The author is Assistant Professor of Economics, Stanford University. The research upon which this article is based has benefitted from the guidance and advice of many individuals including Samuel Bowles, Stanley Engerman, Albert Fishlow, Herbert Gintis, Charles Gulick, Neil Smelser, Richard Sutch, David Tyack, Lloyd Ulman, Gavin Wright, and Jeffrey Williamson. Innumerable conversations with these and other individuals helped greatly to clarify the arguments and interpretations advanced herein.

¹ Katz, Michael B., The Irony of Early School Reform: Educational Innovation in Mid-Nineteenth Century Massachusetts (Cambridge, Mass., 1968), p. 169Google Scholar.

² Field, Alexander J., “Sectoral Shift in Antebellum Massachusetts: A Reconsideration,” Explorations in Economic History, 15 (Apr. 1978), 146–53.CrossRefGoogle Scholar

³ Field, “Sectoral Shift,” 166–67.

⁴ Such a dynamic model of reform would imply that the local determinants of reform at a given time were sufficiently idiosyncratic (a crusading local leader, for example) that one could not expect a general model using broad economic and demographic variables to be of much explanatory value.

⁵ Field, Alexander J., “Industrialization and Skill Intensity: The Case of Massachusetts,” Journal of Human Resources, 14 (1979) (forthcoming)Google Scholar.

⁶ Shorter, Edward, The Making of the Modern Family (New York, 1975), pp. 46, 64–65, 218–27Google Scholar. See also Warner, Sam Bass, Streetcar Suburbs (Cambridge, Mass., 1962), p. 12Google Scholar.

⁷ Field, “Sectoral Shift,” 158–61.

⁸ Given existing levels of rule abidance, changes in punishment levels or probabilities of detection will probably induce changes in the frequency of rule violation. But microeconomic theory is no more successful in explaining why an observed level of rule violation, given enforcement levels, is not higher or lower, than it is in explaining why preferences are what they are. Another way of phrasing this is to ask why, given roughly similar enforcement levels, crime rates in Tokyo are so much lower than they are in New York. See Gary Becker, “Crime and Punishment: An Economic Analysis,” Journal of Political Economy, 76 (Mar.-Apr. 1968), 169–216, for an attempt to deal with crime as an example of rule breaking within the framework of microeconomic theory.

⁹ Contingent because their success depends on others pursuing them. The term “generalized reciprocity” is drawn from the anthropological literature to describe situations in which goods are exchanged over time in the absence of an immediate quid pro quo for each transaction.

¹⁰ Indeed, traditional communities with relatively low levels of occupational specialization tend to have few individuals, such as policemen, whose sole occupation is the maintenance of rule-abiding behavior.

¹¹ These included (for both 1855 and 1865) the length of school session in months, average attendance as a share of the school age population, and a combined measure: total student-days attended as a share of the maximum number of potential student-days (all school-age children attending for twelve months). The reported results concentrate on length of school session because the empirical regularities associated with this variable are so striking. With some exceptions, the attendance variable and (largely as a result) the potential student-day variable are not systematically related to the economic and demographic variables examined here. Results for 1855 rather than 1865 are reported primarily because of the unique political conditions (discussed later in this paper) that prevailed during the former year.

¹² Obviously this control was not unlimited. Owners of real estate as well as those of business and commercial property had to pay local property taxes to support these schools.

¹³ A compulsory education law was passed in 1852, but as the variance in attendance rates indicates, it was of varying effectiveness. It is revealing that of the 93 towns that had hired truant officers by the year 1866, 71 were classified as manufacturing towns in the 1865 census summary. Ninetysix of the 330 cities and towns in Massachusetts were classified as manufacturing in this abstract. If the establishment of a truant officer had been completely unrelated to a locality's economic structure, one would have expected about 27, as compared with 71 of the 93 truant-officer towns to be classed as manufacturing. Establishment of truant officer as reported in the Thirtieth Annual Report of the State Board of Education … 1866–67 (Boston, 1867), p. 74. Manufacturing towns as defined in Warner, Oliver, Abstract of the Census, 1865 (Boston, 1867), pp. 145–65Google Scholar. See Field, Alexander J., “Educational Reform and Manufacturing Development in Mid-Nineteenth Century Massachusetts,” Ph.D. dissertation, University of California, Berkeley, 1974, p. 41Google Scholar.

