/ UiT The Arctic University of Norway
 

Description of the linked censuses 1960, 1970 and 1980

1. DESIGN AND PROCEDURE

2. TERMS AND VARIABLES

3. EVALUATION SURVEYS



1. DESIGN AND PROCEDURE

1.1. Introduction

The Population Census in Norway in 1980 was the nineteenth in the series of nationwide censuses in this country and the ninth in this century. There has always been a wish to compare these censuses, but discrepancies in census design and procedure have complicated the fulfilment of this wish.

This publication presents main figures for the censuses in 1960, 1970 and 1980. It is based on reprocessed material deduced from the basic information in the three censuses. The Division for Population and Housing Censuses has recoded information in order to compare the censuses.

The reprocessed material is composed to provide information on the same person through the censuses. Such a composition gives us the possibility to present figures for the situation at three different points of time, but al so for transitions on the individual level between the three points of time in this publication.

1.2. Coverage

The total material covers persons who are included in at least one of the three censuses 1960, 1970 and 1980. All the censuses comprise all persons, also foreign citizens resident in Norway on the census date, 1 November. In 1970 and 1980 all persons were considered as resident if they were registered according to directions for keeping and arrangement of the National Register. In 1960 the National Register did not exist (the census was employed to create it), but the definition of 'resident' corresponds to the directions that were established later.

The reprocessed information from the housing censuses includes all private dwellings (flats etc.) where at least one person was registered as resident. Empty dwellings are not included in the censuses, neither are dwellings for persons who were registered at another address. We have not included information about the dwelling for persons registered in joint households (old peoples' homes, nursing homes, children's homes, and other institutions).

1.3. Statistical basis

The information in the censuses is derived partly from information collected by the Central Bureau of Statistics directly from the informants themselves on separate forms, partly from various registers.

In 1960 no register was used. Information is taken from questionnaires. National Registration Number has been entered at a later stage.

In 1970 name, address and marital status were collected from the Central Population Register (CPR) and filled in in advance. If the advance filling-in was missing or wrong, the informant was asked to fill in or correct the information. We also developed lists in advance that classified the industry of the establishments in each municipality. All other information came directly from the informant.

In the Population and Housing Census 1980 most information is based on data entered directly into the questionnaires by the informants.

All information on persons under 16 years is taken from the CPR. Data on age, sex and other demographic variables also come from this register. Information on education is taken from data given by the informants themselves to the 1970 census and from the Register of started and completed education in the US. Education abroad was included in the questionnaire as a separate entry.

The necessary information about conscripts to military and compulsory social service was taken from the Defence Headquarters and the Administration of Compulsory Social Service. Income information was supplied by the Taxation Register, the National Insurance, the National Housing Bank and the Government Loan Fund for Education.

The Central Register for Establishments and Enterprises was useful in the coding of industry.

All the censuses have been carried out in pursuance with the Act of 25 April 1907 relating to official statistics, and according to recurrent decisions by Stortinget (Royal Decree 1 April 1960, Decision by Stortinget 14 November 1969, and Decision by Stortinget 16 November 1978).

1.4. Data collection

In 1960 one questionnaire was filled in with person and dwelling information for each dwelling (flat list) and one for each house (house list). The forms were handed out and collected. The informants themselves grouped persons together to households.

In 1970, too, the forms were handed out and collected. The "main person in the household" was asked to fill in a form for each person in the household, a form with information on the dwelling and a form on the house in which the dwelling was situated.

Dwelling and household are defined as one concept, which enabled us to group the person information returned with the dwelling information together as one household.

In 1980 all persons 16 years and over in that year received a person form. The questionnaires and instructions were distributed by mail. Name and address were filled in in advance. The oldest person in the household received å dwelling form and a stamped return envelope. Only one dwelling form was to be filled in for each dwelling and all person forms should be mailed together with the dwelling form. This method of collection enabled us to group into households.

2. TERMS AND VARIABLES

2.1. Comparison

The population and housing censuses in 1960, 1970 and 1980 vary in scope, design and processing procedure. Therefore the Census Division has recoded comprehensively to obtain data sets with comparable figures. In this work we have employed classifications and definitions from the 1980 census as bases, with some exceptions.

As the material has been reprocessed, one can normally not compare directly figures in this publication with information in the publications for each census.

All information is comparable for at least two censuses, except socio-economic status. When we still include it, it is because we consider it an important background and/or consequence variable that is useful disregarded comparison.

This comparable material from three censuses is the result of complicated recoding. A forthcoming publication will present documentation for all variables.

