Information potential of the database on the 1959 All-Union Population Census as a source for historical and demographic studies of the urban environment

Bondar Valerii Aleksandrovich PhD in History Associate Professor; Department of Documentation, Archival Science and History of Public Administration; Department of History; Ural Humanitarian Institute; Ural Federal University named after the first President of Russia B. N. Yeltsin 620083, Russia, Sverdlovsk region, Yekaterinburg, Turgenev str., 4, room 482 bondar1213@yandex.ru

Features of the preparation and conduct of the 1959 census

The middle of the XX century became the time of post-war reconstruction for Russia, which was focused on all available resources, primarily labor. The task was so large-scale that many tasks that seemed secondary were naturally pushed into the background. In particular, this applies to population censuses. The materials on the basis of which the database in question is formed relate to the unique census of 1959, conducted twenty years after the previous one and fourteen years after the end of the Great Patriotic War. This is a long time in comparison with European countries, where censuses were conducted as quickly as possible after the end of the World War (1945-1951) [See: 7].

One of the possible reasons is the fear of the consequences of making public the scale of human losses in the last war. Back in 1947, the head of the Central Statistical Department of the USSR State Planning Committee raised the question of conducting not a census (considering it impractical), but only a special population accounting, data on which were required for economic purposes. The refusal to conduct the census was justified by two reasons: the instability of the territorial distribution of the population and the possibility of not publishing the results of the accounting ^[4]. As a result of consideration of the proposal, it was decided to postpone any such event "until the end of the five-year plan" (1950) ^[7].

The political situation associated with Stalin's death and subsequent changes in the country's leadership also played a role. In 1953, the question of obtaining population data was raised again, without conducting a population survey – on the basis of voter lists, "adding the population under 18 years old in the village according to village councils, and in the city according to house books" ^[5]. These measures were implemented because population data were required for the needs of public administration. A little later, several more information notes were compiled and sent to the country's leadership on issues of population registration and related activities. The official proposal to conduct a census was received only in 1956. Over the next two years, census measures were developed and coordinated, and other preparatory work was carried out.

One of the main events of the census was the All-Union Congress of Statisticians in 1957, during which a number of proposals were made. In particular, it was proposed to expand and improve the planned census program: take into account not the age, but the date of birth, introduce four categories of marital status, dividing unmarried into widows, divorced and never married, take into account the number of children born to women. In order to save money, some of the data was allowed to be collected selectively, and several such proposals were put forward. Some ideas had to be implemented for decades, for example, combining the population census with housing conditions was implemented only in 1989.

Despite the large number of suggestions and wishes, the approved census program differed only slightly from the one developed for the 1939 census. The position "permanently or temporarily lives here" was replaced by two: for permanent residents in their absence, specify how long they are absent, and for temporary residents – a permanent address and time of absence. As a result, it was possible to collect information, including on population movements (migrations). Questions about literacy and graduation were included in the education question, which reduced the number of questions without reducing the amount of data collected. Information about the place of work and occupation at this place of work changed places (in 1939, first they asked about the type of occupation, and then about the place of work). Such a decision did not clarify or add new data to the information collected, but linked the interviewees to their place of work. This trend fully corresponds to the established practice of hiring for as long as possible, with the dependence of social guarantees on the place of work and profession. For those who did not have occupations that were a source of income (in addition to employment, it could be their own farm, work in artels or fishing), it was suggested to indicate another source of livelihood.

Among other preparatory measures of the census was the clarification of the boundaries of all settlements, the ordering of the numbering of blocks (the bypass was carried out quarterly) and households, followed by the compilation of lists of blocks of urban settlements and rural settlements with specified names, as well as the compilation of the most reliable maps of districts. Urban settlements were often divided into blocks specifically for the purpose of conducting a census. Such zoning was designed to cover the population as accurately as possible by the census, as well as to streamline and optimize the work of census takers. As such, persons were selected on the recommendations of party, trade union and Komsomol organizations with exemption from their main work for the period of the census. In addition to maintaining salaries at the place of employment, additional remuneration was paid, the amount of which depended on the function performed during the census ^{[14, pp. 212-213]}.