¹⁴ In using 1860 data for the merchants and the farmers variables, we are assuming that the 1860 values are relatively highly correlated with the unobserved 1855 values. The correlation between population in 1855 and population in 1865, for example, is 0.9898. The intercorrelation of the 1855 and 1860 data for these variables may not be this high, but these observations are the best available proxies for observations not obtainable from published sources. In the results reported in my dissertation, I used agricultural data from the 1865 Massachusetts census. The use of 1860 data is a minor improvement. The full sample for the dissertation runs was 330, rather than 329 towns, due to one town for which 1860 data were unavailable.

¹⁵ Commonwealth of Massachusetts, Abstract of the Census, 1865 (Boston, 1867), pp. 148–64.Google Scholar

¹⁶ Thernstrom wrote in his study of Newburyport: “… Irish working class families were especially successful in accumulating property, but especially unsuccessful in climbing out of the low status manual occupations.… Property mobility often … was achieved by sacrificing the education of the younger generation.” Thernstrom, Stephan, Poverty and Progress: Social Mobility in a Nineteenth Century City (Cambridge, Mass., 1964), p. 156Google Scholar. See also, Warner, Streetcar Suburbs, p. 9.

¹⁷ See Field, “Educational Reform,” p. 54, for the results of an analysis of the occupation of school committee members in Lowell, Massachusetts between 1826 and 1855. Of the eighty men who served at various times during this period, forty-one, or more than one in two, were either clergymen, doctors, or lawyers. Of the 176 person-terms during this period, 103 are accounted for by the above three occupational categories. These results are consistent with studies of the social composition of U.S. school boards conducted at other times and places. See Chambers, W. W. Jr., “Social Class Analysis and the Control of Public Education,” Harvard Educational Review, 23 (Fall 1953), 268–83Google Scholar.

¹⁸ The strong negative correlation of density with the share of the population in farming (−0.4799) confirms this interpretation of the variable. On Massachusetts farms it was rare for unrelated families to live under the same roof.

¹⁹ This conclusion was reached by both Maris Vinovskis and this author after extensive reviews of these debates. The main documents examined in both cases were the annual reports of the State Board of Education and the annual reports of the local school committees. Horace Mann's Fifth Annual Report included the results of a survey of businessmen concerning their views on the economic value of education, in which they repeatedly stressed the socialization function. Discussion of the socialization function is also particularly vivid in the annual reports of towns dominated by large-scale manufacturing, as noted in the text. See Vinovskis, Maris, “Horace Mann on the Economic Productivity of Education,” New England Quarterly, 43 (Dec. 1970), 565CrossRefGoogle Scholar; Field, “Educational Reform,” pp. 33–35; Field, Alexander J., “Educational Expansion in Mid-Nineteenth Century Massachusetts: Human Capital Formation or Structural Reinforcement?” Harvard Educational Review, 46 (Nov. 1976), 550–51.CrossRefGoogle Scholar

²⁰ Field, “Educational Reform,” Table B.4, pp. 323–24. The correlation of PCMN55 with POP55 was even lower (0.0502), confirming that manufacturing in Massachusetts in 1855 was not strictly an urban phenomenon. There are of course a number of other ways whereby this relationship can be captured. The Irish and density variables, for example, are to some extent proxies for industrial composition. The relationship of industrial variables to these variables and to the dependent variables is explored in more detail in Field, “Educational Reform,” pp. 307–14.

