The survey figures indicate less retardation among children in private schools than in public schools at the elementary school and high school levels. A comparison of 1955 and 1950 data shows that there has been a reduction in the amount of retardation during the intervening five years. Although the figures on retardation and acceleration shown in tables F, 5, and 6 indicate the prevailing direction of differences by color, school level, and type of control, the recorded extent of retardation and acceleration would be much greater if the figures were based on data by single years of age and single years of school in which enrolled. Such data are not available from the survey.

ment of Students: October 1955." Statistics on school enrollment for October of years prior to 1955 have been published in the following reports in Series P-20: No. 54 (1954), No. 52 (1953), No. 45 (1952), No. 40 (1951), No. 34 (1950), No. 30 (1949), No. 24 (1948), No. 19 (1947), and No. 1 (1946); and in Series P-S, No. 9 (1945). Enrollment data for April 1947 were published in Series P-20, No. 12. Da ta on educational attainment and illiteracy for persons 14 years old and over in October 1952 were published in Series P-20, No. 45.

The figures for 1950, 1951, and 1952 in this report are revisions of those published in 1950 to 1952 to make them comparable with data for 1953. (See the section on "Revisions in estime ting procedure.") The revised estimating procedure used in 1953 was also used in 1954 and 1955.

1950 Census data.--Statistics on school enrollment and educational attainment for cities, standard metropolitan areas, States, regions, and the United States appear in Volume II of the 1950 Census of Population. Detailed statistics on school enrollment and educational attainment by age and socio-economic characteristics for regions and the United States are presented in a special report of the 1950 Census entitled "Education."

Figures on school enrollment from the October Current Population Surveys differ from 1950 Census da ta for reasons in addition to the difference in the dates. In the first place, the survey data exclude the institutional population and members of the Armed Forces. These two groups were included in the 1950 Census. Second, there were differences in field work. The small group of Current Population Survey enumerators were more experienced and had more intensive training and supervision than the large number of temporary census enumerators and may have more often obtained more accurate answers from respondents. Third, the census was taken in April and relates to enrollment since February 1, whereas the surveys were taken in October and relate to enrollment in the current term. This difference in months of the year affects not only the extent of school enrollment (through "drop outs" during the school year, etc.) but also the level of school in which persons of a given age are enrolled.

Data from school systems.--Information on school enrollment and educational attainment is also collected and published by Federal, State, and local governmental agencies, and by independent research organizations. This informe tion is obtained from reports of school systems and institutions of higher learning, and from other surveys and censuses. These data are only roughly comparable with data collected by the Bureau of the Census by household interviews, however, because of differences in definitions, subject matter covered, time references, and enumeration methods. To illustrate, for comparable grades, the enrollment figures of the Bureau of the Census tend to be lower than those in the Biennial Survey of Education conducted by the United States Office

of Education, largely because the census data refer to shorter time periods and count a person only once, although he may have attended more than one school during the reporting period. In the biennial survey, some persons are included in the enrollment figures more than once, such as those enrolled in both public and private schools and, generally, those enrolled in two different States at any time during the school year. Census data are subject to sampling variability, which may be relatively large for small age groups and for kindergarten and college enrollment.

Perhaps in part because of differences between the methods of collecting enrollment data through interviews and reporting systems, there is a sizable difference between the college enrollment figures from the two sources for 1955. The October 1955 Current Population Survey, conducted by the Bureau of the Census, shows a figure of 2,379,000. The Biennial Survey of Education, conducted by the Office of Education, shows 2,695,340 enrollees in institutions of higher learning in the fall of 1955. It is possible that the college enrollment reported by the Current Population Survey may be smaller than it should have been because of a tendency on the part of enumerators not to obtain information on college enrollment among persons who are employed.

DEFINITIONS AND EXPLANATIONS

Population coverage.--The figures st n for 1945 to 1955 are for the civilian population excluding the relatively small number of inmates of institutions.

Urben and rural residence.--The definition of urban and rural areas which was used in the October 1955 survey was the same as that used in the 1950 Census, but it differed substantially from that used in surveys and censuses before 1950. The territory classified as urban is the same as that in the 1950 Census. According to the definition that was adopted for use in the 1950 Census, the urban population comprises all persons living in (a) places of 2,500 inhabitants or more incorporated as cities, boroughs, and villages; (b) incorporated towns of 2,500 inhabitants or more, except in New England, New York, and Wisconsin, where "towns" are simply minor civil divisions of counties; (c) the densely settled urban fringe, including both incorporated and unincorporated areas, around cities of 50,000 or more; and (d) unincorporated places of 2,500 inhabitants or more outside of any urban fringe. The remaining population is classified as rural.

Farm and nonfarm residence.--The rural population is subdivided into the rural-farm population, which comprises all rural residents living on farms, and the rural-nonfarm population, which comprises the remaining rural population. The method of determining farm and nonfarm residence in the October 1955 Current Population Survey is the same as that

used in the 1950 Census but differs somewhat from that used in census es and surveys before 1950. In the 1950 Census, as in the present survey, persons on "farms" who were paying cash rent for their house and yard only were classified as nonfarm; furthermore, persons in institutions, summer camps, and

tourist courts were classified as nonfarm.

Unmarried persons attending school away from home are enumerated as usually residing in their parents' homes; this fact is especially relevant in the interpretation of data for persons Whose usual residence is on a farm. The 1950 Census enumerated these persons as residents of the communities where the schools and colleges were located.

Age.--The age classification is based on the age of the person at his last birthday.

