What To Expect From the 1950 Census
PAS Report 11
Historic PAS Report Series
PAS published its first Information Report in 1949. To celebrate this history, each month we're presenting a new report from the archives.
We hope you enjoy this fascinating snapshot of a planning issue of yesteryear.
AMERICAN SOCIETY OF PLANNING OFFICIALS
1313 EAST 60TH STREET — CHICAGO 37 ILLINOIS
|Information Report No. 11||February 1950|
What To Expect From the 1950 Census
Data collecting for the decennial census will begin in April of this year. Publication of many of the final volumes of United States summary material is expected by the end of the calendar year of 1952, with preliminary local reports being made available as soon as possible in 1950 and more detailed data being issued throughout 1951.
Changes have been made in the 1950 Census as compared with that of 1940: changes in the data to be collected, in the areas in which they are to be collected, and in the methodology of the collection and presentation of the data. Important changes are stated in the following pages, together with some of the implications of these changes. Some of the plans for publication of data are still unresolved. however, and will depend on the amount of funds available for this purpose. The material for this report was obtained from personal discussion and correspondence with Philip M. Hauser, Acting Director of the Bureau of the Census, and from releases and publications of the Bureau. PLANNING ADVISORY SERVICE, however, assumes full responsibility for the implications drawn from the information.
NEW GEOGRAPHIC CONCEPTS
Population and housing data of 1940 was published for 140 metropolitan districts, a metropolitan district being defined as a central city with 50,000 inhabitants or more and the surrounding urban places and minor civil divisions that had a population of 150 or more persons per square mile. Because of the difficulty, among others, of comparing these data with that obtained from the Census of Manufactures and Distribution, the 1940 definition of metropolitan district will not be used after the 1950 Census, and will be used in the 1950 Census chiefly for historical interest and for comparison with the 1940 data.
Standard metropolitan areas have been established for the 1950 Census. These areas have been determined by interagency cooperation, the Bureau of the Budget and the Bureau of Employment Security among others, actively cooperating with the Bureau of the Census. Data which have been collected for these areas in the 1947 Censuses of Manufactures and Distribution will therefore be comparable to the 1950 Censuses of Population and Housing. Each standard metropolitan area is composed of a single county or a group of two or more contiguous counties which contain at least one city with 50,000 or more inhabitants. These areas may cross state lines. Other distinguishing characteristics of the standard metropolitan areas are:
- Where two cities of 50,000 or over are within 20 miles of each other, they will ordinarily be included in the same metropolitan area.
- Each county included in a standard metropolitan area must have either 10,000 non-agricultural workers or 10 percent of the non-agricultural workers in the area, or more than one-half of the county's population must have been included in 1940 in the "metropolitan district"; as defined above by the Bureau of the Census. In addition, non-agricultural workers must constitute at least two-thirds of the total employed labor force of the county.
- When more than one county comprise the standard metropolitan area, each county included in a standard metropolitan area must be economically and socially integrated with the central county of the area. A county has been regarded as integrated, (a) if 15 percent of the workers living in the county work in the central county of the area, or (b) if 25 percent of those working in the county live in the central county of the area, or (c) if telephone calls from the county to the central county of the area average 4 or more calls per subscriber per month. These criteria were selected largely because investigation indicated that they were the only direct measure of integration on which usable data might be obtained for a substantial proportion of the areas under consideration. A report defining the standard metropolitan areas indicates that where satisfactory data on even these criteria are lacking, considerable reliance was necessarily placed on informed local opinion.
- In New England, where areas have been defined on a township rather than on a county basis, the same principles as outlined above have been applied in so far as they are applicable. Since the county is not used, a population density criterion of 150 persons per square mile in urban-rural areas or minor civil divisions surrounding the central city of 50,000 or more population (or where strong integration is indicated, 100 persons per square mile) has been substituted for the provisions in paragraph (2). Thus in New England, the metropolitan area relies on the 1940 definition of metropolitan district.