The basic material for the three censuses is documented in detail in separate publications (all in Norwegian):

1960: Børke, Sindre: Folketellingen 1960. Ny organisering av datafilen. Interne notater 83/18. Statistisk Sentralbyrå, Oslo/Kongsvinger 1983

1970: Statistisk Sentralbyrå: Folke- og boligtelling 1970. Dokumentasjon. Interne notater 83/12. Statistisk Sentralbyrå, Oslo/Kongsvinger 1983

1980: Børke, Sindre: Folke- og boligtelling 1980. Dokumentasjon. Rapporter 84/15. Statistisk Sentralbyrå, Oslo/Kongsvinger 1984

2.2. The units

Person

All persons, included those with foreign citizenship and those without fixed address, registered as resident in the country as per the census date are included in the census. See above (section 1.2) for definition of registered as resident in 1960.

Family

As a family is defined:

1. Married couple with or without unmarried children living in the same dwelling.

2. Father/mother with unmarried children registered as living in the same dwelling.

3. Each single person not covered by any of the above definitions. Such persons constitute a one-person family separately.

Families consisting of married couples can also include unmarried children of one of the spouses, unmarried adopted children and stepchildren, but not foster children. According to these rules persons (e.g. spouses) who are not registered in the same dwelling are not grouped to the same family. Further, no married, separated or previously married person is grouped to the same family as their parents. When for example a married couple is living together with their divorced daughter, the couple makes up one family and the daughter another. Unmarried siblings living together in a dwelling where none of their parents live, are classified as one family each.

Household

A household is defined as a dwelling household, i.e. all persons registered as dwelling in the same household constitute one household. The household may contain one or more families.

We distinguish between private households and joint households.

Private household comprises persons registered in private dwellings.

As Joint household are classified persons given board in old peoples' homes, children's homes and similar, and also people registered as dwelling in hotels and the like.

Employees resident at their institution are always classified to private households. The same applies to military personnel registered as resident in a camp.

Dwelling

Since the census applies the concept dwelling household, the term of dwelling is to be interpreted as private household. The dwelling variables refer to the place of residence at which the persons are registered as per 1 November in the census year.

There is no information on dwelling for persons living in joint households.

Dwelling is mainly the same as flat, hence we include in dwelling also flat in a semi-detached house, flat in horizontally divided house, flat in a block, bed-sits with separate access, flatlet, and also detached one-family house.

Bed-sits without separate access are normally not classified as dwelling, but in blocks of bed-sits (e.g. residential halls) each bed-sit is counted as dwelling if someone is registered as resident there.

Physically a dwelling is room(s) being built or rebuilt for one person or several persons with joint lodging, and having access to the room(s) directly, not through another dwelling.

Bed-sits without own access, not being parts of blocks, institutions etc. are therefore always counted to the same dwelling as the flat of which they are a part, even if a separate family is registered as resident in the bed-sitter.

House is not a unit of its own in the material, but the dwelling variables include some information on the house in which the dwelling is located.

2.3. The variables

The list below does not give details about all variables in this publication. Further details are given in the documentation mentioned above, and in the text chapters of other census publications. The variables are grouped by standards and definitions valid for the 1980 census, if otherwise is not stated.

2.3.1. Population-variables

Place of-residence

We presume every person to have his or her place of residence where registered by the CPR as per the census date.

The main rule is that a person is to be registered where he or she takes the diurnal rest.

Persons without regular address are normally registered as resident in the municipality in which they had their latest fixed address.

This material does not include information on temporary place of residence.

Municipality/enumeration district

The division of municipalities is as per 1 November 1980. If borders have been altered during the period, the persons have been grouped geographically according to the 1980 division.

The division of enumeration districts has not been updated to 1980 standard. This material therefore contains the original districts as they were as per census dates.

Area-of residence/urban settlement

All dwellings are located in either densely or sparsely populated area. An area is densely populated if it is inhabited by at ]east 200 persons and the distance between the houses normally is 50 metres or less.

Densely populated areas are grouped in urban settlements. The physical demarcation is in some cases estimated and the limit of 200 persons is not always strictly followed.

The 1960 census did not group together into settlements, but the Census Division has gathered densely populated areas into urban settlements later. The details about the newly grouped settlements will be published in the report documenting the recoding.

We have used the concept of settlement as a functional unit more than a geographical one. Hence settlements will vary in their size over time. The rules for the classification of densely and sparsely populated have been kept constant in all censuses, but the area being covered by the settlement has been corrected from one census to another.

Change-of residence

Data on the same person in various censuses give us the opportunity to cross-tabulate this person's place of residence at 10-year intervals. This is being called change of residence in the tables.

