The author is Assistant Professor of Economics at Oregon State University. She thanks Walter Galenson, George Staller, William Goldsmith, Kenneth Fraundorf, and two anonymous referees for their helpful advice and comments.

¹ U.S. Bureau of the Census, Historical Statistics of the United States (Washington, D.C., 1975), p. 133Google Scholar.

² Mincer, Jacob, “Labor Force Participation of Married Women: A Study of Labor Supply,” in Universities-National Bureau Committee for Economic Research, Aspects of Labor Economics (Princeton, 1962), pp. 63–95Google Scholar.

³ Cain, Glen, Married Women in the Labor Force (Chicago, 1966)Google Scholar; Bowen, William and Finegan, T. Aldrich, The Economics of Labor Force Participation (Princeton, 1968)Google Scholar. Recent work on labor force participation makes use of more complex estimation procedures and/or simultaneous models of the labor supply of husband and wife. Examples include Ashenfelter, Orley and Heckman, James, “The Estimation of Income and Substitution Effects in a Model of Family Labor Supply,” Econometrica, 42 (Jan. 1974), 73–85CrossRefGoogle Scholar; Kraft, Arthur, “Preference Orderings as Determinants of the Labor Force Behavior of Married Women,” Western Economic Journal, 11 (Sept. 1973), 270–84Google Scholar; and Wachter, Michael, “A Labor Supply Model for Secondary Workers,” Review of Economics and Statistics, 54 (May 1972), 141–51CrossRefGoogle Scholar. It was felt that the assumptions and data requirements of these models made them inappropriate to the turn of the century.

⁴ If the wives in these families were counted as in the labor force, the mean participation rate would increase substantially—from 9.31 to 30.83. U.S. Commissioner of Labor, Eighteenth Annual Report: Cost of Living and Retail Prices (Washington, D.C., 1904), p. 37Google Scholar.

⁵ Smuts, Robert, Women and Work in America (New York, 1959), p. 14Google Scholar.

⁶ U.S. Commissioner of Labor, Eighteenth Annual Report, p. 57.

⁷ In 1970 there were a total of 2,752,146 lodgers and 63,637,721 households, so the maximum percent of households taking in lodgers is 4.3 percent. U.S. Bureau of the Census, Census of Population, 1970: Vol. 1, Characteristics of the Population, (Washington, D.C., 1973), p. 649Google Scholar.

⁸ The Immigration Commission estimated that in one year, 1907, $275 million was sent to Europe by individuals in the United States. A part of this went to pay for others' passages to the United States; 30 percent of arriving immigrants reported that their passage had been paid for by relatives. U.S. Congress, Immigration Commission, Report of the Commission, 61st Congress, 3rd Session, (1911), Vol. 2, p. 426 and Vol. 3, p. 359.Google Scholar

⁹ These attitudes are documented in Smuts, Women and Work, pp. 110–42.

¹⁰ Ibid., p. 43.

¹¹ Calculation is based on data from U.S. Commissioner of Labor, Eighteenth Annual Report, pp. 29, 57.

¹² For a summary of the attitudes toward child labor as expressed in a large number of oral history interviews, see Hareven, Tamara K., “Work and Family Patterns of Immigrants in a Planned Industrial Town, 1900–1930,” in Ehrlich, Richard L., ed., Immigrants in Industrial America (Charlottesville, 1977), pp. 62–64Google Scholar.

¹³ Virginia Yans-McLaughlin, “A Flexible Tradition: South Italian Immigrants Confront a New York Experience,” in Ehrlich, Immigrants in Industrial America, p. 79.

¹⁴ Ibid., p. 77.

¹⁵ For statistics on levels of education, see the discussion of literacy later in this paper.

¹⁶ Italian consular officials reported that many immigrants “… regretted the fact that their children learned in school the selfishness of their North American classmates.” Harney, Robert F. and Troper, Harold, Immigrants: A Portrait of the Urban Experience, 1890–1930 (Toronto, 1975), p. 90Google Scholar. Just recently, similar fears about the influence of schools on traditional beliefs has led Russian Old Believers in Oregon to face jail sentences rather than send their own children to school past age 12.

¹⁷ Thernstrom, Stephan, Poverty and Progress (Cambridge, 1964), p. 24Google Scholar.

¹⁸ U.S. Bureau of the Census, Historical Statistics, pp. 15, 127, 131, and U.S. Bureau of Census, Census of Population: 1970, General Social and Economic Characteristics (Washington, D.C., 1972), p. 372Google Scholar.