Urban Fringe Areas
Data will be collected for the built-up urban fringe areas contiguous to each city of 50,000 or more inhabitants. This is a new concept. It has been introduced to indicate the degree of urbanization of people, not reflected by the political boundaries of the central city. The boundaries of this built up area will, in general, be delineated by including those areas, contiguous to the central city, which have a density of 500 dwelling units, or 2,000 inhabitants per square mile. This is usually the area that has a regular pattern of city blocks. Tabulations will include figures for the central city and for the urban fringe area. Of course, it will be possible to measure the "degree of influence"; of the central city by subtracting the figures reflecting its political jurisdiction from the figures for the entire urbanized area.
Implication of Metropolitan Area Definition and the Urban Fringe
Cities attempting metropolitan planning have the problem of defining the area in which planning is to take place. For statistical ease, planning officials have frequently relied on the 1940 Census definition of metropolitan district which included, in addition to the central city, contiguous areas or minor civil divisions containing 150 or more persons per square mile. Such areas will most likely differ sharply from the areas which will result from the 1950 Census: the county-city type metropolitan area, or the central city plus urban fringe.
It may be anticipated that, again for statistical ease, the area determined by many communities as that in which metropolitan planning should take place, will be changed to meet the 1950 definitions. Impetus thus may be given to joint city-county planning activities.
Degree of Urbanization
The degree of urbanization of the country will be further depicted by another new concept to be employed in the 1950 Census. Those unincorporated places with 2,500 inhabitants or more, which have been previously omitted in the tally of urbanization because they were not incorporated, will, in the 1950 Census, be considered as urban, and boundaries for these places will be established so that they may be easily identified. Thus, the 1950 classification will conform to the popular concept of urban rural characteristics, because it will include the people in the built-up areas, without regard to the corporate status of the area. The total effect will be to reduce the rural non-farm classification and also to emphasize the urban classification of 1950.
The impact of this change may be felt, for example, by various states or counties depending somewhat on the extent of their rural non-farm population under earlier classifications and somewhat on whether they contain a considerable number of unincorporated places of 2,500 persons or more or of cities of 50,000 persons or more with extensive suburban developments. This total rural-urban population for a particular state is important for certain policy determinations and program administration. For example, if the ratio of total rural population compared to total urban population is used as a basis for the distribution of federal or state funds, a state or county which has few large unincorporated villages and few cities for which an urban fringe will be established, is likely to have a larger percentage of the total rural population in 1950, than on the 1940 basis, while a state or county with many large unincorporated villages, or many cities for which the urban fringe will be established, may have a smaller percentage.
Census tracts have been permanently established in cities with a population of 250,000 persons or more, and in some smaller cities. These tracts ordinarily include from 3,000 to 6,000 people and are demarked primarily by the similarity of the population characteristics within them, such as race, economic status, and living conditions. Boundaries for census tracts were reviewed for the 1950 Census, and were revised in a few cases. In the cities where census tracts had been previously established and where the urban fringe area is to be enumerated, census tracts will be established for the outlying areas contiguous to these cities.
Census tracts are of considerable value in analysis of all types of municipal problems. Although the overall statistical picture for a city may conceivably remain the same from one census to the next, the internal pattern of activities within a city may have changed considerably. Census tracts help isolate the areas of change that would otherwise be obscured in the city totals. For example, the school population may be decreasing for the. community, as a whole, but may he very rapidly increasing in a particular area. Data obtained for census tracts have many valuable applications for throwing light on desirable locations for various types of municipal facilities and services, etc. Tract data may help in indicating the areas in which redevelopment should have top priority. Illustrative of the type of socioeconomic analysis of census tract data that may be made for a city is The Social Areas of Los Angeles by Eshref Shevky and Marilyn Williams, published for the Haynes Foundation by the University of California Press in 1949, in which social status, segregation and degree of urbanization of the population is postulated and compared with trends in the country as a whole.
Statistics for census tracts will be published in a special series of reports. Plans are now being formulated by the Bureau of the Census to establish a set of standard groupings of census tracts as a type of intermediate statistical area — larger than the tract and smaller than the entire community. These groupings of census tracts will have use in presentation of population, housing. income, business, and other data, which would be too largely influenced by specific conditions if viewed on the basis of census tracts alone. For example, the business census indicating the characteristics of several kinds of businesses would find that a specific establishment would influence to too great an extent, the figures for the whole census tract. Therefore, the groupings of census tracts will permit a more careful isolation of change within the city, but still eliminate the possibility of undue influence by a particular type of activity.
