Relation to Standard Industrial Classification.-List C shows for each Population Census category the code designation of the similar category or categories in the Standard Industrial Classification (SIC). This relationship is presented here for general information purposes only and does not imply complete comparability. The SIC, which was developed under the sponsorship of the United States Bureau of the Budget, is designed primarily for the classification of reports on industry obtained from establishments. These reports are, by their nature and degree of detail, considerably different from reports on industry obtained from household enumerations such as the Population Census. As a result, many distinctions called for in the SIC cannot be observed in the Population Census. Furthermore, the needs which the Population Census data are designed to meet frequently differ from the needs which the establishment data meet. Perhaps the most basic difference between the two systems is in the allocation of government workers. The SIC classifies all government agencies in a single major group whereas the Population Census industrial classification system allocates them among the various groups according to type of activity, as explained in the next paragraph. Definition of "Public administration."-The major group "Public administration" includes only those activities which are uniquely governmental functions, such as legislative and judicial activities and most of the activities in the executive agencies. Government agencies engaged in educational and medical services and in activities commonly carried on also by private enterprises, such as transportation and manufacturing, are classified in the appropriate industrial category. For example, persons employed by a hospital are classified in the "Hospitals" category regardless of whether they are paid from private or public funds. The total number of government workers appears here in the data on class of worker; of particular significance in this connection is the crossclassification of industry by class of worker (tables 133 and 161). Relation to certain occupation groups. In the Population Census classification systems, the industry category “Agriculture" is somewhat more inclusive than the total of the two major occupation groups, "Farmers and farm managers" and "Farm laborers and foremen." The industry category includes, in addition to all persons in these two major occupation groups, (a) other persons employed on farms, such as truck drivers, mechanics, and bookkeepers, and (b) persons engaged in agricultural activities other than strictly farm operation, such as crop dusting or spraying, cotton ginning, and landscape gardening. Similarly, the industry category "Private households" is somewhat more inclusive than the major occupation group "Private household workers." In addition to the housekeepers, laundresses, and miscellaneous types of domestic workers covered by the major occupation group, the industry category includes persons in occupations such as chauffeur and secretary, if they work for private households.

Class of Worker

The class-of-worker information, as noted above, refers to the same job as does the occupation and industry information. The allocation of a person to a particular class-of-worker category is basically independent, however, of the occupation or industry in which he worked. The classification by class of worker consists of four categories which are defined as follows:

1. Private wage and salary workers.-Persons who worked for a private employer for wages, salary, commission, tips, pay-in-kind, or at piece rates.

2. Government workers.-Persons who worked for any governmental unit (Federal, State, or local), regardless of the activity which the particular agency carried on.

3. Self-employed workers.-Persons who worked for profit or fees in their own business, profession, or trade, or who operated a farm either as an owner or tenant. Included here are the owneroperators of large stores and manufacturing establishments as well as small merchants, independent craftsmen and professional men, farmers, peddlers, and other persons who conducted enterprises of their own. Persons paid to manage businesses or farms

are classified as private wage and salary workers (or, in some few cases, as government workers).

4. Unpaid family workers.-Persons who worked without pay on a farm or in a business operated by a member of the household to whom they are related by blood or marriage. The great majority of unpaid family workers are farm laborers.

The relatively small number of employed persons for whom class of worker was not reported has been included among private wage and salary workers unless there was evidence on the census schedule that they should have been classified in one of the other class-of-worker categories.

FIGURE 32.-Percent Change, 1940 to 1950, IN THE NUMBER OF EMPLOYED PERSONS IN THE UNITED STATES, BY CLASS OF WORKER

1940 Census.-The changes in schedule design and interviewing techniques for the labor force questions, as explained in the section on "Employment status," do not affect comparability between 1940 and 1950 for most of the occupation, industry, and class-ofworker categories. There is evidence, however, that for the categories which include relatively large proportions of female unpaid family workers ("Farm laborers, unpaid family workers," "Agriculture," and "Unpaid family workers"), the 1940 data are sometimes understated by an appreciable amount relative to 1950.

For experienced unemployed persons the 1950 occupation data are not comparable with the data shown in the 1940 Third Series bulletin for the United States. The occupation data for public emergency workers (one of the two component groups of the unemployed in 1940) refer to "current job," whereas the "last job” of the unemployed was reported in 1950.