The census data were recorded as of the night of January 14-15, 1959, regardless of the actual date of the survey of the census takers, which varied depending on the region and the degree of organization of the census workers. The data obtained were supposed to supplement the information collected by the Registry office, household and household books, and voter lists as of January 1 of each year.

From January 11 to January 14, the counters made a preliminary inspection of all residential premises on their sites. On January 15, the filling in of the census forms began, which was conducted by a survey method and lasted 8 days both in urban settlements and in rural areas. Then, for 10 days (from January 23 to February 1), a continuous control round was conducted. In its course, the absence of factors that could distort the collected primary data was checked: the correctness of accounting, the absence of double counting or omissions. These measures significantly reduced the likelihood of errors and increased the reliability of the collected data, at least in part.

Due to the carefully thought-out organization, the system of selecting employees for conducting the census and the checks of the collected data, the census information in terms of primary documents is largely reliable. The processing and publication of the collected data had a number of features that influenced the characteristics of the published aggregated data. The obtained primary results were supplemented with classified data and adjusted for official publication. Therefore, the published data have different degrees of reliability for different territories. In particular, data on a number of areas were not subject to publication in the open press ^[6]. Therefore, it is advisable to talk about two types of data – primary and published. Moreover, the published data were divided into materials for official use (not subject to publication in the open press), which were received only by state authorities, and public materials (already adjusted), which were included in the official multi-volume edition of the census results, published in 1962-1963 under the title "Results of the All-Union census of 1959". In addition to the consolidated volume under the same name, the mentioned edition included separate volumes devoted to the republics within the USSR. Some data were published in the format of preliminary results and individual indicators [See, e.g.: 12, 15].

The primary census materials were deposited in the funds of the Central Statistical Office, which are stored in the Russian State Archive of Economics (RGAE, formerly the Central Committee of the USSR), and the funds of territorial statistical departments, which were deposited in the relevant regional archives.

Due to the peculiarities of the census and the nature of the processing of published data, it is important to refer to its primary materials. Nevertheless, judging by the currently available publications based on the primary data of the 1959 census [See, e.g.: 3, 8], these materials have not been preserved in full. Often, the researcher is dealing with a sample (either natural or formed according to the principles established by archival institutions). The representativeness of the sample presented in the primary census materials should be checked both using statistical methods and by comparing it with available data from other sources, in particular, published information on the territory in question, presented in acts of civil status, household and household books, and voter lists.

The village of Arti as a census object

The specifics of filling in and data set of primary census materials are considered on the example of a case formed based on the results of the census in the Artinsky district of the Sverdlovsk region, which is stored in the State Archive of the Sverdlovsk Region (hereinafter – GASO). The choice of an array is determined by several factors. Firstly, the relatively small number of residents and their comparative homogeneity in terms of social and national origin, occupation and socio-cultural characteristics. The impact of migration processes is minimized. Secondly, only primary materials related to the district center, the urban–type settlement of Arti, have been preserved. This circumstance implies a number of features, due to which it is possible to assess the potential of both the source as a whole and a specific array within a given framework. More attention is paid to urban-type settlements, since it is on their example that the historical, demographic and social changes that took place during the census are most indicative.

The village of Arti can serve as an example of a settlement that, at the time of the census, was in the process of transformation from a rural-type settlement (factory settlement) to an urban one with a set of demographic characteristics that combined the features of both urban and rural settlements. In the census materials, the city-forming Artinsky Mechanical plant is referred to as a "hardware plant". Based on the available primary materials, it can be concluded that production has developed in several directions. In the structure of the plant during the described period, there were, apart from auxiliary ones (for example, plant management, transport, repair and construction shops), at least a sickle shop, a stagnant shop, a needle shop, a workshop of knives and scissors, a workshop of kerosene. Most of the population belonged to the social group "workers" and, one way or another, was associated with the mentioned hardware factory.

According to the criteria given in the decree of the Presidium of the Supreme Soviet of the RSFSR dated September 12, 1957. "On the procedure for classifying settlements into the category of cities, workers and resort towns", the population of a work settlement (urban-type settlement) should have been at least 3 thousand people, and workers, employees and their family members should have been at least 85% of the total number of residents ^[11]. Although the status of an urban-type settlement was obtained back in 1929 ^[10], the Arti meet these criteria. According to the 1959 census, the population of Arti was 13,016 people ^[1], of which workers, employees and their family members accounted for 93.5%. The last figure was calculated by the author from the processed primary census materials, since in the published data information on the social structure of the population is aggregated by administrative-territorial units (for example, in Sverdlovsk the region as a whole) [See: 9].

