12. Groups of Industries

The Standard Industrial Classification system used in this census groups the 450-odd individual manufacturing industries mentioned above into 141 subgroups comprising one individual industry or a small number of closely related individual industries and

the inclusion in the 1947 census of data for establishments whose scale of activity was so small as to have been eliminated by the $5,000 value of products limitation in 1939 and earlier years. It was, of course, impossible to include in the 1939 tables any figures for the two industries (coffee and spice roasting and

several industry subgroups. Each individual industry is designated by a 4-digit code, each industry subgroup by a 3-digit code identical with the first three digits of its component industries, and each major industry group by a 2-digit code identical with the first two digits of its component industry subgroups.

For brevity, individual industries are sometimes referred to as 4-digit industries, subgroups as 3-digit industries, and major industry groups as 2-digit industries.

In all the summary industry tables as well as in the State or area tables by industry, subtotals for industry subgroups and major groups are shown where possible, and codes for each industry level are shown at the left of the industry name. Where a

ered in 1947 but not in 1939. On the other hand, the statistics for industries covered in 1939 but not in 1947 have been eliminated from the 1939 tables. In some instances the 1939 returns did not supply sufficient information to permit a completely satisfactory manufacturing" activities as defined in 1947; the distinction between "manufacturing" and "non

lack of comparability between 1947 and 1939 statistics arising from this difficulty would be small on an over-all basis, though in small counties or industryState combinations there might well be some distortion of the 1939-1947 relationship introduced by differences in the classification of identical establishments in the two years.

Within the manufacturing area proper, there were

industry, only the 4-digit industry name and code are shown. In order to save space in these tables, shorter titles have been substituted for many of the titles shown in the Standard Industrial Classification.

As indicated in Section 14 below, a more prominent place has been given to groups of industries in this as compared with earlier censuses. The principal reasons for this change in emphasis are threefold: (1) Most important for census purposes, States, or local areas can frequently be compared at broader industry levels, where no comparative data can be published at the 4-digit level; (2) the census data are frequently used as "bench-marks" for other series on manufacturing for which the coverage is not adequate or the production data not sufficiently complete to justify publication of the statistics on a 4-digit basis; and (3) the limitations attaching to year-toyear comparisons of 4-digit statistics (cf. Section 10 above) do not apply in equal degree to 3- and 2-digit statistics. A minor shift in the character of an establishment's production may change its 4-digit industry classification, where its 3-digit classification is frequently, and its 2-digit classification nearly always, unaffected.

13. Statistics for 1939 and Earlier Years

The changes in the scope of the census between 1939 and 1947 have already been discussed in Section 8 above. No attempt was made to eliminate from the 1939 tabulations data for establishments reporting no employees in that year, but, as indicated, this adjust

criteria used for classifying establishments into individual industries. These changes are described in full in Appendix D. As in former years, the most basic general statistics (number of establishments,

production workers and their wages, cost of materi

als, value of products, and value added by manufacture) were retabulated for all industries affected by

changes in definition. Other data such as the number and salaries and wages of all employees, inventories, capital expenditures, etc., were not retabulated for 1939. Total employment and salaries and wages are shown for all industries not affected by changes in definition and estimated figures are shown for States and metropolitan areas; for all other comparable 1939 general statistics, the reader is referred to the 1939

census volumes.

Whenever possible, these basic general statistics, in addition to those for total employment and pay rolls, have been shown for census years back to 1899. In general, the series for each industry was carried back as long as comparability was not affected by more than 10 percent as a result of the change in definition. Comparability was measured in terms of the difference between the original and retabulated value added by manufacture figures for the year in question. If the difference was less than 10 percent, the earlier years were considered comparable with the later years and were shown in the series; if the difference exceeded 10 percent, the earlier years were considered noncomparable and the series was discontinued at the point of change. It was not possible to

the schedules were not retabulated following changes in definition, and in such cases the determination of comparability was usually based on the relative importance of the products involved in the change for the year in question. Figures for years earlier than 1939 were taken from the published volumes. In general, significant differences resulting from changes in definition have been explained in headnotes or footnotes; minor changes are merely indicated by the footnote "Revised." In a few instances, figures for some years have been eliminated to save space.

