1 See, for example, Brown, E. H. Phelps, The Inequality of Pay (Berkeley, 1977), chap. 1;Williamson, Jeffrey G. and Lindert, Peter H., American Inequality: A Macroeconomic History (New York, 1980);Lebergott, Stanley, The American Economy: Income, Wealth and Want (Princeton, 1976);Williamson, Jeffrey G., “The Structure of Pay in Britain, 1710–1911,” Research in Economic History, 7 (1982), pp. 1–54; Peter Lindert, “Who Owned Victorian England?” (Agricultural History Center, University of California-Davis, Working Paper Series, No. 2, Feb. 1983).
2 Taylor, A. J., ed., The Standard of Living in Britain in the Industrial Revolution (London, 1975), pp. xvii–xviii, xxix–xxxvii.
3 Williamson, Jeffrey, “Was the Industrial Revolution Worth It? Disamenities and Death in Nineteenth Century British Towns,” Explorations in Economic History, 19 (1982), pp. 221–45;Fogel, Robert W., Engerman, Stanley L., and Trussell, James, “Exploring the Use of Data on Height: The Analysis of Long-Term Trends in Nutrition, Labor Welfare, and Labor Productivity,” Social Science History, 6 (Fall 1982), pp. 401–21;Hollingsworth, J. Rogers, “Inequality in Levels of Health in England and Wales, 1891–1971,” Journal of Health and Social Behavior, 22 (09 1981), pp. 268–83;Kadin, Miriam L., “Modernization and the Social Inequality of Death in the United States, 1910–1970” (Ph.D. diss., Brown University, 1982), chap. 1;Preston, Samuel H., Haines, Michael R., and Pamuk, Elsie, “Effects of Industrialization and Urbanization on Mortality in Developed Countries,” International Union for the Scientific Study of Population, International Population Conference, Manila, 1981, (Liege, 1981), vol. 2, pp. 233–54.
4 Morawetz, David, Twenty-Five Years of Economic Development (Baltimore, 1977), pp. 47–50, 54–58; United Nations, Levels and Trends of Mortality since 1950 (New York, 1982), pp. 1, 60–64, and passim.
5 Sir Newsholme, Arthur, quoted in Titmuss, R. M., Birth, Poverty, and Wealth (London, 1943), p. 12.
6 Woodbury, Robert M., Infant Mortality and Its Causes (Baltimore, 1926).
7 Morawetz, Twenty-Five Years of Economic Development, chap. 3.
8 Antonovsky, Aaron and Bernstein, Judith, “Social Class and Infant Mortality,” Social Science and Medicine, 11 (05 1977), pp. 453–77; United Nations, Levels and Trends, pp. 60–65, 108–10, 138–41, 163–69, and The Determinants and Consequences of Population Trends: New Summary of Findings on Interactions of Demographic, Economic, and Social Factors (New York, 1973), vol. 1, pp. 132–42.
9 Arthur Newsholme, quoted in Antonovsky and Bernstein, “Social Class and Infant Mortality,” p. 453.
10 Antonovsky, Aaron, “Social Class, Life Expectancy and Overall Mortality,” Milbank Memorial Fund Quarterly, 45 (1967), pp. 66–68; Preston, Haines, and Pamuk, “Effects of Industrialization and Urbanization,” pp. 247–52.
11 United Nations, Levels and Trends, pp. 58–60. Morris, J. N. and Heady, J. A., “Social and Biological Factors in Infant Mortality—V, Mortality in Relation to Father's Occupation, 1911–1950,” The Lancet (03 12, 1955), p. 556; Antonovsky and Bernstein, “Social Class and Infant Mortality,” pp. 458–9.
12 Humphreys, N. A., “Class Mortality Statistics,” Journal of the Royal Statistical Society, 50 (06 1887) p. 282.
13 Seutemann, K., “Kindersterblichkeit sozialer Bevölkerungsgruppen, insbesonders im preussischen Staate und senien Provinzen,” in Neumann, F. J., ed., Beiträge zur Geschichte der Bevölkerung in Deutschland seit dem Anfange dieses Jahrhunderts (Berlin, 1894), vol. 5, pp. 69–167.
14 Morris and Heady, “Social and Biological Factors,” Fig. 3, p. 556; United Nations, Determinants and Consequences, pp. 138–39.
15 United Nations, Determinants and Consequences, p. 139.
16 Woodbury, Infant Mortality, Table 38.
17 United Nations, Determinants and Consequences, p. 137.
18 A problem arises because occupation of father or mother at the time of the record (death certificate or census enumeration) may not have been the same occupation when the child died. Further, the socioeconomic circumstances which contributed to child mortality may have been reflected by previous rather than current occupation. It is true, however, that occupation tended to be stable over short periods of time and that upward occupational mobility was not dramatic over brief periods.
