Comparing the U.S. Decennial Census Coverage Estimates for Children from Demographic Analysis and Coverage Measurement Surveys

Following every U.S. decennial census since 1960, the U.S. Census Bureau has evaluated the completeness of coverage using two different methods. Demographic analysis (DA) compares the census counts to a set of independent population estimates to infer coverage differences by age, sex, and race. The survey-based approach (also called dual system estimation or DSE) provides coverage estimates based on matching data from a post-enumeration survey to census records. This paper reviews the fundamentals of the two methodological approaches and then initially examines the results of these two methods for the 2010 decennial census in terms of consistency and inconsistency for age groups. The authors find that the two methods produce relatively consistent results for all age groups, except for young children. Consequently, the paper focuses on the results for children. Results of the 1990, 2000, and 2010 decennial censuses are shown for the overall population in this age group and by demographic detail (age, race, and Hispanic origin). Among children, the DA and DSE results are most inconsistent for the population aged 0–4 and most consistent for ages 10–17. Results also show that DA and DSE are more consistent for Black than non-Black populations. The authors discuss possible explanations for the differences in the two methods for young children and conclude that the DSE approach may underestimate the net undercount of young children due to correlation bias.     

The US Decennial Census: Political Questions,Scientific Answers

The US decennial census was initiated in 1790 to facilitate nation‐building tasks, especially that of reconfiguring political representation as the population grew and settled new territories. To this basic task of power distribution have been added other key governmental functions, such as the use of census data in guiding revenue sharing and in the enforcement of nondiscriminatory policies. Throughout its history the census has been the focus of partisan clashes. Following the identification of the “differential undercount” a measure of how census coverage differs across demographic groups and geographic areas–the partisan battles intensified, and in recent decades have come to focus not just on how the census counts are used but how the census data are collected. It has been argued that census methodology could be designed to predetermine given partisan outcomes, and for the 2000 census this charge shifted from “could be” to “is.” The Census Bureau has taken extraordinary steps to demonstrate that no partisan considerations have affected the design or implementation of the census, and that its decisions are based solely on the best technical judgment available.     

Demographic comparison between self-response and personal visit interview in Census 2000

During Census 2000, over 95 percent of the housing units were in mailback areas. In these areas either the United States Postal Service or Census Bureau staff delivered the census questionnaire. In urban and suburban areas of the country, the United States Postal Service delivered the census questionnaires between March 13 and 15 of 2000. The addresses in these areas are predominately city style; house number and street name. In more rural areas of the country with predominately non-city style addresses, Census Bureau staff delivered the questionnaire during March of 2000. Respondents completed their census questionnaires and returned them through the mail. Returns from these housing units are classified as self-response. Respondents which did not complete and return their census questionnaire by April 18, 2000 were interviewed during the nonresponse followup operation (personal visit interview). This paper will examine the demographic characteristics of persons enumerated on the mail return questionnaire (self-response) and the persons enumerated during the nonresponse followup operation (personal visit interview).     

AN INVESTIGATION OF GENERALIZED RAKINGIN THE SYNTHETIC ESTIMATION OFPOPULATION SIZE

The problem of counting a population that is cross-classified with respect to demographic and geographic attributes is considered. A census is conducted in which individuals are “captured” with probabilities that are believed to be relatively constant within demographic categories. The census is followed by a random sample in which individuals are “recaptured” independently of the census. Using the two counts, capture-recapture estimates of the demographic category populations are obtained. A synthetic estimate of population size for a geographic entity is obtained by summing the corresponding adjustment factors (capture-recapture estimates divided by census counts) across all individuals captured by the census in the entity. The use of generalized raking is considered as a method for smoothing adjustment factors. It is found that generalized raking differs little from a class of weighted least squares regression models. This suggests that generalized raking does not offer an improvement over regression for smoothing adjustment factors. The efficiency loss of generalized raking relative to the best regression-based procedures can be substantial.     

Evaluating the relative accuracy and significance of net migration estimates

This paper presents the elements of a theory for evaluating the quality of a set of net migration estimates. The total error in a net migration estimate is decomposed into total bias and total variation. The bias is further decomposed into three bias elements-selection bias, estimator bias, and measurement bias. Tables of bounds for measurement and estimator biases in the vital statistics and the forward survival ratio estimates of net intercensal migration are presented. Both net migration levels and net migration ratios are treated, and provision is made for both life table and census survival ratio estimates. Some of the statistical tables are applied illustratively to net migration data for Canadian counties or census divisions during the 1951-61 decade. Tests of significance and confidence intervals are indicated for net migration estimates, and the basic technical notes are presented in appendices.     

