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.     

Making the Housing Unit Method Work: An Evaluation of 2010 Population Estimates in Florida

The housing unit (HU) method—in which population estimates are derived from estimates of occupied HUs—is the most commonly used method for making small-area population estimates in the United States. It is widely used because it is conceptually simple, can utilize a wide variety of data sources, can be applied at virtually any level of geography, and often produces reliable estimates. Yet the HU method is more nearly a general approach to population estimation than it is a specific methodology. In this paper, we describe and evaluate several data sources and estimation techniques that can be used in applying the HU method. Using a set of county and subcounty estimates produced by the Bureau of Economic and Business Research (BEBR) at the University of Florida for April 1, 2010, we analyze errors by population size and growth rate, calculate errors for each of the three components of the HU method, and investigate the accuracy of estimates based on several different data sources and estimation techniques. We compare the accuracy of the 2010 estimates with previous BEBR estimates and estimates produced by the U.S. Census Bureau. We conclude that although some data sources and estimation techniques work better than others, the HU method can be tailored to produce reliable population estimates for a wide variety of geographic areas.     

An Evaluation of Population Estimates in Florida: April 1, 2000

The housing unit (HU) method is the most commonly used method for making small-area population estimates in the United States. These estimates are used for a wide variety of budgeting, planning, and analytical purposes. Given their importance, periodic evaluations of their accuracy are essential. In this article, we evaluate the accuracy of a set of HU population estimates for counties and subcounty areas in Florida, as of April 1, 2000. We investigate the influence of differences in population size and growth rate on estimation errors; compare the accuracy of several alternative techniques for estimating each of the major components of the HU method; compare the accuracy of 2000 estimates with that of estimates produced in 1980 and 1990; compare the accuracy of HU population estimates with that of estimates derived from other estimation methods; consider the role of professional judgment and the use of averaging in the construction of population estimates; and explore the impact of controlling one set of estimates to another. Our results confirm a number of findings that have been reported before and provide empirical evidence on several issues that have received little attention in the literature. We conclude with several observations regarding future directions in population estimation research.     

Report on a multiple regression method for making population estimates

A new method of making population estimates is introduced and contrasted to the standard ratio-correlation method. The new method, called the “difference-correlation method,” produced population estimates for the counties of Michigan which had a smaller mean percentage error than estimates produced by the ratio-correlation method. The differences of proportions which are used in the difference-correlation method will always have means of zero, while the ratios used in the ratio-correlation method have means which vary. Higher intercorrelations among the variables and increased temporal stability of the intercorrelations are two advantages of using differences rather than ratios.     

Death by survey: Estimating adult mortality without selection bias from sibling survival data

The widely used methods for estimating adult mortality rates from sample survey responses about the survival of siblings, parents, spouses, and others depend crucially on an assumption that, as we demonstrate, does not hold in real data. We show that when this assumption is violated so that the mortality rate varies with sibship size, mortality estimates can be massively biased. By using insights from work on the statistical analysis of selection bias, survey weighting, and extrapolation problems, we propose a new and relatively simple method of recovering the mortality rate with both greatly reduced potential for bias and increased clarity about the source of necessary assumptions.     

How does POSSLQ measure up? Historical estimates of cohabitation

We use March Current Population Survey (CPS) data from 1977 to 1997 to produce a new historical series of indirect cohabitation prevalence estimates. We compare our new estimates with those produced by the traditional method and evaluate the new estimates. We then compare the indirect estimates with the new direct estimates to investigate whether biases exist in the indirect estimates. Our findings indicate that the traditional indirect method of estimating cohabitation prevalence underestimates cohabitors in different subpopulations, especially among those with children. We also find that the new indirect measure produces relatively unbiased estimates of cohabitors' characteristics.     

Estimating the Expectation of Life at Old Ages: Comments on the Article by Mitra

Family reconstitution studies assess demographic behaviour for that section of the population that remains in the parish of their birth. Although there has been considerable discussion of whether or not these non-migrants are representative of the population as a whole, the distorting effects of migration censoring have not been adequately considered. This paper shows that even when the demographic behaviour of non-migrants is identical to that of the general population, migration-censoring will significantly bias family reconstitution results. In particular, existing family reconstitution studies underestimate both age at marriage and life expectancy. The paper assesses the potential magnitude of bias introduced by migration, and suggests several new techniques for correcting the errors.     