Practically, the village of Arti was a transitional type from a rural settlement to an urban one. On the one hand, it was characterized by a small population; extended multi-generational families still met in it (parents, their children, grandchildren and relatives lived under one roof). The situation with housing was similar: houses with one or two, possibly conditional "apartments", were widespread. On the other hand, a significant part of the residents worked at the factory and had little to do with farming, typical for rural settlements, as the main form of employment. According to the processed primary census materials, agricultural, forestry and agricultural workers accounted for 4.5% of the employed (including those who indicated personal farming as a source of income).

The Artinsky district, which acts as an object for collecting the considered primary data, was formed in December 1923, initially as part of the Kungursky District of the Ural region. After the administrative reform in 1934, the Artinsky district became part of the Sverdlovsk region, where it remained during the census period.

According to the 1959 census, the population of the district was 26,273 people ^[2], of which, as mentioned above, 13,016 (49.5%) lived in the district center and were classified as urban population. On this basis, the Artinsky district can be classified as rural, although with some reservations. The primary materials of the 1959 census presented in the archive fund do not affect the settlements belonging to the district, therefore it is not possible to extrapolate the patterns revealed during the development to them.

Features of filling in the primary materials of the 1959 census

When analyzing the information potential of the 1959 census, it is necessary to take into account the specifics of filling in primary census materials, since they have a direct impact on the development of the data contained in them. The work is simplified by linking to an administrative division, that is, the completed forms are ordered by localities and districts in which the interviewees lived at the time of data collection.

The primary materials are standard typographic forms specially designed and printed for the 1959 census. In accordance with the quarterly round–up, they are ordered within the framework of the relevant case - according to the numbering of blocks within the settlement, which was assigned by the census department for the convenience of organizing data collection. This was the reason, at first glance, for the implication of the principle by which the sheets are ordered, which does not include either streets in alphabetical order or alphabetically ordered surnames of homeowners or heads of family. In addition to the forms of the individual survey, summary statements were formed on the household or locality (if it was a very small settlement).

On each census form, the numbers of the census department, the instructor's section, the counting section, the block and the summary statement of the household (separate or apartment building) are indicated. The numbers of the sheets in the file are also indicated, presumably with a chemical pencil, when forming censuses in the case, however, they do not always coincide with the actual number of the sheet in the case in the form in which it was during the processing of information and the formation of a database based on them. In the certification sheet of the processed case, it is noted that the sheets are divided into four ranges with the last number 775.

In the upper right corner of the completed form there is a signature or a letter designation, apparently belonging to the scribe or a representative of the site. Such notes are available only on the first sheet for each household.

In total, the primary materials are placed on 711 sheets, including 2,538 rewritten with an average of 3-4 entries. The census forms were filled out by hand, however, almost all entries are legible, including notes and corrections. There is space on each sheet for six individual entries, grouped according to the household principle implemented for statistical purposes. For data processing purposes, the interpretation of the term "household" is taken from the official order of Rosstat ^[13], however, there are other interpretations [for more information, see: 16].

If more than six entries were required for a household, the subsequent ones were transferred to new forms that received the same number of the household summary statement. The sheets were numbered according to the apartment number in the apartment building (No. 1, No. 2), if there was no division into apartments, just the number 1 was put. If more than one apartment or house without division into apartments was needed (seven or more people lived, including temporarily absent), a letter index of the type "a" was added to the number", "b".

The entries were numbered starting with the head of the family; these numbers are used in references when describing the means of livelihood in addition to work (and its payment). For example, "dependent No. 1". In records of dependent relatives, in some cases (most often, in case of separation), the degree of kinship and occupation (profession) of the person who supports the interviewee are indicated.