Change of residence is being measured at municipality level. This must not be confused with migration statistics, because all migration during the 10-year period between censuses is not covered.

2.3.2. Occupation-variables

Economic-activity

When we compare three censuses, we must start with the 1960 definition of economic activity and consider as economically active in 1960 those who had income from own work as their main source of livelihood. In 1970 and 1980 this corresponds to persons who had worked 1 000 hours and over during the year. All three censuses include only persons 16 years and over. There is a fairly reasonable correspondence between these definitions.

The material is organized in such a way that other definitions of economic activity can be applied. If one does not need to include 1960 figures in one's comparisons, the concept can be expanded to include persons 16 years and over with 100 hours and over a year. In cases of 1980 data being grouped by socio-economic status, the limit for economic activity is fixed to 500 hours a year, as the standard requires.

Correct use of words means that economically active is used only when we refer to income from own work as main source of livelihood (1960), or to 1 000 hours and over a year (1970 and 1980). In other cases we write for example economically active with 100 working hours and over.

Industry

The industries are grouped according to the standard of 1980. In a few cases we have used a classification by primary, secondary, and tertiary industries. The connection between this classification and the standard is such:

Primary industries Agriculture, hunting, forestry, fishing, sealing and whaling
Secondary industries Oil extraction, mining and quarrying
Manufacturing
Electricity, gas and water supply
Construction
Tertiary industries Wholesale and retail trade, restaurants and hotels
Transport, storage and communications
Financing, insurance, real estate and business services
Community, social and personal services


Commuting

In this publication a commuter is a person with registered place of residence and place of gathering for work in different municipalities. In the 1980 and 1970 censuses all persons who gathered regularly during the week before the census date/at the census date were given a code for gathering municipality. in 1960 this muster code was given to all economically active in selected municipalities.

Changes in municipality boundaries make it impossible to compare figures for commuting. We can only supply information on a person's place of residence and whether he or she is commuting or not. The variable is yes/no information on commuting.

Socio-economic-grouping

The grouping of socio-economic status classifies persons to various levels. The grouping is an index that takes various personal variables into account. The index splits the material into distinctly defined groups, but it does not label one group as having a higher status than another.

Each person can appear in only one group. The same person can fill the requirements for several groups (both economically active and working at home), but in those cases one group is to be given priority according to certain rules.

The standard distinguishes primarily between persons with economic engagement and those with another main activity. In this grouping we consider a person as economically active if he or she has worked 500 hours or more during the year.

The economically active are further divided according to occupational status (self-employed/ employed), occupation, competence (education), and authority (supervisory status or not).

Those who are not engaged in economic life are divided into groups of activity (working at home, studying, doing military service), or according to their source of livelihood (receiving pension).

The standard is presented in detail in

Statistisk Sentralbyrå: Standard for inndeling etter sosioøkonomisk status Standarder for norsk statistikk 5 Statistisk Sentralbyrå, Oslo/Kongsvinger 1984

The variable socio-economic status is included only for the material from the 1980 census.

2.3.3.- Education variables

Highest education

The classification of the highest education of a person is not comparable between the 1960 census on the one side, and the 1970 and 1980 censuses on the other, except on a very coarse level. We have therefore distinguished only between education on compulsory level, intermediate level and higher level in comparisons between the three censuses.

This comparison pay more attention to type of education than length, because 9-year compulsory basic school is introduced in the period. The term intermediate level covers education beyond compulsory level, but not education leading to a university degree or similar. The term comprises both vocational training and general education.

Continuation school and the tenth year of basic school are considered as intermediate level education, except for 1970, when the tenth year was coded together with 9-year basic school.

High level is all types of education leading to a university degree on cand. mag. level or higher. Accordingly, education on university or college level I in 1970 or 1980 is not high level in these tables.

Both in 1970 and in 1980 the standard classification of education was the basis for coding, which renders more detailed comparisons between the two censuses possible. However, one must keep in mind that some educational codes do not exist in the standard (vocational training at basic school level). The forthcoming documentation of the 60-70-80 material will explain the relation between the code values used in the censuses and the standard.

Pupil/student status

All persons of compulsory school age are considered as pupils, others are pupils/students if they attended school as per the census date, disregarded other activities, as for instance paid work.

2.3.4.- Family and-household-variables

Number-of economically-active in-the-family/household

The number of economically active is the total of persons filling the requirements for economic activity as mentioned in section 2.3.2 above.