The occupational and industrial classification systems used in 1940 are basically the same as those of 1950. There are a number of differences, however, in the title and content for certain items, and in the degree of detail shown for the various major groups. The 1940 classification by class of worker is comparable with the 1950 classification system.

The process of adjusting the 1940 detailed occupation and industry data for comparability with the 1950 figures cannot be explained in this volume because of the great deal of material involved in this process. It appears useful, however, to give here some indication of the comparisons and adjustments to be made in deriving roughly comparable data on the major group level.

The listings which follow show only those major groups where significant adjustments are necessary, or where the comparative groups are not immediately identifiable. Furthermore, the adjustments shown are those which can be estimated fairly readily from the census data. This rough procedure yields figures for employed males and females in each major group which, for the United States as a whole, differ from the figures obtained by the more detailed adjustment procedure by a maximum of about 6 percent and

For major occupation groups, the comparisons are as follows: Professional, technical, and kindred workers.-Comparable with Professional and semiprofessional workers. Add accountants and auditors to 1940 figure.

Managers, officials, and proprietors, except farm.-Comparable with Proprietors, managers, and officials, except farm.

Clerical and kindred workers and Sales workers.-Comparable with Clerical, sales, and kindred workers. Subtract accountants and auditors from 1940 figure.

Craftsmen, foremen, and kindred workers.-Comparable with Craftsmen, foremen, and kindred workers. Add linemen and servicemen, telegraph, telephone, and power; and motion picture projectionists to 1940 figure.

Operatives and kindred workers.-Comparable with Operatives and kindred workers. Subtract linemen and servicemen, telegraph, telephone, and power; and motion picture projectionists from 1940 figure.

Private household workers.-Comparable with Domestic service workers for both sexes combined and for females; not comparable for males.

Service workers, except private household.-Comparable with Protective service workers and Service workers, except domestic and protective. Note that the 1950 occupation data are limited to civilians, thereby necessitating the removal from the 1940 group of soldiers, sailors, marines, and coast guards.

For major industry groups, the comparisons are as follows:

Transportation, communication, and other public utilities.— Comparable with Transportation, communication, and other public utilities. Subtract radio broadcasting and television from 1940 figure.

Business and repair services.-Comparable with Business and repair services. Add accounting, auditing, and bookkeeping services to the 1940 figure.

Entertainment and recreation services.-Comparable with Amusement, recreation, and related services. Add radio broadcasting and television to the 1940 figure.

Professional and related services.-Comparable with Professional and related services. Subtract accounting, auditing, and bookkeeping services from the 1940 figure.

Public administration.-Comparable with Government for both sexes combined and males; not comparable for females. Note that the 1950 industry data are limited to civilians thereby necessitating the removal from the 1940 group of national defense.

The 1940 occupation and industry data shown in this volume include adjustments which take account of the differences between the 1940 and 1950 classification systems. In order to maximize the amount of comparable data, it was sometimes necessary to estimate the adjustments from information which was incomplete or not entirely satisfactory for the purpose. Furthermore, there were certain differences between the 1940 and 1950 coding and editing procedures which could not be measured statistically. Caution should, therefore, be exercised in interpreting small numerical changes. Caution should also be exercised with regard to large relative increases in the numbers of women engaged in occupations which are unusual for women. Although it is certainly true that women have expanded the range of their occupational activities during the last decade, the figures shown here may, in some cases, tend to overstate this expansion because more intensive checking of questionable returns of this type was performed in 1940 than in 1950; this overstatement is particularly true of the railroad occupations.

The 1940 data on occupation, industry, and class of worker shown in this volume have been revised to eliminate members of the armed forces in order to achieve comparability with the 1950 figures for the employed, which are limited to civilians. In the occupation tables of the 1940 reports, the armed forces were mainly included in the major group "Protective service workers." In the industry tables, the armed forces were all included in the major group "Government." In the class-of-worker tables, the armed forces were all included in the category "Government workers" (or in the total "Wage or salary workers").