An Evaluation of Persons per Household (PPH) Estimates Generated by the American Community Survey: A Demographic Perspective

The American Community Survey (ACS) is a U.S. Census Bureau product designed to provide accurate and timely demographic and economic indicators on an annual basis for both large and small geographic areas within the United States. Operational plans call for ACS to serve not only as a substitute for the decennial census long-form, but as a means of providing annual data at the national, state, county, and subcounty levels. In addition to being highly ambitious, this approach represents a major change in how data are collected and interpreted. Two of the major questions facing the ACS are its functionality and usability. This paper explores the latter of these two questions by examining “persons per household (PPH),” a variable of high interest to demographers and others preparing regular post-censal population estimates. The data used in this exploration are taken from 18 of the counties that formed the set of 1999 ACS test sites. The examination proceeds by first comparing 1-year ACS PPH estimates to Census 2010 PPH values along with extrapolated estimates generated using a geometric model based on PPH change between the 1990 and 2000 census counts. Both sets of estimates are then compared to annual 2001–2009 PPH interpolated estimates generated by a geometric model based on PPH from the 2000 census to the 2010 census. The ACS PPH estimates represent what could be called the “statistical perspective” because variations in the estimates of specific variables over time and space are viewed largely by statisticians with an eye toward sample error. The model-based PPH estimates represent a “demographic perspective” because PPH estimates are largely viewed by demographers as varying systematically and changing relatively slowly over time, an orientation stemming from theory and empirical evidence that PPH estimates respond to demographic and related determinants. The comparisons suggest that the ACS PPH estimates exhibit too much “noisy” variation for a given area over time to be usable by demographers and others preparing post-censal population estimates. These findings should be confirmed through further analysis and suggestions are provided for the directions this research could take. We conclude by noting that the statistical and demographic perspectives are not incompatible and that one of the aims of our paper is to encourage the U.S. Census Bureau to consider ways to improve the usability of the 1-year ACS PPH estimates.     

What Data from the 2010 Census Tell Us about the Changing Child Population of the United States

This article provides an overview of changes in the U.S. child population (persons under age 18) based on data released from the 2010 census. Today, the number of children in the United States (74.2 million) is at an all-time high, but the share of the national population who are children (24 %) is at an all-time low. The number of children in the population grew by 1.9 million between 2000 and 2010, but the overall national figure masks many important details and divergent paths. The growing racial and ethnic diversity in the U.S. is more advanced among children than among adults. Some areas of the country and some demographic groups grew significantly over the decade, while the number of children in other areas and in other groups fell.     

Methods for the analysis of mortality risks across heterogeneous small populations: examination of space-time gradients in cancer mortality in North Carolina counties 1970–75

A method of analyzing mortality rates in heterogeneous populations is presented. This method, appropriate for the investigation of mortality rates in small geographic areas (e.g., counties) where the forces of mobility operate to selectively “package” persons, is applied to the determination of whether a spatial west-east gradient in cancer mortality rates existed in North Carolina over the period 1970 to 1975. A significant gradient (as well as a significant temporal trend) is determined to exist in the data, though only for particular race, age and sex-specific demographic groups. Several alternate hypotheses are presented to explain the existence of the spatial gradient in these particular demographic groups.     

The use of sampling in conjunction with population censuses

In Asia and the Pacific, the practice of governments is increasingly to collect information on their populations' size, age and sex composition, geographic distribution, and certain other basic demographic and socioeconomic characteristics on the basis of a complete (100%) enumeration, and to supplement this basic information by collecting information on a larger range of variables on a sample basis. The additional information, which is gathered as part of the census operation, may be related to such population characteristics as migration, employment, fertility, and health. A complete census is generally indispensable for obtaining information about small domains, and also for obtaining politically important data, which must be seen to be free from sampling variability. A complete census is typically confined to obtaining a detailed picture of the number and basic structural characteristics of the entire population, with as much detail as possible about local areas. Sample surveys can quickly obtain a wider variety of more complex data. Sampling applications for the census include 1) using sampling in the design and control of census operations, such as in planning, testing, controlling, and evaluating the census; 2) using sample enumeration to supplement items covered in the complete census; 3) sampling the census results for processing to make the results available more quickly and at lower cost; and 4) extracting samples of microlevel files of detailed census data so as to facilitate dissemination of primary data.     

Double-sampling regression-cum-exponential estimator of the mean of a sensitive variable

ABSTRACT

A flexible scrambled response model using a randomization device for quantitative sensitive data is used to evaluate the protection of respondents’ privacy. A double-sampling regression-cum-exponential estimator is used to estimate the mean of a sensitive variable using the mean of a nonsensitive auxiliary variable under scrambled response. The expected bias, the expected mean square error, and the minimum mean square error of this exponential-type estimator are expressed. Simulations and empirical results show that the proposed estimator under scrambled response model has a lower mean square error and a lower bias than the ratio and the exponential estimators.     

