Skip to main content Accessibility help

Socio-Economic Influences on School Attendance: A Study of a Canadian County in 1871

  • Frank T. Denton (a1) and Peter J. George (a1)


A RECENT SPECIAL ISSUE of the Quarterly (1) was devoted to “Education and Social Change in English-Speaking Canada”. In his interesting and impressive contribution to this issue, Michael B. Katz suggested that many hypotheses (2) concerning the determinants of school attendance may be tested by employing data relating to “the status and structure of the family and household” contained in the manuscript census. “By starting with the census one may study the gross patterns of attendance among the children of any group, religious, ethnic, occupational, or any other into which the census material can be arranged.” (3)



Hide All


1. History of Education Quarterly, 12, No. 3 (Fall, 1972).

2. For example, “school attendance varied inversely with the proportion of Irish immigrants in a community … [and] … directly with the proportion of the workforce employed in professional and commercial occupations.” Katz, Michael B., “Who Went to School?”, History of Education Quarterly XII, No. 3 (Fall, 1972): 435.

3. Ibid., p.437.

4. This project has been supported by the McMaster University Urban Research Unit, and has focussed on variations in family size and school attendance. Various aspects of this research project are discussed in Denton, Frank T. and George, Peter J., “An Exploratory Statistical Analysis of Some Socio-Economic Characteristics of Families in Hamilton, Ontario, 1871,” in Social History, 5 (April 1970): 1644, and “The Influence of Socio-Economic Variables on Family Size in Wentworth County, Ontario, 1871: A Statistical Analysis of Historical Micro-Data”, forthcoming in the Canadian Review of Sociology and Anthropology, (November, 1973).

5. Government of Canada, Census of Canada 1870–71, Volume I (Ottawa, 1873), pp. 89

6. The individual returns for Wentworth County are contained in Ontario Census 1871, Public Archives Microfilm Reel Nos. C-615 and C-616.

7. Surname and given names, sex, age, month of birth for children born within the previous twelve months, country or province of birth, religion, ethnic origin, profession, occupation or trade, marital status, whether or not married in the last twelve months, whether or not going to school, literacy, and infirmities.

8. The sampling was what is known as “random systematic”. That is to say, a random starting point from 1 to 10 was chosen in the case of Hamilton and from 1 to 5 in the case of the rest of the county. This defined the initial household in each case. Thereafter, every 10th or every 5th household was selected.

9. By far the largest proportion of dwellings remaining in the sample were single-family dwellings. However, dwellings containing two or more families were occasionally encountered and it was decided to include two or more “normal” families living in the same dwelling only if they were enumerated separately. Once the decision rule to include only “normal” families enumerated separately had been invoked, there still remained some difficulties with a few observations—problems of partial illegibility of surnames and given names, incomplete returns for the characteristics of a husband or wife in a family, apparent transcription errors in the recording of answers by enumerators, children of different surnames from the husband and wife, etc. Families were discarded from the sample when illegibility precluded the interpretation of data pertinent to the study, when the returns were incomplete for either the husband or wife, or when the returns were apparently incorrect in some respects. Families with children of different surnames were retained in the sample, and any children not over sixteen years were placed in a separate category referred to as “other dependents sixteen years and under”.

10. Blishen, Bernard R., “A Socio-Economic Index for Occupations in Canada”, Canadian Review of Sociology and Anthropology, 4 (1967): 4153. Also see Blishen, , “The Construction and Use of an Occupational Class Scale,” Canadian Journal of Economics and Political Science 24 (November, 1958): 519–31.

11. Whereas Blishen calculated an index of the socio-economic rank of occupations to two decimal points, we are mainly concerned with defining broad groups and have employed summary categories with arbitrary bounds. We certainly do not regard the 1961 occupational categorization as strictly appropriate for 1871; rather, it represents merely a rough initial guide for classifying 1871 occupations, the application of which has been tempered in many cases by our “judgment.” It may be noted that Katz used a different approach in devising a scale of occupations which he believed to be representative of mid-nineteenth century conditions. See Katz, Michael B., “Social Structure in Hamilton, Ontario,” in Thernstrom, Stephan and Sennett, Richard, eds. Nineteenth Century Cities: Essays in the New Urban History (New Haven, 1969), p. 25, and “Occupational Classification in History,” in The Journal of Interdisciplinary History, 3 (Summer, 1972): 63–88.

12. A general discussion of the use of dummy variables in regression analysis can be found in a number of places. A treatment that is particularly relevant to the present analysis is given in Melichar, Emanuel, “Least-Squares Analysis of Economic Survey Data”, Proceedings of the Business and Economic Statistics Section of the American Statistical Association, 1965: 373385.

13. See Melichar, , “Least Squares Analysis of Economic Survey Data”, for a discus-ion of this test. When the dependent variable is a dummy variable the assumption of normality of errors in the regression equation is violated and the F and t-tests, which are based on this assumption, must be regarded as approximate only. For a discussion of the problem, see the note by Orley Ashenfelter in Appendix A of Bowen, William G. and Finegan, T. Aldrich, The Economics of Labor Force Participation (Princeton, 1969), pp. 644648.

14. The true value of GTS must be zero or one. However, the value of GTS calculated from a regression equation will probably not be one of these values. Following common practice, we may interpret the calculated value as the probability that a person with given characteristics will be going to school, on the assumption that the value lies between zero and one. Although there is no guarantee that the calculated value will lie between these bounds in every case, in practice the assumption is a reasonable one for most cases.

15. Katz, , “Who Went to School?”, pp. 438, 443.

16. Ibid., p. 441.

17. Ibid., p. 439.

18. Katz's methodological dilemma is summarized in the following: “The obvious conclusion is that Irish Catholicism and Free Church Presbyterianism, respectively, retarded and promoted school attendance. However, this explanation takes no account of other factors; it could be that Irish Catholics were poor and Scottish Presbyterians prosperous, and therein lay the difference. The point is of some importance, for it raises the question to what extent school attendance was a product of religion and to what extent a result of economic factors.” Ibid., p. 439.


Full text views

Total number of HTML views: 0
Total number of PDF views: 0 *
Loading metrics...

Abstract views

Total abstract views: 0 *
Loading metrics...

* Views captured on Cambridge Core between <date>. This data will be updated every 24 hours.

Usage data cannot currently be displayed