The 1940 major occupation group figures presented in Chapter C of this volume may differ in some cases from the corresponding

Chapter C were developed by a more detailed analysis of the 19401950 classification differences than were the figures in Chapter B. 1930 and earlier censuses.-Prior to 1940, the census data on the economically active population referred to "gainful workers" rather than to "labor force." The differences between these two concepts are described in the section on "Employment status." The effects of this variation in approach on the various occupation and industry categories are virtually impossible to measure. For most categories, the number of gainful workers is probably equivalent to the number of persons in the experienced civilian labor force. For certain categories, however, particularly those with relatively large numbers of seasonal workers, the gainful worker concept probably yields somewhat larger figures than does the labor force concept.

The occupational and industrial classification systems used in the 1930 Census and earlier censuses were markedly different from the 1950 systems. The relationships between the present and earlier systems are being analyzed, and the results of the study will be made available by the Bureau of the Census. For information on occupation and industry data from 1930 and earlier censuses, see the publication, U. S. Bureau of the Census, Sixteenth Census Reports, Population, Comparative Occupation Statistics for the United States, 1870 to 1940, Washington, D. C., 1943.

Other data. Comparability between the statistics presented in this volume and statistics from other sources is frequently affected by the use of different classification systems, as well as by many of the factors described in the paragraphs on comparability with Current Population Survey data and other data in the section on "Employment status." In regard to comparisons between occupation figures from the Population Census and those based on data from government licensing agencies, professional associations, trade unions, etc., comparability may not be as direct as would first appear. Among the sources of difference may be the inclusion in the organizational listing of retired persons or persons devoting all or most of their time to another occupation, the inclusion of the same person in two or more different listings, and the fact that relatively few organizations attain complete coverage or membership in an occupation field. Differences in definition and concept, as well as errors in census data, may also be contributing factors to lack of comparability.

Quality of Data

The omission from the labor force of some marginal workers (mainly part-time and occasional workers), as explained in the section on "Employment status," has probably resulted in some understatement in many of the occupation, industry, and class-ofworker figures. Another factor to be considered in the interpretation of these data is that enumerators sometimes returned occupation and industry designations which were not sufficiently specific for precise allocation. One cause may have been the enumerator's lack of knowledge of how to describe a particular job on the census schedule. Another possible cause was inadequate information about the worker's job on the part of the housewife or other person from whom the enumerator obtained the report. Indefinite occupation and industry returns can frequently be assigned, however, to the appropriate category through the use of supplementary information. For example, in the case of occupa tion, the industry return on the census schedule is often of great assistance. In the case of indefinite industry returns, helpful information can frequently be obtained from outside sources regarding the types of industrial activity in the given area. The basic document used in the allocation of the schedule returns of occupation and industry to the appropriate categories of the classification systems is the publication, U. S. Bureau of the Census, 1950 Census of Population, Alphabetical Index of Occupations and Industries, Washington, D. C., 1950.

The application of detailed occupational and industrial classifications to approximately 60 million workers is obviously subject

ably does not have any serious effect on the usefulness of most of the data, there are a few cases where relatively small numbers of erroneous returns may produce what might be regarded as a serious misstatement of the facts. These cases relate mainly to the numbers of men, women, and children shown in occupations which are unusual for such persons, and to the government workers shown in industries that are ordinarily not carried on by government agencies. Some of the more obvious misclassifications have been adjusted, but it was not possible to perform a complete review of the data for all discrepancies.

Information on certain aspects of the quality of the occupation, industry, and class-of-worker data is available from the PostEnumeration Survey. The Post-Enumeration Survey material refers only to employed persons and, in the case of occupation and industry, is as yet tabulated only for major groups. One important factor which affects the quality of the data on occupation, industry, and class of worker-the reporting of employment status was not investigated, mainly because the time interval between the census and Post-Enumeration Survey enumerations appeared too great to yield adequate information on an item so subject to change as employment status.

In general, the percent distributions by major occupation group, major industry group, and class of worker were only slightly affected by errors of coverage, errors in reporting a job description, and errors in reporting age. With few exceptions, the percentage of employed persons in each such category was affected by less than one percentage point. The stability in the percentage was found both when corrections were made for all three types of errors combined and when corrections were made for each type separately. This stability also existed with regard to the distributions for both sexes combined and for males and females separately.