In a number of instances, comparable 1939 industry data are not available because of differences in census procedure in the two years, e.g., in 1947 separate reports were required of all manufacturers producing paper and board and further converting them into paper and board products in an integrated operation. In 1939, many such manufacturers filed consolidated reports and the integrated operation was classified in the Paper and Board Industry. It was not, however, feasible to retabulate the 1939 reports. Consequently, relative to 1947, the 1939 employment and other industry figures are overstated for the Paper and Board Industry and understated for some of the converting industries. The general statistics for the major industry (2-digit) group, Paper and Paper Products, are, however, comparable for the two years. All such instances are noted in the text for the industries affected.

Two instances of changes in census procedure should be noted particularly since they affect the industry figures for nearly all States. In the bakery industries, 1939 production-worker employment, the only 1939 figure shown by State and by industry, is high relative to 1947 because in the earlier year driver-salesmen were classified as production workers whereas in 1947 they have been included with salaried employees. The effect of this change may be indicated by noting that for the country as a whole production-worker employment "declined" from 180,000 to 150,000 between 1939 and 1947, whereas total employment rose from 217,000 to 233,000. Value added by manufacture was not, of course, affected by this change in the method of reporting. In the dairy products industries the increase in both value added and employment from 1939 to 1947 is somewhat understated because in 1939 separate reports were obtained from the manufacturing departments of plants distributing fluid milk whereas in 1947 the whole plant was treated as "nonmanufacturing" in accordance with standard classification procedures, and excluded from the tables presented in the State and General Summary Volumes. In consequence, the 1939 figures may be too high by 25 percent as compared with 1947. For the Food and Kindred Prod

two factors is to overstate the 1939 productionworker employment by 7 to 8 percent relative to 1947. It was not feasible to retabulate 1939 data for the State Volume even in the detail employed for industries. On an individual industry basis, it was feasible to retabulate only production-worker employment and that only for the State (table 4); consequently, no 1939 industry comparisons are available for metropolitan areas or principal industrial counties. Value added by manufacture was also retabulated for the State by major industry group (table 3) and for each of its subdivisions on an over-all basis. For the State and its standard metropolitan areas (table 1), wages of production workers were retabulated and the number and salaries and wages of all employees were estimated for 1939 on the basis of their relationship to production workers and wages in that

year.

14. The Census Report Form

Although only one form was required from each establishment, the diversity of manufacturing activities made it necessary to use more than 200 different report forms to canvass the various industries included in the 1947 census. For convenience, however, these may be classified into four types:

(a) Most of the 200 forms represented variations of the "standard" form. This form called for detailed general statistics, and in addition contained inquiries adapted to the requirements for information on production of commodities, materials consumption, and equipment in use for one or several manufacturing industries (cf. Appendix B).

The production and materials section of each form was the result of intensive study of the 1939 census reports and extensive consultation with other Federal agencies, manufacturers, and trade associations in an effort to provide the most useful information consistent with the ability of manufacturers to report the desired data.

Nearly two-thirds of all manufacturers reporting in the census used one variety or other of the standard form. One variety of the standard form was the "general" schedule which was used in a number of industries producing commodities which could not be specified in significant detail.