Further, individual questionnaires were usually filled out in the order of kinship and property: head, spouse, children (from eldest to youngest), grandchildren (from eldest to youngest), parents of the head, parents of the spouse, relatives of the head, relatives of the spouse [See, for example.: GASO, F. r1813, op. 12, D. 5, L. 76—77ob; GASO, F. r1813, Op. 12, D. 5, L. 593-594]. There are also deviations from this order [See, for example: GASO, F. r1813, op. 12, D. 5, L. 665-666; GASO, F. r1813, op. 12, D. 5, L. 673-674], but additional research is needed to understand whether this is an exception or the peculiarity of filling out forms by a specific scribe. Singles also received a separate number (1), sometimes without specifying "single". The household address is clearly divided into urban and rural with its own form for each of them. The peculiarity of filling in the address can be considered the fact that rural settlements directly subordinate to the city councils or village councils of nearby urban settlements could have an address filled in according to the urban type.

Since the census data were processed using a specially designed computer "Census" for this purpose, most of the features of the form, in order to adapt to a machine-readable form, had numerical designations that encoded a variant of one or another feature. The magnitude and type of designation are related to the type and characteristics of the encoded feature. For example, for the sign "Is he currently married?", "0" and "1" were used for the answers "no" and "yes", respectively. The sign of widowhood for unmarried people was not additionally encoded, it is still the same "0". Since a note about widowhood was not a requirement when filling out forms, it was left to the discretion of the counter.

A difficult option is provided for questions about employment and the source of livelihood. Since there are many employment options identified, this feature was encoded with a three-digit number. The place of work had two indexes: one to encode the answer to the question whether he works in an organized collective (counting collective farms and artels), or on his own farm, or does not work at all; the second index encoded the place of work for those who work in an organized collective, for other options, dashes were put in the census materials instead. The source of livelihood was indicated only for those who did not have an occupation that was a source of income, and those working on the site of this feature had a dash, which was encoded as "0".

The exceptions were age (originally in the form of a number), citizenship, native language, nationality and attitude to the head of the family. Citizenship initially had no place for a numerical designation (this is how the form was drawn up). Nationality and native language are an ambiguous graph. Russian Russian has a place for encoding in the form, but in the processed array the most common label is XX, which corresponds to Russian and Russian language. If we talk about other nationalities and native languages, they had their own encodings (for example, Tatar – 37 [See, for example: GASO, F. r1813, op. 12, D. 5, L. 119ob; GASO, F. r1813, op. 12, D. 5, L. 164], Mari – 26 [See, for example: GASO, F. r1813, op. 12, D. 5, L. 196; GASO, F. r1813, op. 12, D. 5, L. 690], Jewish – 44 [See: GASO, F. R1813, op. 12, 5, l. 393], Belorussky – 02 [See: GASO, F. R1813, op. 12, d. 5, l. 139–139ob].

The column "relation to the head of the family" also did not initially provide a numerical designation, however, in each census sheet, the number of its members is indicated in the column with a mark indicating the head of the family. The rest of its members are marked X.

Numerical designations in the census form can be used when encoding information from primary materials, if subsequent aggregation and processing using mathematical methods are assumed. However, in their current form they are insufficient and need to be improved in order to maximize the coverage of the described features.

Features of the database on the primary materials of the 1959 census as an electronic source

The database based on the studied complex of primary materials was formed at the International Center for Demographic Research (ICDI) UrFU is initially focused on historical and demographic research. While the database does not have an approved name, the temporary one is "The All–Union Census of 1959: Artinsky district of the Sverdlovsk region of the RSFSR". According to the mode of use in terms of individual records, the database is of a reference nature, since it is focused on transmitting the characteristics of the source. The volume of the database coincides with the number of individual records identified as a result of processing an array of primary materials (2,538).

The available sample, according to the table of sufficiently large numbers used in demographic and sociological studies, meets the sampling criteria with a marginal error of ? 2% with an accuracy of 95% for any trait, the probability of which is considered to be 50% (for example, gender).

The situation is complicated by the unclear nature of the selection of primary census materials. Close to the truth is the assumption that the researcher is dealing with a natural sample. A comparison of published data ^[2] and database data on population size and gender shows a very high similarity in the distribution of this trait with a deviation of 0.52% (Table 1). This confirms the hypothesis that the sample is representative of features with a 50% probability of occurrence.

Table 1.

Artinsky district: the representativeness of the natural sample by gender

Source: compiled by the author based on the materials of the database and ^[2]

The available published data for the Sverdlovsk region ^[9] can also be used to verify the similarity of the distribution of features, but the accuracy of the assumptions in this case will be somewhat less, since we are talking about aggregated data for the region as a whole. There may be differences at the level of individual settlements, depending on their specifics. Nevertheless, this comparison makes sense.