The accuracy of the count of persons in each major group was also measured by the Post-Enumeration Survey. The absolute number of employed persons in most of the major occupation, major industry, and class-of-worker groups was in error by less than 4 percent as a result of the three types of error measured.

INCOME

Definitions

Components of income.-Income, as defined in the 1950 Census, is the sum of money received from wages or salaries, net income (or loss) from self-employment, and income other than earnings. The figures in this report represent the amount of income received before deductions for personal income taxes, social security, bond purchases, union dues, etc.

Receipts from the following sources were not included as income: money received from the sale of property, unless the recipient was engaged in the business of selling such property; the value of income "in kind," such as food produced and consumed in the home or free living quarters; withdrawals of bank deposits; money borrowed; tax refunds; gifts and lump-sum inheritances or insurance payments.

Information was requested of a 20-percent sample of persons 14 years of age and over on the following income categories: (a) The amount of money wages or salary received in 1949; (b) the amount of net money income received from self-employment in 1949; and

the amount of other money income received in 1949. If the sam; le person was the head of a family, these three questions were repeated for the other family members as a group in order to obtain the income of the whole family. The composition of families is that of the time of interview, although the time period covered by the income statistics is the calendar year 1949. Specific definitions of these three categories of income are as follows:

Wages or salary. This is defined as the total money earnings received for work performed as an employee. It includes wages, salary, armed forces pay, commissions, tips, piece-rate payments,

Self-employment income.-This is defined as net money income (gross receipts minus operating expenses) from a business, farm, or professional enterprise in which the person was engaged on his own account or as an unincorporated employer. Gross receipts include the value of all goods sold and services rendered. Expenses include the costs of goods purchased, rent, heat, light, power, depreciation charges, wages and salaries paid, business taxes, etc.

Income other than earnings.-This includes money income received from sources other than wages or salary and self-employment, such as net income (or loss) from rents or receipts from roomers or boarders; royalties; interest, dividends, and periodic income from estates and trust funds; pensions; veterans' payments, armed forces allotments for dependents, and other governmental payments or assistance; and other income such as contributions for support from persons who are not members of the household, alimony, and periodic receipts from insurance policies or annuities. Information on the income distribution of families and unrelated individuals is presented in Chapter B. Statistics on the income of all persons 14 years of age and over are presented in Chapter C. When comparisons among States or other areas are made, account should be taken of the fact that this report covers money income only and not the value of goods produced and consumed on the farm. The money income distributions of families or persons in areas containing a relatively large proportion of farm residents may not be an adequate indicator of income levels in those areas. It should also be remembered that changes in money income between 1939 and 1949 were accompanied by changes in prices; therefore, an increase in wage or salary income does not represent a similar change in economic well-being.

Comparability

1940 Census. In 1940 all persons 14 years of age and over were asked to report (a) the amount of money wages or salary received in 1939 and (b) whether income amounting to $50 or more was received in 1939 from sources other than money wages or salaries. Comparable wage or salary income distributions for 1940 and 1950 are presented in tables 144 and 166 of this report. All of the other 1950 Census income data shown in this report relate to total money income and are not comparable with statistics from the 1940 Census.

Office of Business Economics income payments series.-The Office of Business Economics of the Department of Commerce publishes data on the aggregate income received by the population in each State. If the aggregate income were estimated from the census income distributions, it would be smaller than that shown in the State income payments series for several reasons. The Bureau of the Census data are obtained by household interview, whereas the State income payments series is estimated largely on the basis of data derived from business and governmental sources. Moreover, the definitions of income are different. The State income payments series includes some items not included in the income data shown in this report, such as income in kind, the value of the services of banks and other financial intermediaries rendered to persons without the assessment of specific charges, and the income of persons who died or emigrated prior to April 1, 1950. On the other hand, the Bureau of the Census income data include contributions for support received from persons not residing in the same living quarters.

Federal Reserve Board Survey of Consumer Finances.-The Federal Reserve Board Survey of Consumer Finances obtained, among other data, information on the size distribution of income for 1949. The survey was based on a nationwide sample of approximately 3,500 units in 66 areas, covering all persons in private households. Interviews were taken and most of the results were presented on a "spending unit" basis, those household members who had incomes which they used primarily for their own purposes constituting separate spending units. Tabulations of certain financial data, however, were presented on a "family unit" basis as well. The "family unit" is identical with the concept of family or individual used by the Bureau of the Census. Differ

and the statistics in this volume are subject to sampling variability and arise from differences in estimating procedure as well as from differences in the number and wording of the questions asked.