¹⁹ Van Kleeck, Mary, A Seasonal Industry: A Study of the Millinery Trade (New York, 1917), p. 60Google Scholar.

²⁰ Smuts, Women and Work in America, p. 43.

²¹ Caroline Golab, “The Impact of the Industrial Experience on the Immigrant Family: The Huddled Masses Reconsidered,” in Ehrlich, Immigrants in Industrial America, pp. 20–21.

²² One woman clothing worker said, “‘I couldn't do as well if it wasn't for Jenny and Mame there. Mame has learned to sew on buttons first rate and Jenny is doing almost as well … Mame's seven and Jenny's going on six,’ said the mother, ‘but Jenny's the smartest. She could sew on buttons when she wasn't much over four.’” Campbell, Helen, Prisoners of Poverty (Boston, 1887), pp. 201–2.Google Scholar

²³ Butler, Elizabeth, Saleswomen in Mercantile Stores (New York, 1912), p. 150Google Scholar.

²⁴ Abbott, Edith, Women in Industry (New York, 1916), pp. 346–48.Google Scholar

²⁵ Manning, Caroline, The Immigrant Woman and Her Job (Washington, D.C., 1930), pp. 40–41Google Scholar.

²⁶ In a later study, Day Monroe found that 16.2 percent of the working women with children under age two, 39.2 percent of the women with children aged two to seven, and 48.2 percent of working women with children aged seven to fourteen were employed away from home. Monroe, Day, Chicago Families (Chicago, 1932), p. 220Google Scholar.

²⁷ Monroe notes, “Mothers with unbroken marital ties who had children sixteen or older earned slightly less often than did those having no children of this age, 12 percent of the former and 15 percent of latter being breadwinners. However, as the number of older sons and daughters increased the proportion of mothers earning fell, being but 8 percent of those having three or more children of this age at home. … This decrease in the likelihood of employment of the mother as the number of potential earners increased was due, undoubtedly, in part to a lessened need for her contributions to the income, though her age also may have been a factor.” Monroe, Chicago Families, pp. 223–24.

²⁸ Valerie Kincaid Oppenheimer's calculations show that occupations that were 90 percent or more female employed 38 percent of the female labor force in 1900 (versus 29 percent in 1960), whereas occupations that were 70 percent or more female employed 54 percent of all women workers (versus 50 percent in 1960). See Oppenheimer, , The Female Labor Force in the United States (Berkeley, 1970), p. 71Google Scholar.

²⁹ Although the importance of job opportunities was first noted by Nedra Belloc in her study of the labor force participation of women in 1940, it was Bowen and Finegan who first introduced a complete industrial structure variable into the regression. To calculate this variable, the proportion of female employment to total employment in each industry nationally was multiplied by the employment in that industry in the urban areas of each state. The results were summed over all industries and the total was divided by the total state urban employment. This variable is similar to that developed by Bowen and Finegan. Bowen and Finegan, Labor Force Participation, pp. 773–76.

³⁰ Both Cain and Bowen and Finegan find that the overall unemployment rate had a significant and negative effect on participation. Cain, Married Women, pp. 161, 191. However, Mincer pointed out the possibility of a spurious correlation between the overall unemployment rate and female labor force participation. Mincer, Jacob, “Labor Force Participation and Unemployment: A Review of Recent Evidence,” in Robert, and Gordon, Margaret, eds., Prosperity and Unemployment (New York, 1966)Google Scholar. Cain then experimented with other unemployment variables, including the male unemployment rate and still found significant negative correlations. Cain, Glen, “Unemployment and the Labor Force Participation of Secondary Workers,” Industrial and Labor Relations Review, 21 (Jan. 1967), 280, 282.Google Scholar

³¹ U.S. Bureau of the Census, Historical Statistics, pp. 379, 381.

³² Manning found that of 2,000 women interviewed about the reason for their choice of their first job, 1,035 cited the influence of relatives and friends, 346 said it was “the easiest to get,” and 73 valued the nearness of the job to their home. Manning, Immigrant Women, p. 103.

³³ U.S. Commissioner of Labor, Eighteenth Annual Report, pp. 110–31, 236–37, 244–49, 286–87, 366–67, and 370–71.