Thailand’s urban population reconsidered

In lieu of any official definition of “urban population” in Thailand, researchers and policy makers have generally used a simple urban-rural dichotomy, relying upon the administratively designated municipal areas as urban and the remainder of the kingdom as rural. For administrative purposes, however, a more refined set of residence categories is available: the “urban” segment is subdivided into three categories of municipal areas; within the “rural” group, units of population concentration classified as sanitary districts are distinguished from the remainder of the rural population. Data from a 1970 census sample tape have been used to analyze a number of socioeconomic characteristics, as well as migration and fertility, for each of these five residence categories separately in an attempt to determine if the formerly used rural-urban dichotomy is valid. For each of the characteristics and for migration and fertility, a clear urban-rural continuum emerges, with the sanitary districts generally resembling the smaller urban places more than they do the rural areas under which they are usually subsumed. The evaluation thus suggests the importance to research and policy either of using an urban-rural continuum or of grouping sanitary districts with municipal areas when a dichotomy must be used. Doing so should facilitate evaluation of the interrelations between urbanization and demographic and development factors.     

Establishing trends in atsic regional council populations using census data: A cautionary note

This paper points out that limitations in official census data for Aborigines and Torres Strait Islanders draw into question the validity of trend analysis based on time series data for Aboriginal and Torres Strait Islander Commission (ATSIC) regional council areas. Accordingly, the meaningful application of projection techniques to estimate future population profiles using existing census data is severely restricted. Among the difficulties encountered in reconstructing council area populations are census boundary changes over time, changes in enumeration techniques and coverage, the problems posed by self-identification and associated population growth, and, in some cases, the difficulty of matching ATSIC regional council boundaries with census geography. Following discussion of these problems, detailed figures showing changes in the size of the Aboriginal and Islander populations and labour force in each council area are presented using 1976 as the base year. As expected, geographic patterns of population and labour-force change are difficult to discern and exact reasons for comparative growth or decline are impossible to determine. The paper concludes that reverse projections for regional council areas using 1991 Census data would provide a more reliable basis for establishing demographic trends. Although not entirely adequate, these reconstructions for ATSIC regional councils are the only estimates of these populations that have been undertaken to date.     

A demographic history of the Indo-Dutch population, 1930–2001

Indonesia was a Dutch colony until 1949. In the aftermath of World War II and the independence of the former Dutch East Indies, many people migrated from Indonesia to the Netherlands or other Western countries. This migrant population, known as the Indo-Dutch population, consists of Europeans, Asians, and persons of mixed European-Asian blood. These groups have all associated themselves with and experienced the colonial culture of the former Dutch Indies, and have carried this cultural experience elsewhere through migration. This paper provides a demographic history of the Indo-Dutch population, using a variety of data sources and methods. Starting from the population of ‘Europeans’ according to the 1930 census of the Dutch Indies, a demographic projection is made covering the period 1930–2001. By the beginning of 2001, the estimated number of Indo-Dutch persons is 582,000, including the second generation. Of these 582,000, an estimated 458,000 are living in the Netherlands and 124,000 elsewhere. The composition by age, sex and generation very clearly reflects the demographic history of the population.     

International Migration and City Growth in the Global South: An Analysis of IPUMS Data for Seven Countries, 1992–2013

Evidence on the demographic components of city growth in the global South is scarce, and the role played by international migration is neglected. We analyze the importance of recent international migration in cities, compare it with that of internal movements, and evaluate the growth contribution across national contexts and the urban hierarchy. Combining individual-level census data and geographic master files of metropolitan areas with indirect demographic estimation techniques, we cover 377 cities in seven countries. It is found that, in almost one third of cities, population change and replacement has been mainly determined by migration. The international component was larger than the internal one in more than half of cities. Whereas internal migration tends to decrease with rising city size, international movements tend to increase. Positive net international migration substitutes for the net losses from domestic movements in large cities, but complements the gains in intermediate-sized cities.     

Some New Techniques for Applying the Housing Unit Method of Local Population Estimation: Further Evidence

The housing unit method of population estimation is often characterized as being imprecise and having an upward bias. In an earlier paper we argued that the method itself cannot be properly characterized by a particular level of precision or direction of bias. Only specific techniques of applying the method can have such characteristics. In that paper we presented several new techniques for estimating the number of households and average number of persons per household (PPH). However, the testing of these new techniques was limited by the lack of census results against which the estimates could be compared. Complete census data on population, households, and PPH are now available and can be used to test alternate estimation techniques. In this paper we replicate the tests reported in our earlier paper using 1980 census data for Florida’s 67 counties. These tests provide further evidence that the new techniques produce more precise, less biased estimates than previously used techniques.     