2.3.5.- Dwelling-variables

Rooms

Data on number of rooms concern dwellings where someone is registered as resident, and comprise rooms of 6m2 and over fit for human occupation all year round. Rooms for other purposes (kitchen, bathrooms, laundry rooms, and rooms for economic activity (offices» are not included.

The publication does not specify the number of rooms for dwellings with 8 rooms or more, as the 1980 census did not collect this information. In our calculation of average number of rooms in the dwelling for the 1980 census, average figures from the 1970 census on dwellings with more than 8 rooms have been taken as starting point.

Type-of dwelling

It proved difficult to achieve a fair comparison for the variable type of dwelling, and we could not stick to the classification of 1980. Instead we have produced a less detailed standard that apply to none of the censuses, but is an extract of all three.

The new classification has been established in co-operation with institutions that use dwelling statistics. These institutions already have good experience in using the standard.

The relation between the new classification and the original ones from the three censuses is:

New classification Corresponding type of dwelling in previous censuses
Dwelling house on farm

1960

Dwelling building in connection with farming

1970

Dwelling building in connection with farming

1980

Dwelling building in connection with farming
Detached one-family house

1960

Detached one-family house

1970

Detached one-family house
Other one-family house
Other dwelling-building with one flat

1980

One-family house with one flat
Other dwelling-building with one flat
Small houses with more than one flat

1960

Building in row, chain, terrace house or vertically divided two-family house
Other small house with less than 5 flats

1970

Chained one-family house etc.
Horizontally divided two-family house
Other dwelling building with 2 or more flats and less than 3 floors

1980

Building in chain etc.
Horizontally divided two-family house
One-family house with 2 or more flats
Other dwelling building with 2 or more flats
Block

1960

House with 5 flats or more

1970

Other dwelling building with more than 2 flats

1980

Block etc.
Business building

1960

Business building with less than half the area used for dwelling

1970

Business building with less than half the area used for dwelling

1980

Business building with less than half the area used for dwelling
Dwelling in house for joint household

1960

Dwelling in house for joint household
joint household

1970

Dwelling in house for joint household

1980

Dwelling in house for joint household
Dwelling not comparable 1960/1970 - 1980

1960

Provisionally built house or house being temporarily occupied as dwelling

1970

Dwelling to be used 10 years or less



3. EVALUATION SURVEYS

3.1. General

The questionnaires to population censuses and similar surveys will always be encumbered with random and systematical errors. In addition, many informants will fail to answer one or more questions. For the censuses in 1970 and 1980 we therefore performed separate evaluation surveys (ES) to map the scope of errors. No such survey was carried out for the 1960 census, but we have reason to believe that the conclusions for the ES70 is applicable for the 1960 results. The censuses in 1960 and 1970 were quite similar in formulation of the questions, forms' construction, and collection and processing of data.

The evaluation surveys are interview surveys of a representative sample of the population, which enables us to compare two sets of information on a part of the population.

Data from the ES is collected and controlled in a more elaborate way than the census itself. An ES is therefore generally recognized as being of a better quality, coming closer to "the truth" than information in a census.

Errors in the census material vary according to what kind of table we study. In a table showing the total population at two or more points of time, the net error in the material will decide whether it makes sense to compare figures from different points. The net error, or net deviation, in data material is the difference between the total figure of the material and the "true" material.

An apparent change from one census to the next does not necessarily originate from a real change. Even if the world is unchanged, the censuses may display changes if they contain data of various quality. And vice versa, changes in the real world may appear as stable conditions if the censuses have varying net deviation.

Persons totalling up to a net result may have given wrong answers separately, but their answers may counterbalance each other so that the total figure comes close to "the truth" and shows little net deviation. Such a table is of great significance even if it contains individual deviations.

In a table comparing the same individual at different points of time also the gross error in the material will influence the quality. The reliability of a change/non-change depends on how probable it is that a person has answered correct once, both times or not at all. Such errors at the individual level are called the gross error of the material.

3.2 Net errors

In 1970 the ES discussed the variables education and economic activity. The survey proved that the census had produced too few persons with basic education only and correspondingly too many with longer education. Besides, too few had vocational education.

It is notable that these quality conclusions about education variables to a certain degree are valid for the 1980 census as well. In 1980 the information on education is based on a register in the US. This register was established on data from the 1970 census, which implies that for persons with the same education in 1980 as in 1970 this information originates in the 1970 questionnaire.

The ES in 1970 also shows that the census underestimated the number of economically active. The underestimation increased with decreasing working hours per year and was most significant for the group with less than 500 hours. The deviation was more marked for women than for men.