19 Johnson, R. Christian, “The 1900 Census Sampling Project: Methods and Procedures for Sampling and Data Entry,” Historical Methods (Fall 1978), pp. 147–51; England and Wales, Registrar General, Census of England and Wales, 1911, vol. 13, “Fertility of Marriage,” Part 11 (London, 1923).
20 For example, Innes, J. W., Class Fertility Differentials in England and Wales, 1876–1934 (Princeton, 1938), pp. 41–52.
21 England and Wales, “Fertility of Marriage,” pp. vii–xi;Preston, Samuel H. and Haines, Michael R., “New Estimates of Child Mortality in the United States at the Turn of the Century,” Journal of the American Statistical Association, 79 (06 1984), pp. 272, 277–78. For a favorable evaluation of another sample from the 1900 American census, see Smith, Daniel Scott, “Differential Mortality in the United States before 1900,” Journal of Interdisciplinary History, 13 (Spring 1983), pp. 736–46.
22 United Nations, Indirect Techniques for Demographic Estimation, Manual X (New York, 1983), pp. 73–85.
23 Coale, Ansley and Demeny, Paul, Regional Model Life Tables and Stable Populations (Princeton, 1966), pp. 5–41.
24 Smith, Daniel Scott, “Differential Mortality in the United States before 1900,” Journal of Interdisciplinary History, 13 (Spring 1983), p. 14. The West Model works well for the total and white populations, but may not be as suitable for the black population. Preston and Haines, “New Estimates of Child Mortality,” p. 274;Zelnik, Melvin, “Age Patterns of Mortality of American Negroes: 1900–02 to 1959–61,” Journal of the American Statistical Association, 64 (06 1969), pp. 433–46.
25 For England and Wales, the standard life table chosen gives a mortality level slightly too favorable. The mortality index for the country as a whole is actually 1.06 instead of 1.00. But that is really only important as a scaling matter, on the assumption of linear relationships. Linear relationsips are a close approximation in the intermediate range of Coale and Demeny model life tables. For the United States, the standard life table chosen yields a mortality index for the country as a whole of 1.0088, which is very close to 1.00.
26 As a check on the data for England and Wales, the differentials were compared to those from the first vital statistics tabulation (1911) of the infant mortality rate by this social class categorization. (Great Britain, Registrar General, Seventy-Fourth Annual Report of the Register General of Births, Deaths, and Marriages in England and Wales, 1911 [London, 1913], pp. 73, 88.) The results are as follows:
The comparison is quite close, despite the fact that the census results apply to children as well as infants, while the vital statistics results apply only to infants. Further, the vital statistics data come from only one year of experience (1911) and are thus subject to more variability. In addition, a zero- order correlation between 116 detailed occupational categories of the census mortality index and the 1911 vital statistics infant mortality rate gave a quite close relationship (r = 0.960)
27 United Nations, Indirect Techniques, pp. 74–75, 83.
28 For example, Laurie, Bruce, Hershberg, Theodore, and Alter, George, “Immigrants and Industry: The Philadelphia Experience, 1850–1880,” Journal of Social History, 9 (12 1975), pp. 219–48;Haines, Michael R., “Fertility and Marriage in a Nineteenth-Century Industrial City: Philadelphia, 1850–1880,” This JOURNAL, 40 (03 1980), Tables 1 and 2.
29 For a discussion of problems with the 1911 English census social class categories, see Armstrong, W. A., “The Use of Information about Occupation,” in Wrigley, E. A., ed., Nineteenth Century Society: Essays in the Use of Quantitive Methods for the Study of Social Data (Cambridge, 1972), pp. 203–6.
30 Preston, Haines, and Parnuk, “Effects of Industrialization and Urbanization,” pp. 237–46.
31 Antonovsky and Bernstein, “Social Class and Infant Mortality,” Table 2.
32 England and Wales, Registrar General, Census of England and Wales, 1911, vol. 10, “Occupations and Industries,” Part II (London, 1914), Table 13.
33 Preston, Haines, and Pamuk, “Effects of Industrialization and Urbanization,” Tables I and 2. For England and Wales, the mortality index values were: London 1.06; county boroughs 1.23; other urban districts 1.03; rural districts .84.
34 United States Bureau of the Census, Twenty Censuses: Population and Housing Questions, 1790–1980 (Washington, D.C., 1979), pp. 35–36; Armstrong, “The Use of Information about Occupation,” pp. 191–92, 203–11, 226–28.
35 Income measures between the United States and England are not strictly comparable. American incomes were largely actual yearly incomes from the 1901 Commissioner of Labor Survey, while English incomes were full-time equivalent annual earnings. Nonetheless, the relative incomes, as given by the index, should be independent of level. The index reflects higher relative incomes among most English professionals.
36 For example, all teachers in the United States had a mortality index of 1.00, while native white teachers had an index of .79, still above the .57 value for English teachers. For physicians and surgeons, the index was .97 overall and .84 for native whites, in contrast to the .43 for England and Wales.