(b) A "short" form, designed to cover very small plants outside the lumber industry was originally mailed to 15,000 establishments which reported in the precanvass that they employed only one person during 1947 or went out of business in that year, but was finally used for approximately 40,000 establishments accounting for 200,000 employees or 1.4 percent of total employment in manufacturing. This form (cf. Appendix B) contained a product section similar to that on the "general" schedule but requested in addition only the following general information: Total employment as of four months of the year, salaries and wages, and total cost of materials. In order to simplify the presentation of the main census tables, production workers and wages were estimated by assuming that there were no salaried employees in establishments with fewer than ten employees, and by applying ratios derived from the 1939 census in the few cases in which employment was ten or more. Man-hours were derived by assuming a

(c) A short form was specifically designed for sawmills producing between 200 M ft. b.m. and 1,000 M ft. b.m. of lumber during 1947. The form contained a product inquiry somewhat less detailed than that used by larger mills, and a materials consumed inquiry identical with that on the standard lumber form, but the general information requested was the same as that requested on the short form, except that total employment was requested for each month of the year. This form was used by 10,000 sawmills accounting for 75,000 employees. All employees reported on this form were assumed to be production workers and man-hours were again assumed to average 2,000 per employee-year.

(d) A short form was also designed for the use of smaller establishments in all printing and publishing industries, except book and magazine publishing. This form was used by some 15,000 establishments with 75,000 employees, most of which had reported in the precanvass that they employed more than one but less than ten employees in 1947. (Establishments employing one person in the printing and publishing industries received the regular short form.) The product section of this form was again less detailed than the standard form for these industries, and the general information requested was the same as that contained in the regular short form. Since, however, employment is generally stable in these industries employment data were estimated and tabulated for the eight months not reported on the form. The ratio of production workers to total employees obtaining for the smaller establishments in each of these industries in 1939 was used to derive the number of production workers, their wages and hours of work.

The practical importance of the use of these short forms is as follows: The statistics on number of establishments, average employment, and pay rolls, cost of materials, value of shipments, and value added by manufacture, where shown, represent all establishments covered in the census; the monthly employment statistics represent all establishments reporting on the standard forms and the short sawmill and printing forms; all other general statistics cover only establishments reporting on the standard forms. Where figures for which coverage is not complete are shown, the tables either include an appropriate measure of coverage or a cross-reference to another table which does include such a measure.

The principal exception to this last statement is that for expenditures for plant and equipment. Since this item was tabulated by State and by industry it

was not considered feasible to include a coverage measure in every instance in which this figure is shown. In using the expenditures figures it should, therefore, be noted that establishments reporting on this item accounted for 97.5 percent of the employment in all census establishments, and that for the United States as a whole the percent coverage is better than 95 percent for all major industry groups, except Lumber, and Printing and Publishing, and on an individual industry basis is less than 90 percent only for: Sawmills (81 percent), Commercial Printing (85 percent), Machine Shops (84 percent) and for 22 smaller industries whose aggregate employ

Product cards were not punched for establishments reporting on the short form, and in consequence all product detail was lost in tabulation. This accounts for a substantial portion of the "products not specified by kind" in certain industries (table 6 of the Industry Volume).

Although short forms were received from establishments in nearly all industries, it was considered desirable for purposes of tabulation and presentation, to eliminate them in those industries containing only a handful of such forms. In all such instances a standard report was built up from the basic information reported on the short form. This accounts for the fact that only about 150 industries are shown as including short form reports. The percentage of estimate involved in this procedure was usually a fraction of 1 percent.

15. Disclosure of Data for Individual Companies

The Bureau of the Census is prohibited by law from publishing any statistics that disclose information reported by individual companies. In this as in all recent censuses the practice has been as follows: (a) It is not a disclosure to show the number of establishments, even down to one, as to locality, industry classification, or approximate size;

(b) Employment figures may be shown for any industry or locality represented by three or more companies;

(c) Value figures may be shown for any industry, locality or product represented by three or more companies; provided also that one or two companies do not produce a very large proportion of the combined output of these companies;

(d) Care is taken to provide against the probability of disclosure by subtraction. If, for example, the subtraction of the statistics for a city from that for a county in which it is located would isolate a group of establishments not meeting the tests indicated in (b) and (c), the data for the city are withheld.

