Human Reproduction, Vol. 15, No. 6, 1422-1423,
June 2000
© 2000 European Society of Human Reproduction and Embryology
Letters to the Editor |
1 Department of Epidemiology, Institute of Public Health, Faculty of Health Sciences – University of Copenhagen, Panum Institute Blegdamsvej 3, DK-2200 Copenhagen N, 2 Institute of Cancer Epidemiology, The Danish Cancer Society, Copenhagen, 3 Department of Epidemiology and Social Medicine, Aarhus University, Aarhus, 4 Centre for Research in Health & Social Statistics, The Danish National Research Foundation, Copenhagen, Denmark, and 5 Thames Cancer Registry, Guy's, King's and St Thomas' School of Medicine, London, UK
Dear Sir,
Dr James' comment addresses a statement that was made in the passing in Jacobsen et al. (1999a) on the absence of Lexis variation in our data set of 614 000 Danish children born 1980–1993, with information on biological maternal and paternal identities derived from registrations at the time of birth. The analysis has subsequently been published in full (Jacobsen et al., 1999b
). In contrast to our analysis, where only paternal age emerges as a significant determinant of the sex of singleton infants, the larger study of Biggar et al. (1999), using data on 1.4 million Danish children born 1960–1994, provides strong evidence for Lexis variation. Certainly, the larger study includes all the births of our investigation, and must hence be the more informative, despite some variation in the definition of parenthood and in the methods of analysis. As the period of sampling increases, not only are more children included, but the families tend to become larger and more complete, which increases the power to detect Lexis variation.
The issue of analysing a data set of largely incomplete families was central to our choice of using regression analysis of birth rates to study composition preference and sex preference in Jacobsen et al. (1999a) and to the discussion in that paper. We have more recently analysed a data set of 25 631 essentially complete maternal families, comprising 56009 children. These data were obtained in the context of a prospective study of diet and other factors on cancer risk; the interviewed mothers were born in the period 1929–1947 (Tjønneland et al., 2000). As it is conventional when working with data on completed families, we examined the ratio of families that continued reproduction to families that stopped reproduction (the continue/stop ratio), conditional on the sex-composition of children in the family. The results are plotted in Figure 1, in a format similar to Figure 1 in Jacobsen et al. (1999a).
|
Figure 1
shows the same combination preference as observed previously with the highest continue/stop ratios in even-sexed families (MM, FF, MMM, FFF) and the lowest continue/stop ratios in families that once were even-sexed, until the birth of a child of the missing sex (MMF, FFM). Sex preference for daughters is indicated by a higher continue/stop ratio in M and MM families compared with F and FF families respectively. As demonstrated in Jacobsen et al. (1999a), particularly in one-child families, this may, however, in part reflect the confounding effect of paternal age, which is associated with both low fertility and with a female-biased offspring sex ratio. The results of the analysis of completed families using the continue/stop ratio as the parameter of interest are thus consistent with the previous analysis based on incomplete families. It is of interest to note that the new analysis concerns events about one generation earlier than our first analysis. The composition preference for mixed sibships and the sex preference for daughters thus seem to have been established for some time in the Danish population.
Notes
6 To whom correspondence should be addressed 
References
Biggar, R.J., Wohlfahrt, J., Westergaard, T. and Melbye, M.(1999) Sex ratios, family size, and birth order. Am. J. Epidemiol., 150, 957–962.
Jacobsen, R., Møller, H. and Engholm, G. (1999a) Fertility rates in Denmark in relation to the sexes of preceding children in the family. Hum. Reprod., 14, 1127–1130.
Jacobsen, R., Møller, H. and Mouritsen, A. (1999b) Natural variation in the human sex ratio. Hum. Reprod., 14, 3120–3125.
Tjønneland, A.M. and Overvad, O.K. (2000) `Kost, kræft og helbred' – en befolkningsundersøgelse og biologisk bank i Danmark. Ugeskr Læger, 162, 350–354 .[Medline]
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||