37 The difference between the child mortality indices was tested for statistical significance for these selected occupational groups. To do this, several steps were necessary. First, the index values were normalized so that the national average values were 1.0000. Second, these normalized values were converted to q(5) values by multiplying them by the q(5) value in the American model table (West model, level 13, both sexes combined assuming a sex ratio of 1.05 at birth; q(5) = .19119). This basically reduced both sets of differentials to a common mortality level. Third, it was assumed that these q(5) values approximately followed a binomial process where q(5) = p and σ = , where n was number of children ever born to that occupational group. Finally, the statistical difference of these values was tested using the formula: This yielded the following t-statistics:
All but textile workers and native white agricultural laborers showed differences which were significant at least at a 5 percent level (two-tailed test)
38 Preston and Haines, “New Estimates of Child Mortality,” pp. 278–79.
39 The index of dissimilarity and the Atkinson index are already weighted.
40 Higher fertility, of course, promotes higher child mortality through the adverse effects of close spacing and high parity on child survival.
41 The effect of social class mortality rates versus the distribution of children ever born (that is, the “weights”) may be seen in the following example:
The use of weights reduces but does not eliminate the gap in the inequality measure. On the other hand the use of English social class child mortality rates with American weights results in a coefficient of variation (.30) far above that for the case with English weights and American rates (.11). The same is true when the 1911 English classes are used to calculate the coefficient of variation.
42 Keyfltz, Nathan and Flieger, Wilhelm, World Population: An Analysis of Vital Data (Chicago, 1968), p. 526;Mitchell, B. R. and Deane, Phyllis, Abstract of British Historical Statistics (Cambridge, 1971), pp. 36–43;Haines, Michael R., “The Use of Model Life Tables to Estimate Mortality for the United States in the Nineteenth Century,” Demography, 16 (05 1979), Table 7, gives an estimate of 46.7 for the U.S. model and 47.8 for the West model. Preston and Haines, “New Estimates,” Table 1, gives an estimate of expectation of life at birth of 50.1, consistent with child mortality in the West Model system.
43 Kuznets, Simon, Modern Economic Growth: Rate, Structure, and Spread (New Haven, 1966), pp. 107–8, 272.
44 Williamson and Lindert, American Inequality, pp. 281–85; Williamson, “The Structure of Pay in Britain,” pp. 22–24.
45 Preston, Haines, and Pamuk, “Effects of Industrialization and Urbanization,” pp. 251–52; United Nations, Determinants and Consequences, pp. 279–81.
46 Hollingsworth, “Inequality in Levels of Health,” pp. 279–81.
47 Shryock, Richard H., The Development of Modern Medicine: An Interpretation of the Social and Scientific Factors Involved (New York, 1947), chap. 15;Benjamin, B., “The Urban Background to Public Health Changes in England and Wales, 1900–1950,” Population Studies, 17 (1964), pp. 225 and passim.
48 Williamson, “The Structure of Pay in Britain,” pp. 36–37.Routh, Guy, Occupation and Pay in Great Britain, 1906–1960 (Cambridge, 1965), pp. 60–61.
49 United States, Bureau of the Census, Historical Statistics of the United States: Colonial Times to 1970 (Washington, D.C., 1975), vol. I, Series D, pp. 85–86.
50 The groups actually included were: Farmer's Sons; Officers of Local Authorities; Goldsmiths and Silversmiths; Watchmakers and Clockmakers; Bakers and Confectioners. Missing were such groups as: Coffee and Eating House Keepers; Inn and Hotel Keepers; Publicans; Boarding and Lodging House Keepers; Dealers and Merchants in various products (for example, coal, timber, wood, cork, bark, boots and shoes, corn, flour, seed); Drapers; General Shopkeepers; Grocers; Greengrocers; Tobacconists; Milksellers and Dairymen; Cheesemongers and Buttermen; Fishmongers, Poulterers and Game Dealers; Clothiers and Outfitters.
51 Shryock, The Development of Modern Medicine, chap. 15.
52 A more precise correction for heteroscedasticity is to use the square roots of children ever born as weights. This would be less valuable in obtaining weights representative of the population. The use of square roots actually makes little difference in the final results.
53 Regressions were run using the two different social class groups as independent variables and three earnings variables (earnings, inverse of earnings, and log of earnings) as dependent variables for both ordinary and weighted least squares. The results, in terms of R2, ranged from .452 to .867. Only I of the 12 equations had an R2 below 0.50.
54 The 1950 American occupational groupings were not included in Table 5 since the 1911 English social classes performed so well.
55 Regressions similar to those mentioned in footnote 53 were run for England and Wales using the 1911 social class group dummies as independent variables and three earnings variables (earnings, inverse of earnings, and log of earnings) as dependent variables, using both ordinary and weighted least squares. R2 values ranged from .546 to .707.
56 The coefficients of variation for the mortality index in the regression models were:
57 Grossman, Michael, The Demand for Health: A Theoretical Investigation (New York, 1972), pp. xiii–xvii.