Mapping the diversity of gender preferences and sex imbalances in Indonesia in 2010

Indonesia is usually viewed as a country free of the acute forms of gender discrimination observed elsewhere in East or South Asia, a situation often ascribed to Indonesia's bilateral kinship system. I re-examine this hypothesis by focusing on ethnic and regional variations in sex differentials. New indicators of marriage practices and gender bias derived from 2010 census microdata highlight the presence of patrilocal patterns as well as a distinct presence of son preference in fertility behaviour in many parts of the archipelago. I also present evidence for excessive child sex ratios and excess mortality of females in some areas that appear to be related to son preference and patrilocal residence systems. The findings confirm the association between son preference, sex differentials in mortality, prenatal sex selection, and kinship systems. I conclude with a more regional perspective on demographic vulnerability of females, distinguishing bilateral South East Asia from more patrilineal Melanesia.     

The growing American Indian population, 1960–1990: Beyond demography

Since 1960, the Native American population has exhibited explosive growth, increasing from 552,000 to 1,959,000, or 255 percent. The average annual growth rate of 4.3 percent, extending over a 30-year period, is demographically impossible without immigration – in fact, of the 1.4 million growth only 762,000 comes from natural increase, whereas 645,000 comes from non-demographic factors. This paper expands on previous work to illustrate with demographic techniques how such extraordinary growth was achieved through changing patterns of racial self-identification on the part of people with only partial or distant American Indian ancestry, coupled with relatively high fertility and improving mortality. It also provides some basic demographic background on the size, growth, and geographic structure of the American Indian population, while exploring both demographic and geographic underpinnings of the changing population. Data on race from the 1970, 1980, and 1990 Censuses are based on self-identification. A person choosing the American Indian racial response does not have to provide any substantiation or documentation. While this method aids overall census taking by permitting respondents to fill out their own census forms, self-identification adds a temporal component to the data because responses elicited from the same individual (or group of individuals) may change over time in reaction to social, political, or economic conditions or variations in question wording. New identities may emerge or old ones may disappear. This effect has had a noteworthy impact on data for American Indians. Analysts and other data users must be aware of underlying response patterns to interpret changes correctly.     

Missing Minorities? The Phases of IRCA Legislation and Relative Net Undercounts of the 1990 vis-à-vis 2000 Decennial Census for Foreign-born Cohorts

The quality of the decennial census of the United States is compromised by population undercount, which often misses immigrants and racial/ethnic minorities, thereby diminishing federal resources allocated to such groups. Using a modified version of demographic analysis and informed by the latest contributions of emigration scholarship, this research estimates net undercount for the 1990 census relative to the 2000 census by age, sex, year-of-entry, and place-of-birth cohorts. Ordinary least squares estimates suggest that males, recent arrivals, and cohorts aged 15–44 had higher relative net undercount for 1990 compared with 2000. Much higher relative net undercount was found for cohorts from Africa, Latin America, and the Caribbean (excluding Cuba and Puerto Rico) who were ineligible for amnesty under the Immigration Reform and Control Act of 1986 (i.e., those fitting the profile of an undocumented immigrant). Larger implications of these findings suggest that the political climate in which a person is embedded—particularly for persons who may feel threatened or marginalized by the government and/or the public—affects that person’s willingness to respond to the census.     

Group-based differences in intra-generational earnings mobility in Israel

High earnings inequality, by itself, is less of a problem if a society is characterized by high earnings mobility rates. Using the matched 1983–1995 census file created by the Israeli Central Bureau of Statistics, which includes data for individuals who were included in both the 1983 and 1995 demographic samples of the census, this study focuses on the differential mobility rates of the various gender–ethnic groups in the Israeli labor market, as well as on the determinants of upward earnings mobility. The findings indicate that rising inequality between the dominant and subordinate groups slows down the impact of high rates of earnings mobility in Israel. Specifically, native-born Ashkenazi men (the most advantaged group) have a firm hold on the high earnings positions. Furthermore, the data show that the earnings gap between Ashkenazi men and most other groups has increased over time. Although the other groups, particularly women, also experienced some upward mobility, this mobility took place across the lower quintiles, where earnings are low.     

Female Farm Operators: Attitudes About Social Support in Their Retirement Years

SUMMARY

In a traditionally patriarchal occupation, women farm operators may be particularly at risk of inadequate planning and support in later years. This study examined the acceptability of a set of in-home and live-in social support options potentially available to 151 women farm operators. For the married women, help from a spouse was most acceptable. The only other option rated as very acceptable by many of the women was having a child come by to help. The study also examined factors differentiating between women who indicated a particular social support option as acceptable and those who did not. Few of the demographic or attitude variables distinguished between the groups. The discussion concludes with recommendations for future research and with the practice implications of the findings.