An important change was, however, made in the handling of data which could not be shown in full detail in this census as compared with earlier censuses. Previously, preference was given to State-byindustry figures. Consequently, regional figures for

an industry were available only if all States within a region were shown separately, or if the sum of the States which could not be shown separately did not constitute a disclosure. Conversely, industry group data were available for a State only if all industries within the group were shown separately or if the sum of the industries which could not be shown did not constitute a disclosure.

In the 1947 census, preference has been given to regions and the nine Census divisions over individual States in table 2 of each Industry report and to industry groups over individual industries in table 4

made because (1) it was considered preferable to show regional data for an industry even at the cost of withholding data for another, usually the least important, State in a region, rather than to show all publishable data for the individual States and group the disclosure States throughout the United States in an "all other" category; and (2) by giving preference to industry group statistics over individual industries within each State, the extent to which individual States can be compared on an industry classification basis is greatly increased. For example, it is frequently possible to compare data for two States at the 3-digit industry group level when no comparative data can be published for any 4-digit industry component of this group.

Table 2 of each Industry report shows summary general statistics for each geographic Census division and State for which this information could be shown. For industries with few establishments or considerable geographic concentration it has frequently been necessary to use geographic regions broader than the traditional nine Census divisions, in order to show all plants by some geographic location. For this purpose the following combinations of Census divisions have been used in the Industry reports:

For each producing State not shown separately in table 2 of the Industry reports, a footnote indicates the number of establishments and, ordinarily, in parentheses, the number of employees for States represented by three or more companies. In some instances, however, the number of employees also had to be withheld in order to avoid disclosing (by subtraction) the number of employees in other States represented by only one or two companies.

Table 4 of each State report shows the industry data for the State. As indicated above, in constructing this table preference was given primarily to the Major Industry (2-digit) Group, for example, Food and Kindred Products; secondarily, to the 3-digit industry group, Dairy Products; and lastly, to the individual (4-digit) industry, Creamery Butter. Where information cannot be shown for a 4-digit industry, it is also withheld for another industry, usually the least important, in the same 3-digit group, in order to avoid disclosure by subtraction. Similarly, information for a 3-digit group will often be withheld to

group in the same State. When 3-digit information is withheld for this reason, it will sometimes happen that information for one or more 4-digit industries within that 3-digit group can be shown without prejudice to the 2-digit major group. This will explain many cases in which a 4-digit industry is shown and the corresponding 3-digit industry suppressed in many tables. In other instances, data for a 4-digit industry only are shown because this was the only 4-digit industry for which reports were received within the 3-digit group.

Table 6 of Chapter I of each census volume shows the number of production workers and value added by manufacture for industries and industry groups and major groups by geographic division and region. It is obvious that table 2 of the Industry reports and table 4 of the State reports could not be prepared independently. For example, figures for a 4-digit industry are frequently eliminated from table 4 of the State report, not to protect the 3-digit information for the State but to protect the 4-digit information for the region (table 2 of the Industry reports). Similarly, 2- and 3-digit information for the State is sometimes withheld to protect 2- or 3-digit figures for the division in table 6, Chapter I of each volume.

In addition to the preference of regions over States, the following priorities were assigned to the lesser geographic areas: States were given preference over standard metropolitan areas (table 5 in the State reports), standard metropolitan areas over principal industrial counties (table 6 in the State reports), principal industrial counties over other counties (table 2 in the State reports), and counties over urban places (table 2 in the State reports).

