Research Guide to the Russian and Soviet Censuses

Ralph S. Clem, Ralph S. Clem

• Publisher: Cornell University Press
• Year: 2016
• ISBN:
  9781501707070
• Topic: History
• Subtopic: Russian History
• Index: History

Taken together, the Russian census of 1897 and the Soviet censuses of 1926, 1959, 1970, and 1979 constitute the largest collection of empirical data available on that country, but until the publication of this book in 1986, the daunting complexity of that material prevented Western scholars from exploiting the censuses fully. This book is both a guide to the use of and a detailed index to these censuses. The first part of the book consists of eight essays by specialist on the USSR, six of them dealing with the use of census materials and the availability of data for research on ethnicity and language, marriage and the family, education and literacy, migration and organization, age structure, and occupations. The second part, a comprehensive index for all the published census, presents more than six hundred annotated entries for the census tables, a keyword index that enables researchers to find census data by subject, and a list of political-administrative units covered in each census.

Part One

GENERAL AND
TOPICAL ESSAYS

Chapter 1

On the Use of Russian and Soviet Censuses for Research

Ralph S. Clem

The quest for more data and better data is a perennial one in demography, as in all science. The improvement in demographic and related data will undoubtedly, however, continue to be limited by the changing conceptual framework and research interests of the demographer; the techniques available for obtaining the information with desired reliability, validity, and precision; and the values of differing cultures which dictate what information may be politic or impolitic to obtain. (Hauser and Duncan, 1959b: 55)

Seeking better to understand human societies, social scientists constantly confront the question of how to investigate socioeconomic and political issues. If these issues suggest themselves and are properly framed, the main problem then becomes the manner in which evidence can be marshaled, evaluated, and used analytically.

Although there are several generic types of social science data (e.g., surveys, opinion polls, registration systems), in practice the primary source of empirical information on a society is almost always the national census of population. The United Nations defines a census of population as “the total process of collecting, compiling and publishing demographic, economic and social data pertaining, at a specified time or times, to all persons in a country or delimited territory” (1958: 3). In fact, the attributes that define a census, universality, simultaneity, enumeration and compilation of individual characteristics, and specified territory, are collectively the reason why census data are so useful.

As Kingsley Davis (1966) put it, the census typically “reveals not only the basic demographic trends, such as population growth, internal population redistribution, urbanization and alterations in the age and sex structure, but also contributes indispensably to a knowledge of changes in the nation’s occupational and industrial composition, in its level of living, education and employment.” To this list of census topics one might add ethnic and racial composition, marriage and family structure, religion, and language use. It should be evident from the range of subjects covered in censuses that the information in them is of use not only in the study of demography (narrowly or broadly defined) but also in the study of anthropology, economics, education, ethnic studies, geography, history, political science, sociology—indeed any discipline concerned with the numbers and characteristics of a society or socioeconomic group (Hauser and Duncan, 1959a: 37–43).

As the social sciences have matured in the era after World War II, the demand for more and better data referred to by Hauser and Duncan has grown commensurately. In addition, bureaucrats and politicians have increasingly realized that knowledge, if not a complete understanding, of socioeconomic conditions and trends as manifested in census data is a useful tool for policymakers. In the postwar period there has accordingly been a rapid expansion and technical upgrading of census taking worldwide, an activity promoted and aided by the UN through its various agencies (UN, 1958; 1967; 1977). Consequently, the volume of census information available to the researcher has grown tremendously over the last twenty years (University of Texas, 1965; Goyer and Domschke, 1983). Likewise, greater interest has focused on censuses taken before World War II, as more and more researchers have come to appreciate the value of such enumerations for period or longitudinal studies. In both the historical and contemporary contexts, the acquisition, cataloging, and preservation of national censuses, such as the program at the University of Texas Population Research Center and the Census Library Project of the Library of Congress and Bureau of the Census, have been invaluable (Goyer, 1980; Dubester, 1969).

The use of census data is not without problems, however; in many respects, just the opposite is true. Despite the potential implied by the growing quantity of empirical information in the censuses of various countries, the researcher must take a wide array of troublesome difficulties into account when using census figures. Hauser and Duncan (1959b) characterize these difficulties as threefold: coverage, comparability, and quality. In the first instance, they refer to “the completeness with which the target population or class of events is enumerated or registered and the range of information about the unit [ordinarily, the individual] of observation” (Hauser and Duncan, 1959b: 65). Thus, one might be concerned about the accuracy of the count in terms of individual enumeration and also about the coverage across the range of subject matter.

Problems of comparability among censuses and between censuses and other sources of data are often so extensive as to render impossible meaningful direct comparisons over time and space. Unfortunately, such pitfalls as different definitions for categories and units are not always recognized, and unwarranted conclusions based on an imperfect understanding of the data are not unknown. At the very least, considerable work is often required to establish some useful standard of data comparability.

Finally, the issue of data quality—the extent to which the data contain errors of coverage, response, recording, and processing—forces the user to evaluate the figures given and assess their accuracy (if only generally) before drawing judgments from them. Recognizing the immensity of effort involved in conducting a national enumeration, the difficulties in making such concepts as “migration” operational in census questionnaires, and the possibilities for misunderstanding questions or misstating answers, one should approach even the most technically sophisticated census with skepticism.

Over the years demographers and other social scientists have developed many methods for recognizing and, in some cases, overcoming census-data problems. The variety and complexity of these methods are far too vast to be recounted here; instead, the reader is referred to several of the standard works on the subject (Jaffe, 1951; Barclay, 1958; Shryock and Siegel, 1980; Coale and Demeny, 1966).