The ES in 1980 looked into dwelling and economic activities. In general the net errors in 1980 are less than in 1970.

We find a high degree of correspondence between the census and the-evaluation survey as to the total number of economically active persons, but there is significant deviation in the groups for working hours, in particular in the group economically active with 1 300 working hours and over during the year.

The dwelling variable Number of rooms in the dwelling ends up differently in the ES and in the census, and there is a certain deviation in the ownership/tenancy relation. Too few persons have stated that they own their dwelling.

Table III.1 below illustrates how net errors in the material from two sources influence a seemingly clear tendency. The population censuses in 1970 and 1980 show a marked increase in the number of economically active with 100 man-hours and over a year from 1970 to 1980, but according to the ES for the two censuses the reason for this increase is underestimation in the 1970 census. The two censuses reveal a difference of 5.9 percentage points, as the ES70 and ES80 appear with nearly identical figures.

This publication applies a limit of 1 000 man-hours and over a year to the definition of economic activity (see paragraph 2.3.2). For this group both ES and censuses show decrease from 1970 to 1980. The decrease is real (cfr. e.g. a lowering of pension age from 70 to 67 years in the period), but not as big as it might seem according to the censuses. Conclusions on development trends therefore must make more of a point of direction than of quantity.

Table III.1 Persons 16 years and over by economic activity and working-hours per year. Population census and quality control survey. Percentages and percentage points. The whole country 1970 and 1980

Economic activity and working hours PHC70 PHC80 ES70 E580 PHC80-PHC70 ES80-ES70
Total 100 100 100 100 - -
Not economically active 40,7 34,9 33,7 34,5 -5,8 -0,8
Economically active, total 59,3 65,1 66,3 65,5 5,9 (-0,8)
100 - 499 hours 4,2 9,6 7,7 8,2 5,4 -0,4
500 - 999 5,8 7,6 7,2 8,7 1,9 1,5
1 000 - 49,3 44 51,4 47,8 -5,3 -3,5
Not reported 0 3,9 0 0,8 3,9 0,8

Figures in brackets are not statistically significant.

Source: Evaluation Surveys for the Population Censuses in 1970 and 1980.

The table also shows that changes in the groups for working hours per year are either non-real (100 - 499 hours) or too big (1 000 hours and over).

3.3 Gross errors

When we compare data about the same individual from various material, we need information about the net deviation in both/all sources. In section 3.2 above we have seen that a major net deviation may effect a "change" which does not originate in realities, but in a technical transition from "false" to "true" value (poor and good sources), or in a transition from one "false" value to another (two poor sources).

Irrespective of the net deviation in two data sources, we must also consider the gross deviation when we compare at individual level. We must know bow likely a single informant is to give correct answers, and whether this likelihood varies between the sources. A major gross deviation in one or more sources will produce interpretation difficulties for tables showing transitions at the individual level. Difficulties will grow if the sources have different gross deviation, or if the gross deviation appears in connection with a major net deviation.

In table 111.2 below we have extracted from the evaluation surveys variables and values where we found the errors significant.

A forthcoming report (1986) comparing the population censuses in 1960, 1970 and 1980, will discuss the quality of all variables covered by the ES70 and/or ES80.

Table 111.2. Gross and net deviation between population census and evaluation control survey for selected variables [Not reproduced in the web version]



Figures in brackets are not statistically significant. i 100 per cent = All persons in the census. 2 The net deviation measured in percentage points. 100 per cent = All persons in the evaluation survey. Source: The Evaluation Surveys for the Population Censuses in 1970 and 1980.



3.4. Tables we did not produce

In this publication we have not excluded from total-figure tables any variable of the comparable material on the ground of bad quality, because we consider the ES-information about net errors so unambiguous that the users can include the necessary corrections.

Transition tables at the individual level have not been produced for the variables Economic activity of the year and Working hours per year. For these variables the combination of net and gross errors will give so misleading figures that a table will be of little or no significance.

The publication presents the figures as they appear in the population censuses in 1960, 1970 and 1980. The results from the ES are not used to correct the tables.

3.5 The sources

Both evaluation surveys have been published:

1970: Statistisk Sentralbyrå: Folke- og boligtelling 1970. Hefte VI Kontrollundersøkelse. NOS A 823. Statistisk Sentralbyrå, Oslo 1976

1980: Heldal, Johan: Kvalitetskontrollundersøkelsen for Folke- og boligtelling 1980. Rapporter 84/21. Statistisk Sentralbyrå, Oslo/Kongsvinger 1984

Parts of the ES 1980 have been included in the publications from the 1980 census.