Information has also been withheld in all the tables mentioned in this section for those 4-digit industries and 3-digit groups in which less than 100 manufacturing employees were reported. The reasons for this are threefold: Earlier census experience shows that as often as not information cannot be shown for smaller "publication cells" without disclosing the operations of individual establishments and some time and money were saved by limiting the “disclosure analysis" required for such cells. Secondly, it was considered desirable to omit lines of relatively little significance in order to make the State tables easier to use. Finally, a not-inconsiderable number of small establishments were inadvertently omitted from the 1947 census (sec. 4) and, it was not possible to review the reports of smaller establishments with the care given to the larger plants; on an industry or State basis, the effect of these considerations is negligible, but the data for any particular "cell" consisting of a small number of small plants are subject to

If, as in earlier years, preference had been given to States and individual industries, statistics could have been shown for 327 California industries accounting for 97 percent of the value added by manufacture for the State; whereas, statistics have actually been shown for 201 industries accounting for 79 percent of the State's value added. On the other hand, 3-digit information has been published for 90 percent as against 70 percent and 2-digit information for 97 percent as against 33 percent of the State's value added under the procedures followed in earlier years. At the same time, industry information has frequently been shown for the Pacific Division, where otherwise it would have been necessary to include the data for Washington or Oregon or both States and in some cases the data for one or more of the Mountain States in an "All other States" total. In the case of California, the most important industrial State in a division consisting of only three States, most of the "suppressions" have been made in order to protect the divisional or regional data for individual industries, rather than to protect the industry group data for the State. Since Maryland is included in a division consisting of nine States, most of the "suppressions" in this State have been made to protect the industry group data for the State itself. In these three States the number of lines omitted because they covered fewer than 100 employees ranged from 16 to 23, but in no case was the value added by manufacture for all omitted lines as much as one-half of one percent of the State total.

Under the 1947 disclosure procedures, comparative industry group data for these three States are available except for three industry groups accounting for less than 3 percent of California's value added, whereas under the procedures used in earlier years, comparative data at any industry level would have been available in less than half the cases in terms of

16. Standard Metropolitan Areas

In addition to the legally constituted units-States, counties, cities-Census of Manufactures statistics have been tabulated for those standard metropolitan areas having 40,000 or more manufacturing employees in 1947. The standard metropolitan areas replace the industrial areas shown in earlier censuses. Industrial areas, as defined in previous years, were groups of contiguous counties having large numbers of manufacturing wage earners. The standard metropolitan areas, on the other hand, are integrated economic and social entities which contain not only highly industrialized counties but also adjoining counties which, though primarily residential in character, contribute significantly to the industrial counties' labor force.

The areas used in this census represent a part of a larger list of standard metropolitan areas defined by the Census Bureau for use in the Censuses of Manufactures, Business, and Population and are recommended by an Inter-Agency Committee on Standard Metropolitan Areas of the Budget Bureau for use throughout the Federal Government. The general concept applied was that an area should be an integrated economic and social entity, with an attendant large volume of daily travel and communication between the central city and the outlying parts of the area. (This usually resulted in a standard metropolitan area considerably more restricted than the market or trading area.) The following principles were used in applying this general concept to the definition of individual areas:

(a) Each standard metropolitan area must include at least one city of 50,000 or more; the area as a whole must have a total population of at least 100,000. Areas may cross State lines.

(b) Where two cities of 50,000 or over are within 20 miles of each other, they will ordinarily be included in the same area. (c) Each county included in a standard metropolitan area must have either 10,000 nonagricultural workers or 10 percent of the nonagricultural workers in the area, or more than one-half of the county's population must have been included in the "metropolitan district" as defined by the Bureau of the Census. In addition, nonagricultural workers must constitute at least two-thirds of the total employed labor force of the county.

(d) Each peripheral county included in a standard metropolitan area must be economically and socially integrated with the central counties of the area. A peripheral county has been regarded as integrated (1) if 15 percent of the workers living in the county work in the central county of the area, or (2) if 25 percent of those working in the county live in the central county of the area, or (3) if telephone calls from the county to the central county of the area average 4 or more toll calls per subscriber per month.

In New England, where areas have been defined on a town rather than on a county basis, the same principles were applied insofar as they were applicable. Item (c) was not applicable to towns. As a substi