Census Data in Russian and Soviet Studies

As in most countries, the published censuses of Russia (1897) and the USSR (1926, 1959, 1970, and 1979) constitute by far the largest and potentially the most useful collection of data on that society.¹ Modern Russian and Soviet enumerations since 1897 cover about one-sixth of the world’s land area and provide considerable information on many aspects of the country’s population, including age and sex composition, ethnic and language identification, marriage and family structure, migration patterns, urban and rural residence, educational attainment and literacy levels, and occupations. There are, of course, other important sources of quantitative data in Russian and Soviet studies, many of which relate to and are used in conjunction with census data; these sources will be described below.

Given the wealth of information in these volumes, it may at first seem surprising that the Russian and Soviet censuses have been substantially underutilized by researchers, both in works that focus exclusively on Russia or the Soviet Union and in comparative studies. This situation is in part attributable to the fact that in the West the field of Russian and Soviet studies is overwhelmingly nonempirical in its topical foci and research methodology; the field is dominated by nonquantitative history, literature, politics, and language studies. A review of American and Canadian doctoral dissertations for the 1970s revealed that only about 10 percent dealt with subjects even remotely quantitative in nature; in fact, almost five times as many dissertations were written in literature as in economics, geography, and sociology combined (Dossick, 1971–80).² Thus, a relatively small number of researchers are trained in the empirical social sciences in Russian and Soviet area studies. This pronounced nonquantitative bias reduces the demand for and utilization of census data (or other types of socioeconomic data). Moreover, this topical tendency is inertial, for each new generation of scholars finds coursework and mentors concentrated in the dominant nonquantitative disciplines.

One can surmise that language barriers, the added work involved, and a general unfamiliarity with sources have contributed to the dearth of comparative or conceptual studies—the most likely types to require the use of census data—that treat Russia and/or the USSR. Beyond these practical concerns, however, is the important, yet subtle influence of ideology and politics on the objective analysis of Soviet society. As Lewis, Rowland, and Clem (1976: v) explained, “when one specializes in a region or country in any of the social sciences, there seems to be a natural inclination to treat that region as if it were a special or unique case in terms of societal processes, and this inclination is particularly strong if that region has a totalitarian government.” With the nonquantitative orientation of Russian and Soviet studies and the unwillingness of most scholars to engage in comparative studies across cultural, national, and ideological boundaries, it is perhaps not surprising that few have chosen to delve into the Russian and Soviet census volumes.

Using Russian and Soviet Census Data: General Problems

Although the Russian and Soviet censuses hold great promise for research into historical and contemporary socioeconomic issues, major problems are connected with the use of these data. Of course, these problems may have discouraged potential users and contributed to the chronic underutilization of the censuses. Nevertheless, the use of census data in any context involves difficulties, and those confronting the researcher in Russian and Soviet studies are generally not insurmountable.

Of the census problems included in Hauser and Duncan’s typology, by far the most serious involving the Russian and Soviet enumerations are those of comparability. As Robert Lewis discusses at length in Chapter 2, problems of temporal, definitional, and geographic comparability must be dealt with before one can use the Russian and Soviet data meaningfully. Temporal comparability is vexatious owing to the irregular intercensal periods in Russia and the USSR, which are an impediment especially to cohort analysis. Definitional comparability problems arise when questions and categories of responses fail to translate reality into operational terms. The investigator must then find some way to reorder the data into comparable (or at least similar) definitions or commit the sin of comparing “apples and oranges.” For example, in their study of historical and geographic trends in urbanization in Russia and the USSR, Lewis and Rowland (1969) found it necessary to reorder all data on urban places into a standard operational definition of “urban” to take into account major differences in usage from census to census. To be sure, in some cases such a reordering is simply not possible, and the data must be used with appropriate caveats for changing definitions and an assessment of the extent to which these differences influence the analysis.

Finally, changes in the national territory and the internal political-administrative unit structure of Russia and the USSR have created significant problems of geographic comparability in the use of census data. As detailed by Lewis, such difficulties can be overcome only through determined efforts. In several studies researchers have compiled data based on a consistent unit structure to facilitate historical and longitudinal analysis of demographic and socioeconomic characteristics and trends. Lorimer (1946: 241–49) reordered data from the 1897 and 1926 censuses into a set of “study areas” based on 1939 political-administrative units. With these figures, he analyzed regional population changes (including both the urban and rural components) for the periods 1897–1926 and 1926–39 (Lorimer, 1946: 150–72). However, because the territory of the USSR was appreciably smaller in 1939 than it was in the census years of 1897, 1959, 1970, and 1979, Lorimer’s units and figures cannot be used for comparisons among all censuses; nevertheless, they are of great value as a check on the validity of other methods.

In the most extensive undertaking of its kind, Leasure and Lewis (1966) derived population estimates for nineteen geographically comparable regions (the major economic regions of 1961) for the censuses of 1897, 1926, and 1959; with a few modifications, data from the 1970 and 1979 censuses can be retrofitted into these units. In addition to the total population, the Leasure and Lewis project also generated estimates for the urban population (based on a standard definition), for the rural population, and for a variety of socioeconomic characteristics. These data were utilized in the study of migration (Leasure and Lewis, 1967; 1968), urbanization (Lewis and Rowland, 1969), ethnicity (Lewis, Rowland, and Clem, 1976), and population redistribution (Lewis and Rowland, 1979). Using essentially the same method, Clem (1977) developed a set of population estimates for 1926, 1959, and 1970 which had the 142 oblast and equivalent level units of 1959 as the standard geographic framework. These units can be aggregated into the larger economic regions employed in the Leasure and Lewis study.

In another major study, conducted at the Office of Population Research at Princeton University, Coale, Anderson, and Härm (1979) reconstructed the population in 1926 within both 1897 and 1959/70 unit boundaries. Th...