Published online by Cambridge University Press: 15 May 2009
1. The Introduction (pp. 67–73) describes the course of infant mortality in England and Wales over the past century, and critically reviews arguments advanced to prove that variations in the infant mortality rate (i.m.) are caused by genetic differences with respect to viability.
2. The i.m. in county boroughs shows a 2-yearly cycle of variability, affecting places with high mortalities.
3. We have devised and tested various social indices and have selected five which gave the highest joint covariance with infant mortality in county boroughs during the 11 years 1928–38. These indices are:
H, percentage of families living more than 1 person per room.
U, percentage of men unemployed.
P, percentage of occupied males in the Registrar-General's Social Classes IV and V.
F, percentage of women employed on manufacturing processes.
L, latitude.
4. We have computed multiple regression equations involving i.m. and the five indices for each of the 11 years, and two summarizing equations. The regressions plus sampling variance account for about 80% of the total variance in infant mortality. The regression is linear.
5. Latitude does not affect infant mortality in Class I. For this and other reasons we regard the latitude effect as expressing miscellaneous poverty indices omitted from our equations.
6. The regression equations enable us to divide the population into various strata with characteristic average infant mortality rates. These include:
‘Better off’ (all poverty indices = 0) i.m.=23·1
Overcrowded poor i.m. = 108
Unemployed overcrowded poor i.m. = 153
Babies whose mothers work in industry suffer an additional mortality risk of at least 35 per 1000, and possibly more. The figure 23·1 is the i.m. rate that would prevail if our five poverty symptoms could be eliminated.
7. In county boroughs two-thirds of infant deaths would be avoided by the abolition of conditions defined by our indices. Of the preventable deaths, one-third are associated with overcrowding, one-quarter with low-paid occupations, one-fifth with unemployment, and one-eighth with industrial employment of women. In England and Wales, over 250,000 deaths in 11 years, about 63% of the total, can be attributed to adverse social conditions.
We have to thank the Rockefeller Foundation for a personal grant to one of us (J. W.) out of an allocation for research work in Prof. Lancelot Hogben's Department, and the Halley Stewart Trust for a grant for mechanical computing equipment. Our thanks are also due to the various members of the Zoology Department in the University of Birmingham who assisted in computing at various times, and especially to Prof. Hogben for his unfailing interest, advice and support.