The use of census tract statistics in the case of certain types of data will be affected by the substitution of sample surveys for complete enumeration. For example, in 1950, data on income will be obtained from a 20 percent sample of persons of 14 years and over, rather than from every person of 14 years and over. Only one out of every fifth male of 14 years of age or over will be asked whether he served in the U.S. Armed Forces during World War I, World War II, or other war. A sample of 3 and 1/3 percent of all females of 14 years of age and over, and ever married, will be asked the number of children ever borne. These sample surveys particularly in the last example above, may be most applicable to the community as a whole, and not be meaningful on a tract basis. The Census schedule attached to this report indicates which questions shall be asked of all persons, and which for a sample of the population.
Block statistics for housing data were available in the 1940 Census for communities of 50,000 inhabitants or more; it seems unlikely that funds will be available to present, as had been done previously, these block statistics in graphic form. Cooperative arrangements have been made by local housing authorities and other authorities with the Bureau of Census for making available block tabulations on housing characteristics (see the section in this report on special services of the Bureau of the Census.)
The definition of a dwelling unit to be used in the 1950 Housing Census will assure consistent treatment of living quarters in all areas covered by the Census. A dwelling unit is defined: a group of rooms or a single room, occupied or intended for occupancy by a family or group of persons living together, or by a person living alone, is a dwelling unit if it has: (a) separate cooking equipment, or (b) two or more rooms and a separate entrance. An exception is: a one-room apartment in a regular apartment house, or one room which is the only living quarters in the structure, is a dwelling unit, even though it does not have separate cooking equipment. Previously, the definition of dwelling unit was not modified by provisions (a) and (b) above.
Mortgage Characteristics and Rent Figures
The mortgage characteristics of owner-occupied non-farm homes will not be included in the 1950 Housing Census. The figures in previous censuses were found to be frequently unreliable, since the housewife, who is usually the respondent, is frequently unacquainted with financial matters.
Therefore, it has been proposed that a sample survey, instead of a complete survey, be conducted after the regular census enumeration has been completed, and that the questions on whether each owner-occupied home is mortgaged and the characteristics of the mortgage, be asked at that time of a person who is best acquainted with this information. This survey of residential financing will cover non-farm residential properties owned by non-resident landlords and by corporations, as well as owner-occupied homes.
In 1940, information was obtained on contract or estimated rent for all non-farm units. However, in 1950, information relating to rent will be obtained only for the non-farm tenant-occupied units and for vacant units that are for rent. It is believed that the estimated rent for owner-occupied units would be quite unrealistic for 1950 because of various changes in the housing market and because of the various effects of rent control. In 1940, contract or estimated rent for all non-farm units had been obtained. Current market value will be enumerated for non-farm owner-occupied units and for vacant units that are available for sale.
State of Repair; Dilapidation
ln the 1940 Census, information was given on the state of repair of dwelling units. Structures were reported as "needing major repairs"; when parts of the structure such as floors, roof, plaster, walls, or foundation required repairs or replacements, the contiued neglect of which would impair the soundness of the structure, and create a hazard to its safety as a place of residence.
In the 1950 Census, the enumerator is required to report on whether or not each dwelling unit is dilapidated. Full instruction, particularly visual aids, will be given to the enumerators so that there may be more uniformity in judging the state of dilapidation of each dwelling unit and, therefore, increased usefulness of the information gathered.