³⁴ Since this study was not specifically designed to focus upon the factors that influence wives to work, the sample population from which the observations were drawn was not universal—only urban working class families were included. Although at first this may seem to be too small a base to provide useful information, further consideration illustrates that it is not so restrictive. First because of the close intertwining of housework and production activities on family farms, such households have been excluded from all cross-section studies of the participation of married women. This study focuses, as have all others, on urban areas, where there are many more job alternatives for women. Second, the exclusive use of working class families eliminates only a relatively small part of the 1890–1920 urban population. Although it would be preferable to include all types of families, it was felt that the omissions were not so large as to impair seriously any conclusions about the labor force participation of married women.

³⁵ For more complete information on data sources and values, see the data appendix.

³⁶ This elasticity does not change much if wives who took in boarders and lodgers are included in the labor force. The income elasticity, evaluated at the mean, is then -1.418. Cain, Married Women, pp. 48, 54, 58–59 and Mincer, “Labor Force Participation of Married Women,” p. 72.

³⁷ Ben-Porath, Yoram, “Labor Force Participation Rates and the Supply of Labor,” Journal of Political Economy, 81 (May/June 1973), 697–704CrossRefGoogle Scholar. See also Gronau, Reuben, “The Effect of Children on the Housewife's Value of Time,” Journal of Political Economy, 81 (March/ April 1973), 530–47.CrossRefGoogle Scholar

³⁸ Heckman, James, “A Partial Survey of Recent Research on the Labor Supply of Women,” American Economic Review, 68 (May 1978), 204–05.Google Scholar

³⁹ Parker, John and Shaw, Louis, “Labor Force Participation Within Metropolitan Areas,” Southern Economic Journal, 34 (Apr. 1968), 545CrossRefGoogle Scholar.

⁴⁰ Butler, Elizabeth, Women and the Trades (New York, 1909), p. 225Google Scholar.

⁴¹ Mincer, “Labor Force Participation of Married Women,” p. 96.

⁴² Smuts, Women and Work, p. 155.

⁴³ Mary Van Kleeck observed this fall in wages whenever there was excess supply in the labor market. In her study of millinery workers she cites instances of women who had to accept pay cuts of up to $2 per week as “a premium to the firm for not laying her off.” Mary Van Kleeck, A Seasonal Industry, p. 44. In her study of flower makers, she notes that “going from shop to shop in search of a job they then begin to underbid each other.” Van Kleeck, Mary, Artificial Flower Makers (New York, 1913), pp. 88–89Google Scholar.

⁴⁴ In the regression, the coefficient on the mean income from boarders was -0.0519, with a standard error of 0.0285 and a t-value of 1.819. Compared to the regression on the entire sample, in the regression for the northern states, only the size of coefficient and the level of significance were reduced for “other family income” but raised for the male unemployment and homeowner variables. The remainder of the estimates were about the same as in the regression on the entire sample.

⁴⁵ Mincer, “Labor Force Participation of Married Women,” p. 75.

⁴⁶ No information was given on the number of families in the survey having children of different ages. However, those children who require the most care are young children who are home all day. Although the survey's “children at home” includes some older children (mostly daughters) who had completed their schooling but were not working, over 80 percent of the “children at home” were under five years of age.

⁴⁷ In 1890–91, the average earnings of working-class men rose from an average of $465.58 per year for men 20–24 to $592.44 per year for men 35–39. This increase of 27 percent compares to an increase of over 144 percent for men in 1970 (from $4,129 to $10,066). Figures for 1890–91 are based on data from U.S. Commissioner of Labor, Sixth Annual Report: Cost of Production: Iron, Steel, Coal, Etc. (Washington, D.C., 1892), pp. 705–28, 826–50, 943–54, 1005–25, 1099–1106, 1139–44Google Scholar and U.S. Commissioner of Labor, Seventh Annual Report: Cost of Production: The Textiles and Glass (Washington, D.C., 1892), pp. 905–81, 1253–87, 1422–59Google Scholar. Figures for 1970 from U.S. Bureau of the Census, Census of Population 1970: Vol. 1, Characteristics of the Population (Washington, D.C., 1973), p. 835Google Scholar.

⁴⁸ For example, Bowen and Finegan find that the presence of children under six has the largest and most consistently negative effect on women's participation. The presence of children 6–13 also results in lower participation rates compared to those of women with no children or those with only older children. However, the presence of children 6–13 has little additional effect on participation for those women with children under six too. The presence of children 14–17 alone also decreases participation, but by a smaller amount than do younger children alone. However, the presence of children 14–17 tends to increase the participation of women with younger children as well. From this, Bowen and Finegan conclude that older children are partial substitutes (or “assistants”) in housework. Bowen and Finegan, Labor Force Participation, pp. 98–100.