Color.--The term "color" refers to the division of the population into two groups, white and nonwhite. The nonwhite group includes Negroes, Indians, Japanese, Chinese, and other nonwhite races.

School enrollment.--The school enrollment statistics from the current surveys are based on replies to the enumerator's inquiry as to whether the person had been enrolled at any time during the current term or school year in any type of day or night school, public, parochial, or other private school in the regular school system. Such schools include kindergartens, elementary schools, high schools, colleges, universities, and professional schools. ginning with 1954, children enrolled in kindergarten are included in the enrollment figures for "regular" schools, and are also shown separately. Figures shown in this report on school enrollment for 1950 to 1953 have been revised to include children in kindergarten.

Be

Persons attending "special" schools not in the regular school system, such as trade schools or business colleges, are not included in the enrollment figures but are shown separately. Persons enrolled in classes which do not require physical presence in school, such as correspondence courses or other courses of independent study, and in training courses given directly on the job, are neither reported as enrolled in school nor included in the "special" school category.

Туре of school.--The statistics on type of school indicate the number of persons enrolled at each of four levels: Kindergarten, elementary school (generally first to eighth grades), high school (generally ninth to twelfth grades), and college or professional school. The last group includes graduate students in colleges or universities. Persons enrolled in junior high school through the eighth grade are classified as in elementary school, and the others as in high school.

Public or private school.--In this report, a public school is defined as any educational institution under the legal control of a public body.

SOURCE AND RELIABILITY OF THE ESTIMATES

The estimates presented in this report are based on data obtained in connection with the monthly population sample survey of the Bureau of the Census. The statistics for 1955 are based on a sample design instituted in January 1954. The sample is spread

over

230 sample areas comprising 453 counties and independent cities. A total of 24,000 to 26,000 dwelling units and other living quarters are designeted for the sample at any time, and completed interviews are obtained each month from about 20,000 to 22,000 households. Of the remainder, about 500 to 1,000 are households for which information should obtained but is not, and the rest are vacant dwellings or others that are not to be enumerated for the survey. The data for 1947 through 1953 were based on a different sample, which consisted of about the same number of units but which covered only 68 sample areas.

Revisions in estimating procedure.--The estimating procedure used in the surveys involves, as a final step, the inflation of weighted sample results to independent estimates of the civilian noninstitutional population of the United States. Prior to 1953 these independent estimates were made separately for various age, sex, and veteran status classes, and were based on statistics from the 1940 Census of Population; statistics of births, deaths, immigration, and emigration; and statistics on the strength of the Armed Forces and separation records. Beginning with the 1953 data, the independent estimates were by age, color, and sex and were based on the results of the 1950 Census adjusted for later births, deaths, etc.

Although the introduction of the new population levels should improve the accuracy of the statistics, it creates some discontinuities in the published series. The figures for 1950, 1951, and 1952 which are presented in this report have been revised to make them consistent with the new population estimates for those years. The estimated number of persons enrolled for 1952 on the new basis is 900,000 less than on the former basis; for 1951 and

1950, the difference is 800,000. If comparisons are made with data for these years contained in earlier publications, these differences should be taken into consideration. For years prior to 1950, revisions have not been made; in general, they would be smaller than those for 1950 to 1952. For the entire period from 1945 to 1952, the percentages of persons enrolled for each detailed age-and-sex group based on the sample surveys are not subject to revision; however, the percentages of persons enrolled for combined groups and the absolute numbers on population and school enrollment previously published for these years are subject to revision. The percentages for 1947 to 1949 in table B are subject to very slight revision.

The standard error is a measure of sampling variability. The chances are about 68 out of 100 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 95 out of 100 that the difference is less than twice the standard error and 99 out of 100 that it is less than 2 times the standard error. For the statements in the text of this report, the criterion of twice the standard error (odds of 95 in 100 or better) has generally been used in determining that differences between sample estimates are not likely to have occurred by chance.

The standard error of an estimated percentage depends upon both the size of the percentage and the size of the total on which it is based. The following are examples of the sampling variability of estimated percentages published in this report. The percent of noninstitutional population 5 years old enrolled in kindergarten was estimated at 42.9 percent. The chances are about 68 out of 100 that a complete census would have yielded a figure between 41.2 and 44.6 percent. The percent of civilian noninstitutional population 14 and 15 years of age

living in urban areas enrolled in school was estimated at 97.5. The chances are about 68 out of 100 that a complete census would have yielded a figure between 97.0 and 98.0 percent.

The standard error of an absolute estimate based on the current survey depends not only on the size of the estimate but also on the distribution of the estimate by the age, color, and sex control groups. The following are examples of the sampling variability of absolute estimates. The number of persons 5 to 34 years of age enrolled in school in 1955 was estimated at 37,426,000. The chances are about 68 out of 100 that a complete census would have yielded a figure between 37,136,000 and 37,725,000. The number of persons 14 and 15 years of age enrolled in school was estimated at 4,516,000. The chances are about 68 out of 100 that a complete census would have yielded a figure between 4,492,000 and 4,539,000.

The standard errors shown above are not directly applicable to differences obtained by subtracting one figure from another. The standard error of an observed difference between two estimates depends upon the standard error of each of the estimates and the correlation between them.

In addition to sampling variation, the estimates are subject to biases due to errors of response and to nonreporting, but the possible effect of such biases is not included in the above measures of reliability.

In general, the individual figures in this report are rounded to the nearest thousand without being adjusted to group totals, which are independently rounded.