²¹ The strongest simple correlation between PCAT55 and a great many measures of manufacturing and industrial composition was 0.17. This particular measure aggregated employment in a number of high skill artisan industries: metal working, machinery, transport equipment, jewelry and woodworking, and expressed this as a fraction of the total manufacturing work force. Industrial employment levels were tried individually (at about the 2-digit SIC level) and using various aggregations, with the denominator in some cases population, and in some cases the total manufacturing work force.

²² Kaestle, Carl and Vinovskis, Maris, “Education and Social Change in Nineteenth Century Massachusetts: Quantitative Studies,” National Institute of Education Final Report (Dec. 1976)Google Scholar; a revised version of this report is to be published in 1979 by Cambridge University Press.

²³ “However, there were no very strong relationships between school attendance and any one of the independent variables.” Kaestle and Vinovskis, “Quantitative Studies,” p. 96.

²⁴ Field, “Educational Reform,” Appendix B, pp. 307–27. See also Field, Alexander J., “Educational Reform and Manufacturing Development in Mid-Nineteenth Century Massachusetts,” this Journal, 36 (Mar. 1976), 265Google Scholar.

²⁵ Kaestle and Vinovskis, “Quantitative Studies,” Table No. IV–21.

²⁶ Kaestle and Vinovskis do not have actual data on percent foreign born for 1860, and thus use a linear interpolation based on 1855 and 1865 data. But I found a correlation of 0.2634 between SES65 and PCIR65, confirming the distinctiveness of the 1855 observation.

²⁷ DeWitt, Francis, Abstract of the Census, 1855 (Boston, 1856)Google Scholar.

²⁸ Mulkern, John R., “The Know-Nothing Party in Massachusetts.” Ph.D. dissertation, Boston University, 1963, pp. 1–5, 113Google Scholar. The Know-Nothing Party acquired its name because its members initially denied knowing anything about the party's activities.

²⁹ Mulkern, “Know-Nothing Party,” pp. 37, 51.

³⁰ Mulkern, “Know-Nothing Party,” pp. 112–15.

³¹ See Messerli, Jonathan, “Controversy and Consensus in Common School Revival,” Teacher's College Record, 66 (May 1965), 749–58.Google Scholar

³² Katz, Irony, p. 1.

³³ Bowles, Samuel and Gintis, Herbert, Schooling in Capitalist America: Educational Reform and the Contradictions of Economic Life (New York, 1976), p. 235Google Scholar.

³⁴ Carlton, Frank Tracy, Economic Influences upon Educational Progress in the United States, 1820–1850 (Madison, 1908)Google Scholar; Simons, A. M., Class Struggles in America (Chicago, 1909)Google Scholar.

³⁵ Katz, Irony, pp. 19–22.

³⁶ When a region is occupied by a foreign nation, the local inhabitants often resent the schools the occupier may impose as much as they do the troops. This seems to be the analogy behind some of the objections to socialization: children were the victims of an ideological blitzkrieg.

³⁷ “Schools cannot be considered repressive merely because they induce children to undergo experiences they would not choose on their own, or because they impose forms of regimentation which stifle immediate spontaneity. Schools … are intrinsically constraining.” Bowles and Gintis, Schooling, p. 272.

³⁸ Gitelman, H. M., “No Irish Need Apply,” Labor History, 14 (1973), 56–65CrossRefGoogle Scholar.

³⁹ Thernstrom, Stephan, The Other Bostonians (Cambridge, Mass., 1973), p. 259CrossRefGoogle Scholar.

⁴⁰ Logically, one can set forth a morality play, or argue inevitably, but not both, for if people have no choice, then one can make no moral judgments on their actions.

⁴¹ For Latin America, See Rockland, Michael Aaron, introduction to Domingo Sarmiento, Travels in the United States in 1847, translated by Rockland, Michael Aaron (Princeton, 1970), pp. 19–75Google Scholar.

⁴² Ravitch, Diane, The Revisionists Revised: A Critique of the Radical Attack on the Schools (New York, 1978)Google Scholar.

⁴³ Curti, Merle, The Social Ideas of American Educators (New York, 1935), pp. 113–14.Google Scholar