Income tax data. For several reasons, the income data shown in this report are not directly comparable with those which may be obtained from statistical summaries of income tax returns. Income as defined for tax purposes differs somewhat from the Bureau of the Census concept. Moreover, the coverage of income tax statistics is less inclusive because of the exemptions of persons having small amounts of income. Furthermore, some of the income tax returns are filed as separate returns and others as joint returns; and, consequently, the income reporting unit is not consistently either a family or a person.

Bureau of Old Age and Survivors Insurance wage record data.The wage or salary data shown in this report are not directly comparable with those which may be obtained from the wage records of the Bureau of Old Age and Survivors Insurance for several reasons. The coverage of the wage record data for 1949 is less inclusive than the 1950 Census data because of the exclusion of the wages or salaries of such groups as domestic servants, farm laborers, governmental employees, and employees of nonprofit institutions. Furthermore, no wages or salaries received from any one employer in excess of $3,000 in 1949 are covered by the wage record data. Finally, as the Bureau of the Census data are obtained by household interviews, they will differ from the Old Age and Survivors Insurance wage record data which are based upon employers' reports.

Quality of Data

The figures in this census, as in all field surveys of income, are only approximately accurate because of errors of response and nonreporting. In most cases the schedule entries for income are

Other errors of reporting are due to misunderstanding of the income questions or to misrepresentation. In addition, the number of persons shown at each income level is subject to error because of the omission of persons who were not enumerated in the census. Despite these sources of error, however, the various checks which have been made on the quality of the 1950 Census income data suggest that they may be sufficiently reliable for the analysis of the characteristics of the population at different income levels. A preliminary analysis of the results obtained in the PostEnumeration Survey indicates that, despite considerable variability of response, there is no important difference between the census and the survey in the percent distribution of income recip ients by income levels. For persons, the median income from the census was $1,917 and from the check, $1,840. Moreover, the Post-Enumeration Survey results suggest that the 1950 Census income distributions would have been substantially the same hai information been available for persons who were not enumerated in the census or who were enumerated but for whom the income questions were left blank. The Post-Enumeration Survey also found a relatively larger number of income recipients than did the census; about 64 percent of the population 14 years old and over enumerated in the census are shown as income recipients in the census as compared with 68 percent in the Post-Enumeration Survey. Most of the additional income recipients reported in the Post-Enumeration Survey were in the lowest income levels and their inclusion affects only moderately the distribution by income level.

A comparison of the income distributions based on the 1950 Census and on the March 1950 Current Population Survey shows no significant difference for families and for persons. Moreover, a comparison of the incomes reported by a matched sample of about 5,000 persons interviewed in both the 1950 Census and the

supports strongly the conclusions based on the Post-Enumeration Survey. This matched sample, like the Post-Enumeration Survey, reveals considerable variability of response but little net difference between the income distributions.

Although the primary purpose of the income questions in the 1950 Census was to provide distributions of families and of persons 14 years of age and over by income levels, rough estimates of aggregate income can be derived from these data. A comparison of aggregate income estimated from preliminary samples of 1950 Census returns with those prepared by the National Income Division of the Department of Commerce indicates that aggregate total money income derived from the 1950 Census income data for persons was 92 percent of the National Income Division estimates after the latter have been adjusted to make them as nearly comparable as possible with the income concept used by the Bureau of the Census. The comparable ratio for wage or salary income was over 95 percent.

The estimated census aggregate derived from the data for families and unrelated individuals was slightly over 80 percent of the comparable National Income Division estimate. The rela tively greater understatement of income for families and unrelated

SAMPLE DESIGN

individuals is attributable in part to the fact that the method used to obtain family income data in the census was inferior to the method used to obtain income data for persons. The coding procedures may have also created a downward bias in the family income data. In coding it was assumed that there was no other income in the family when only the head's income was reported and income information was not reported for other family members. It is estimated that this editing assumption was made for about 5 percent of the families. This procedure was adopted in order to make maximum use of the information obtained. In a large majority of the fully reported cases, the head's income constituted all or most of the total family income.