The criteria used in judging dilapidation include evaluation of: (a) weather tightness, (b) makeshift construction, (c) extent of disrepair, and (d) hazard to the physical safety of occupants. Examples of critical deficiencies which would be serious enough alone to warrant classifying the unit as dilapidated are (a) seriously cracked or broken foundations, (b) holes, open cracks, loose or missing materials (bricks, plaster, shingles and clapboards) over a considerable area of the outside walls, roof, chimney or interior floors, walls, or ceilings, (c) substantial sagging of floors, walls, roof or chimney, extensive damage by fire, storm, flood, (e) makeshift chimney (a stove-pipe or other uninsulated material leading directly from the stove to the outside through a hole in the roof, wall or window). Dwelling units also will be classified as dilapidated when other deficiencies are present in sufficient number and extent to give evidence that the dwelling unit does not provide adequate shelter or protection against the elements or is physically unsafe. Examples of these deficiencies are: holes, open cracks, loose or missing material in foundation, walls, floors, or ceilings, but less extensive than described in Item b above, (b) inside stair treads or rises, balusters or railings that are loose, broken or missing, (c) deep wear on door sills, door frames, outside or inside steps or floors, (d) broken or missing window panes (often holes will be plugged with paper, cloth or tin), (e) rotted or loose window frames which are no longer windproof, (f) shaky or unsafe outside steps or railings.
Not all of the deficiencies which would make a dwelling unit dilapidated, however, can be observed by an enumerator. Certain deficiencies such as enclosed joists or beams which are termite-ridden or rotted could be revealed only by an engineering survey. Other defects, such as lack of light or ventilation or the presence of dampness of infestation, could ordinarily be observed only be a more thorough check of the dwelling that than is possible in the survey. However, it is believed, there are a number of easily observable deficiencies which enable determination of whether the unit is dilapidated. These deficiencies are not only significant in themselves, but also generally indicate that the dwelling unit has serious defects that the enumerator cannot readily observe. Census enumerators are cautioned that all dilapidated units are not necessarily old. Many one room shacks, for example, are seriously deficient when they are brand new.
The new definition of dilapidated dwelling units may result in a reduction of the total number of dwelling units that are considered to be most in need of replacement or repair in 1950, as contrasted with the number that was considered in need of major repair in the 1940 Census, although an intervening period of ten years very likely would have produced a number of additional units in need of repair or replacement.
Housing Information Added, Omitted
Information that will not be included in the 1950 Census but was included in 1940 is that pertaining to the exterior material of the dwelling structure and whether each structure contained any converted units or not.
Information will be secured for a 20 percent sample of all dwelling units for the following items: the year that the dwelling unit was built, whether a kitchen sink is present and is for the unit's exclusive use, what type of heating equipment is used, what fuel is used for heating and what is used for cooking, what type of refrigeration is used and whether the dwelling unit has a radio, television set, and electric lighting.
OTHER PERTINENT CHANGES IN THE 1950 CENSUS
Definition of Family
In 1950, a family will be defined as any group of two or more related persons living together; all related persons in a dwelling unit will be counted as members of the same family, regardless of the degree of relationship. A single family, therefore, may contain two or more married couples who are related to each other; however, two or more family groups that are not related to each other are reported as separate families, even though they may share the same dwelling unit. Under the revised terminology,* a family must comprise two or more related persons — therefore, a household contains no family when it consists of one person, or when, if there are several persons in the household, none of the persons in the household are related to each other. A household, however, will continue to have the same meaning as it did in 1940 — that is, the person or group of persons a occupying a dwelling unit. The members of a household may or may not be related to each other; one person living entirely alone in a dwelling unit constitutes a separate household.
This change in the definition of family will tend to minimize the number of families in comparison with those enumerated for 1940. There will be no more "one-person"; families — and two or more married couples, if related to each other, and "doubled-up"; will count as one family. In comparing the number of families with the total number of dwelling units of a community — one of the frequent methods of estimating additional needed housing accommodations in "housing market analysis"; — there will be a tendency to underestimate the number of persons in need of dwelling units, as compared with that from previous years. Unless it is carefully noted, and interpreted, the definition of family may mean that the housing needs of single persons and doubled up families may be statistically glossed over. Hans Blumenfeld's article "A Neglected Factor in Estimating Housing Demand"; in the August 1944 issue of the Journal of Land and Public Utility Economics indicates the potential demand of "incomplete"; families for housing accommodations.
The major difference in the questions for 1950 as compared to those of 1940 pertaining to consumer income, will be that the amount of income received in 1950 will be judged not only from that received from wages or salary, but from self-employment and from other sources, and that data will be obtained on a 20 percent sample. In 1940, questions were restricted to income from wages and salary and to whether or not the person had $50 or more income from other sources. A more realistic analysis of the consumer income of different families will therefore be possible. It should be anticipated that the statistics for both individual and family income will be higher in the 1950 Census than in the 1940 Census, reflecting not only increased employment and inflation, but also the more inclusive definition of income.
