosen an unbalanced number of children, there is some indication for a predominance of sons over daughters. Therefore parity matters when gender preferences are analyzed (see Gray [11], Jacobsen, M°ller and Engholm [16]).

Is there any possible explanation for the persistence of a slight son preference in some modern societies? Although the ‘structural’ conditions in which son preference was originated have eroded, the related ‘cultural’ idea of boys providing higher utility for the family, etc., may have survived. Arnold [1] finds a high persistence of son preference even in the face of rapid modernization in developing countries. Bongaarts [3], on the other hand, presumes that as societies develop, son preference will decline and girls will be treated increasingly more equal.

In the reviewed studies, there are only scarce hints for girl preference. Some indication for a slight girl preference in Denmark is given by Jacobsen, M°ller and Engholm [16]. Such a finding might be explained by a new and more positive evaluation of the role of women in society in recent decades. Two other studies, one conducted during the Vietnam war in the United States [23], the other one in Israel [26], suggest that in times of military crisis, there is a slight preference towards daughters as to avoid loosing a son in combat.

Obviously, one needs to take into account regionally and historically specific characteristics of the populations analyzed. Brockmann [4], for example, argues that welfare policies mattered for the development of gender preferences in post-war Germany. As findings by Waller [27] substantiate, it should also be considered that generally the magnitude of the observed influences is rather small, even if they turn out to be statistically significant.

Reasons for gender preference Data and method

logo70.gif (2450 bytes)

Gender Preferences for Children in Europe: Empirical Results from 17 FFS Countries
Karsten Hank and Hans-Peter Kohler
© 2000 Max-Planck-Gesellschaft ISSN 1435-9871