The income tables for the United States include in the lowest income group (under $500) about 1.6 million families and about 1.8 million unrelated individuals who were classified as having no 1949 income, as defined in the census. Many of these were living on income "in kind," savings, or gifts, or were newly created families or families in which the sole breadwinner had recently died or left the household. A relatively large proportion, however, probably had some money income which was not recorded in the

census.

Some of the data in the tables which follow are indicated by asterisks or by headnotes as being based on a representative 20-percent sample of the population. A separate line was provided on the population schedules for each person enumerated, with every fifth line designated as a sample line. Within each enumeration district, the schedules were divided approximately equally among five versions. On each version the sample constituted a different set of lines so that each line on the schedule was in the sample on one of the five versions. The persons enumerated on these sample lines were asked all of the pertinent sample questions." In addition, this sample served as the basis for some of the tabulations of data obtained for the entire population.

Although the sampling plan did not automatically insure an exact 20-percent sample of persons in each locality, it was unbiased and for large areas the deviation from 20-percent was expected to be quite small. Small biases, however, arose when the enumerator failed to follow his sampling instructions exactly. For the United States as a whole, tabulations for Chapter B show that the proportion of the total population enumerated in the sample was 19.95 percent. The proportion of the total population in the sample, by regions, was 19.94 percent in the Northeast, 19.93 in the North Central States, 19.97 in the South and 19.96 in the West. Among States, the proportions in the sample ranged from 19.86 percent to 20.00 percent.

The small biases which arose as a result of enumerator error were usually in the direction of a slight underrepresentation of adult males, particularly heads of households. The proportion of bousehold heads in the sample was 19.73 percent and the proportion of all other persons was 20.04 percent.12 Among males 25 years of age and over, 19.71 percent were enumerated in the sample; among females of comparable age, the proportion in the sample was 20.03 percent. As a consequence of this bias, sample statistics on such items as the size and composition of the male labor force are somewhat affected.

Estimates of the number of persons with specified characteristics based on sample data have in all cases been obtained by

"In 19 counties of Michigan and Ohio, the sample consisted basically of every fifth household and the pertinent sample questions were directed to all persons in the housebaki. Such a household sample was used as an experiment to determine the feasibility of this type of sample in future censuses of population.

In the experimental areas of Michigan and Ohio, biases due to the underenumeration of household beads did not exist, although other problems arose because some

multiplying the number of persons in the sample with these characteristics by five. Estimates of percentages have been obtained in each case by using the sample values for both the numerator and denominator.

SAMPLING VARIABILITY

The figures based on the 20-percent sample are subject to sampling variability which can be estimated from the standard errors shown in tables R and S. These tables do not reflect the effects of the biases mentioned above. The standard error is a measure of sampling variability. The chances are about 2 out of 3 that the difference due to sampling variability between an estimate and the figure that would have been obtained from a complete count of the population is less than the standard error. The amount by which the standard error must be multiplied to obtain other odds deemed more appropriate can be found in most statistical textbooks. For example, the chances are about 19 out of 20 that the difference is less than twice the standard error, and 99 out of 100 that it is less than 21⁄2 times the standard error.

Illustration: Table 42 shows that there were an estimated 181,850 rural-farm persons 18 and 19 years of age enrolled in school in the United States in April 1950 (25.1 percent of the estimated 725,295 rural-farm persons in this age group). Table 34 indicates that the total rural population of the United States was 54,229,675. By linear interpolation, from table R, the standard error for an estimate of 182,000 in an area with 54,000,000 inhabitants may be estimated at about 860. Consequently, the chances are about 2 out of 3 that the figure which would have been obtained from a complete count of the number of rural-farm persons 18 and 19 years who were enrolled in school would have differed by less than 860 from the sample estimate. It also follows that there is only about 1 chance in 100 that a complete census result would have differed by as much as 2,150, that is, by about 21⁄2 times the number obtained from the table. Table S shows that the standard error of the 25.1 percent on a base of 725,000 is about 0.1 percent. For most estimates, linear interpolation will provide reasonably accurate results.13