Index of Mobility
In the 1950 Census, persons will be asked where they lived in 1949 or one year before the census is taken. This compares to the 5-year gap which was tested in 1940 when people were asked where they lived in 1935. This period of gauging mobility has been shortened by the Bureau of the Census, because more accurate information can be obtained for the shorter period and because if April 1945, a war year, were chosen, it would reflect a time of unusual disturbance in the location of the population. Also, migration information in 1950 will be restricted to naming the county where persons lived one year previous with no attempt to determine the city or to classify this former place of residence as urban or rural. In 1940, there was an attempt to classify as urban or rural, the previous address of the person interviewed, thus obtaining trends in population location. However, it was found that these resultant data were inaccurate and expensive to code: person tended to report a city as their place of residence, even though they lived outside the city limits. A question, however, on whether the former residence was on a farm, will still be used in the 1950 schedule.
Thus, information on movements of population will be more accurate than was obtained in the previous census, but will be less detailed. No movements from city to suburb, or vice-versa, if located in the same county, will be differentiated, according to the 1950 Census schedule; these movements are, of course, extremely important to the planner, Also, no internal or intra-city movements will be distinguished by this type of question.
In pre-tests for the 1950 Census, self-enumeration was attempted to see if there would be any savings in cost without any loss in accuracy. It was found that there was no over-all significant difference between the schedules filled out by enumerators and by the ones that were completed by the persons themselves in completeness of coverage of the population or in accuracy of the information obtained. Although some time of census employees was saved by the respondents filling in the schedules themselves, there was not a sufficient savings to offset the additional costs of printing, distributing and processing the more numerous and more complex questionnaires that were required with self-enumeration. Therefore, it was decided not to use self-enumeration in the Census of Population and Housing of 1950. However, self-enumeration will be used in the Census of Agriculture, since enumeration by census employees in rural areas involves considerable travel time and expense. Also, the data obtained by self-enumeration are more accurate, since the farmers and others have time to examine their records and fill out the information in more detail than they would in conversation.
SPECIAL SERVICES OBTAINABLE FROM THE BUREAU OF THE CENSUS
Tabulation of Data for Submission of Applications Under Title I of Housing Act of 1949
The Housing and Home Finance Agency has made arrangements with the Bureau of the Census so that cities or local public agencies which could meet the time schedule and supply the funds necessary for the tabulation of data on the dilapidated condition of dwelling units, could submit these data as evidence of blight and slum conditions when applying for funds under Title I of the Housing Act of 1949. Timing was an important element in the making of these arrangements: cities wishing an early processing of their applications for funds under the Housing Act must meet deadlines set by the HHFA, and if these cities wish census tabulations as supporting evidence for their applications, they need to secure the data in time to meet the deadlines.
The following cities are among those which have requested advance block statistics: Cleveland, Ohio; Knoxville. Tennessee; New Haven, Connecticut; Easton, Pennsylvania; Buffalo, New York; San Bernardino California; Pensacola, Florida, New York, New York; Providence, Rhode Island; Ventura, California; St. Louis, Missouri; Little Rock, Arkansas. The estimated cost of these block tabulations will be approximately 45¢ per dwelling unit. The expense of procuring these data, however, may be charged against a preliminary advance for surveys or as a project cost provided a number of conditions set by HHFA are met. The deadline for this arrangement was the middle part of January.
Special Tabulations and Surveys for Local Public Housing Authorities Under the Housing Act of 1949
Arrangements were also made by the Public Housing Administration with the Bureau of the Census to permit local housing authorities to obtain data prior to the regular publication of the 1950 Census on the characteristics of low-income families and on substandard housing, this information to be used as a basis for planning local public housing programs. Since the 20 percent sample on which income data is to be secured in the 1950 Census does not give detailed information which is significant for cities of under 60,000 population (this is especially true for cities of under 10,000 population), the Bureau of the Census will supplement the data and make special tabulations, provided that the local authorities have entered into contract with the Bureau by March 15, 1950. (These requests for special tabulations and supplementations must be channeled through PHA field offices).