For example, the social structure of the urban population (distribution by social groups) in the given sample is similar to the official census data (Table 2). Fewer employees and more collective farmers and those working on their farms are explained by the specifics of the locality – relatively few institutions, the area is more rural, some features of a factory settlement with a subsidiary are preserved the economy of a part of the workers. 

Table 2.

Artinsky district: the representativeness of the natural sample by social groups (in% of the total)

Source: compiled by the author based on the materials of the database and ^{[9, p. 37]}

Taking into account the above data, it can be argued that the sample of primary data that has been formed, and therefore the database compiled on its basis, has the accuracy necessary for conducting research.

The orientation of the database to historical and demographic research was the main reason for the formation of a certain database structure, which, in addition to accurately transmitting the source records related to personalities, contains a separate table with information about the family.

The first table, entitled "Personal data", contains individual data on the re-registered citizens. Most of the field names repeat the wording of the questionnaire, although these names have been simplified as much as possible for ease of filling in and using the database. In the form, the names of many fields are supplemented with brief instructions on how to fill in, which is unnecessary for an electronic database. Significant changes compared to the source concerned information about permanent residence and absence at the place of the census – they are divided into columns "Place of permanent residence" and "Time of absence at the place of the census". The first column marks the place of permanent residence of the interviewee. If it coincides with the place of the census, then the name of the settlement (Arti village) was put, in other cases the address from the census sheet was reproduced. The time of absence at this address was entered in the "Notes" column, and the absence at the address where the census took place was indicated in the "Time of absence at the census place" column.

The added fields include both technical information ("Individual record number" and "Family Code") and substantive information ("Housing conditions", "Archive cipher", "Notes"). An individual record number is assigned to it when it is entered into the database. The "Family code" field contains a conditional identical designation for all family members. The family code is the main identifier for the second table called "Family".

The "Housing conditions" field is filled in according to indirect data, which distinguishes it from all others. Nevertheless, data of this kind are important, including from the point of view of comparison with materials from later censuses. Taking into account the peculiarities of the settlement, in most cases the housing was a house or part of it (a conditional "apartment").

The characteristics of housing conditions were given on the basis of several factors: the address (with or without the designation of the apartment number), the number of families in the census unit (conditional household), the source of livelihood, the presence of both families and singles in the census unit; additionally, it was taken into account whether the correspondents were temporary residents or not. If the apartment number was indicated with the accommodation of members of the same family, then the housing was characterized as a separate apartment, when two or more families (including one person) lived, a conditional communal apartment, in other cases – a separate (for one family) or a joint (for several families or singles) house. The "temporary housing" type was used for temporary residents with a permanent address in the appropriate field, except for those living in the same area where the census was conducted. The type of "rented housing" is rare, as a rule, in relation to singles living in the same household as families, or in relation to a family or a single person at a young age in a household where an elderly citizen with the status of a "loner" is the first to be re-registered. Housing could also be rented, where one of the residents, as a rule, the first to be rewritten, indicated as a source of livelihood income from renting premises in one form or another [See: GASO, F. r1813, op. 12, d. 5, l. 154ob; GASO, F. r1813, op. 12, D. 5, L. 296].

The "Archive cipher" field contains an abbreviated indication of the source of the data, according to which information is filled in on the transcribed, in the format of links to archival documents: the name of the archive, the number of the inventory, the case and the sheet. The "Notes" field contains additional information on the respondents and has the following structure: age data for children under one year old; extended information about education – the number of classes completed (or from which class he left); time of absence at his permanent place of residence (in months); additional marks by belonging to a social group and nationality; marks about the separation of family members; features of filling out the form.

The second table "Family" contains the characteristics of families and singles, information about which is given in the first table. As mentioned earlier, the main identifier is the designation in the "Family code" field. The following six fields are devoted to the characteristics of the head of the family: surname, gender, age, marital status at the time of the census, nationality and level of education. These data are imported from the "Personal data" table for respondents with the degree of kinship "head" and "single", as well as an unknown degree of kinship without obvious relatives.

The characteristics of the family as a whole include aggregated numerical characteristics (the number of family members, children under 18, employees, dependents, consumers) and qualitative parameters that describe it from a socio-demographic point of view. Earlier, the feature of filling in the column "Attitude to the head of the family" was mentioned – the number of family members is indicated in the census form next to the designation of the head of the family. However, it needs to be verified by counting, since errors in filling out and incomplete accounting of family members are possible. The number of children under the age of 18 was counted when filling out. The number of employees was calculated when filling out, based on the source of livelihood, designated as "work" and "household" (for heads of families and singles). As dependents, those rewritten with the mark "dependent <degree of kinship>", "alimony" and "household" were taken into account (the latter provided that they were not considered heads of families or singles). These categories did not include pensioners, people receiving scholarships, as well as those living on savings and income from renting out premises. The number of eaters was calculated relative to the average volume for an adult male who is not engaged in physical labor with coefficients by age, gender and occupation (all working professions and employment in agriculture and forestry are attributed to physical labor).

The qualitative parameters conditionally include such parameters as the presence of members absent at the time of the census, the demographic type of the family, the type of family according to P. Laslett^[17], the structure of the family, the social group (for the family as a whole). The presence of members missing at the time of the census is a logical type field that assumes only the "yes" and "no" marks. The demographic type of the family was determined based on whether the head of the family was married. The type of family according to Laslett was determined based on such parameters as the state of marriage, cohabitation or separation of relatives, the number of generations in the family. Entries about separated families are necessarily accompanied by comments about the identified family structure in the notes. The family structure is described based on the members living together using standardized description options that identify singles, nuclear and extended family types with subtypes. The social group of the family as a whole is determined based on the membership of the social group of the simple majority of its members, if there are only two members in the family, by the affiliation of the head of the family with a mandatory reservation in the notes of other social groups.

As additional fields, the table includes fields characterizing the place of residence (city or district, added to allow combining data from several databases compiled using primary materials from the 1959 census), housing conditions (imported from the "Personal Data" table), notes, the "Archive cipher" field, which is filled in the same order the same as in the "Personal data" table. The "Notes" field contains data with the following structure: a note on the structure for a divided family; a note on the interethnic marriage of a family member and the nationality of the spouse; features of classifying family members to social groups; time of absence from permanent residence (in months); differences in the surname or nationality of family members; remarks on peculiarities of filling in or doubts about the reliability of records on the degree of kinship.

It is also necessary to take into account the specifics of filling in primary materials and their reflection in the database. In particular, the spelling of surnames, first names and patronymics was given as in the source. Patronymics were not always spelled out in full, when filling out the database, the omitted endings of patronymics were reconstructed wherever it was possible to do this in an unambiguous way, however, this was not possible in all cases. In addition, the spelling given in the source was preserved, in particular, the spelling of the letter "e" in the names, surnames and patronymics where it stood in the source, and the peculiarities of spelling, for example, "Trapezdnikov". Not all counters prescribed the letter "e" when filling out forms, this can be attributed to the individual characteristics of filling.

Data reflecting the place of work and employment at this place were slightly adjusted, several spelling variants were reduced to one for convenience of accounting. For the same purpose, the writing of the time of absence is uniformly designed. In general, the database, despite its initial orientation towards historical and demographic research, adequately reflects the information contained in the source and can be used in as wide a range of historical research as the primary census materials.

Conclusions

The presented database, formed on the basis of processed primary materials, can be used in research on social history, historical demography and migration, especially family history; partly in the study of socio–economic history and the history of factories and plants, as well as in other types of historical, demographic, economic, sociological and interdisciplinary research. Examples include the study of migration processes and migration history, marital behavior, family history, employment structure, and demographic transition processes using examples of settlements and localities of various types.

The main problem that has to be solved when referring to the source is the representativeness of the sample of preserved primary census materials. A comparison of the distribution of some features based on the materials of published data and information obtained using the database gives reason to believe that the presented sample reflects the information of the census materials for the considered locality with sufficient accuracy.

If it is necessary to verify or expand the information involved, it may be advisable to refer to the materials of the registry office, passport offices to clarify information about registration and movements of citizens, as well as other materials containing population data for 1959. Nevertheless, the database created on the basis of primary census materials has an undoubted potential for use in scientific research.