Local housing authorities, in their requests for reservation of funds under the Housing Act of 1949, must submit evidence indicating that the proposed programs relate to the needs of low-income families in terms of the numbers of such families, the racial groups to be served, the rent-paying capacities of the families, the size of families, etc. The more speedy tabulation of data and/or the extension of the number of units to be included in the enumeration helps local housing authorities to meet the requirements set by PHA in defining the housing needs of the community.
The Bureau of the Census has informed us it has already undertaken a number of housing surveys in moderate-sized cities for local housing authorities. These surveys will provide data in response to the questions to be asked in the 1950 Censuses of Population and Housing but in quicker and more complete form than that anticipated from the Census. The cost of these surveys range from $8,000 to $12,000 depending on the size of the city.
The Bureau of the Census is equipped to undertake special surveys for cities. Because of its long experience in the acquisition of and processing of data, some cities may prefer to make contracts with the Bureau rather than with other survey groups.
The census conducts a monthly current population survey throughout the country covering some 25,000 households in 68 areas. It is possible, in some cases, to expand the sample of households in cities included in this monthly current population survey to provide estimates for the city itself instead of merely presenting a total applicable to the country as a whole and not to the individual city. Items covered include school enrollment, migration, consumer income, labor force participation. As a part of the census program, a number of these population and labor force surveys were made in 1946 and 1947 for particular cities.
A special sample survey was made in Chicago in 1949 under the auspices of the Chicago Community Inventory supported in part by the Chicago Association of Commerce and Industry, the Health Department of Chicago, the Public Schools of Chicago, the Chicago Urban League, the Cook County Housing Authority, Federal Reserve Board of Chicago, Illinois Department of Labor, Illinois Bell Telephone Company, Michael Reese Hospital Planning Staff, and the U. S. Bureau of Labor Statistics. This survey was designed to provide information similar to that obtained in the decennial census — including all the information obtained by that census. About 3,700 households were included in this survey at a total cost of about $ 20,000.
Another type of survey which was done on a reimbursable basis, was the Washington Metropolitan District Income Survey in February, 1948. This survey obtained data on family income for the use of the National Capital Park and Planning Commission. The sample consisted of approximately 3,600 households; the total cost was approximately $20,000.
There are many cities that are not covered by the current population survey samples. These cities might possibly wish to make a special local survey or to extend the type of question that is asked in the current population survey to their cities. Such special surveys have usually been undertaken by cities in the period between decennial censuses. The Bureau of the Census is not in a position to undertake any special surveys for cities until the field work connected with the 1950 Census has been completed.
Obtaining of Unpublished Tabulations of Census Data
In 1940, there was a large amount of unpublished detail which was available on request. It is anticipated, although plans for publishing the 1950 Census are not yet completed, that there will be considerably more detailed information that will be in published form than was the case of the 1940 Census. However, some data will be presented in a great amount of detail for the county and for the total urban and total rural areas of the county, but not for the separate cities within the county.
If a particular city wished to obtain data for its own jurisdiction which is not published in separate form, it could request this information from the Census. The Bureau of the Census will make available unpublished data, either in photostatic copies or in microfilm copies of the worksheets used by the Census. Photostatic copies are generally about 40¢ per sheet (18 inches by 24 inches) while microfilm copies can be supplied for about 2¢ per sheet, for the same size sheet. Occasionally, the original machine sheets are photostated or microfilmed; data supplied on this basis are often costly because they include on the same sheet many items or areas not of interest. On the other hand, if these other items or areas are to be excluded, then the cost of special transcription sheets of tabulated data for particular area would have to be figured into the estimates of cost, if such transcriptions were required. Of course, the tabulation would be the responsibility of the particular person or city requesting the information.
*The 1930 and 1940 Census definition was: a private family comprises the family head and all other persons in the home who are related to the head by blood, marriage or adoption, and who live together and share common housekeeping arrangements. A person living alone counted as a 1-person private family. A family head sharing his living accommodations with one or more unrelated persons is also counted as a one person private family (with one or more lodgers).
Copyright 1950 American Society of Planning Officials