Analyzing fertility histories: Do restrictions bias results?

The straightforward tests we have conducted lead to two major conclusions. First, parameter estimates, such as the proportions that practice contraception or that breastfeed, can be biased in data restricted to the last closed and open interval. This is particularly true the further back in time one goes. However, the second conclusion is that these restrictions do not bias estimates of the structure of the relationships predicting fertility. This may seem surprising, and perhaps even magical. The reason is that multivariate life table techniques allow one to reach the same conclusion even if the proportions in various categories are altered by a criterion such as limiting the analysis to intervals begun by the last and next-to-last live births. Limiting the analysis in this way means that there are fewer short intervals and thus fewer cases of intervals with characteristics associated with short intervals (e.g., no contraceptive use, no breastfeeding, or infant mortality). As long as the model specified in the multivariate life table is an appropriate one, that is, it is not misspecified, and as long as the skew produced by the WFS restriction is not too extreme, then the multivariate life table procedures can produce unbiased estimates of the structure of the relationships predicting birth interval dynamics. Thus even though the WFS data are in fact inappropriate for some simple parameter estimation procedures, they appear to be adequate for the more complex multivariate procedures of the sort used here. Several caveats must be added to the foregoing results. First, we have performed this test in only one country, Korea; it is possible that the same results might not be obtained in other countries. We expect, however, that they would. Second, our procedure only looks at the first 40 months of experience in the birth interval. A procedure that incorporates the long tails of the birth interval distribution may obtain different results. In fact, we caution against analyzing the tail of the distributions using data from the normal WFS sample, since these would be most affected by the restriction to last closed and open intervals. Third, the extent to which these results are generalizable to other types of substantive problems is unknown at present. We suspect, however, that examining the determinants of lengths of breastfeeding will produce similar results. Finally, even with multivariate procedures, it would be highly misleading to impose the WFS restrictions and then examine trends in the length of birth intervals.(ABSTRACT TRUNCATED AT 400 WORDS)     

Has Income Segregation Really Increased? Bias and Bias Correction in Sample-Based Segregation Estimates

Several recent studies have concluded that residential segregation by income in the United States has increased in the decades since 1970, including a significant increase after 2000. Income segregation measures, however, are biased upward when based on sample data. This is a potential concern because the sampling rate of the American Community Survey (ACS)—from which post-2000 income segregation estimates are constructed—was lower than that of the earlier decennial censuses. Thus, the apparent increase in income segregation post-2000 may simply reflect larger upward bias in the estimates from the ACS, and the estimated trend may therefore be inaccurate. In this study, we first derive formulas describing the approximate sampling bias in two measures of segregation. Next, using Monte Carlo simulations, we show that the bias-corrected estimators eliminate virtually all of the bias in segregation estimates in most cases of practical interest, although the correction fails to eliminate bias in some cases when the population is unevenly distributed among geographic units and the average within-unit samples are very small. We then use the bias-corrected estimators to produce unbiased estimates of the trends in income segregation over the last four decades in large U.S. metropolitan areas. Using these corrected estimates, we replicate the central analyses in four prior studies on income segregation. We find that the primary conclusions from these studies remain unchanged, although the true increase in income segregation among families after 2000 was only half as large as that reported in earlier work. Despite this revision, our replications confirm that income segregation has increased sharply in recent decades among families with children and that income inequality is a strong and consistent predictor of income segregation.     

The Network Survival Method for Estimating Adult Mortality: Evidence From a Survey Experiment in Rwanda

Adult death rates are a critical indicator of population health and well-being. Wealthy countries have high-quality vital registration systems, but poor countries lack this infrastructure and must rely on estimates that are often problematic. In this article, we introduce the network survival method, a new approach for estimating adult death rates. We derive the precise conditions under which it produces consistent and unbiased estimates. Further, we develop an analytical framework for sensitivity analysis. To assess the performance of the network survival method in a realistic setting, we conducted a nationally representative survey experiment in Rwanda (n = 4,669). Network survival estimates were similar to estimates from other methods, even though the network survival estimates were made with substantially smaller samples and are based entirely on data from Rwanda, with no need for model life tables or pooling of data from other countries. Our analytic results demonstrate that the network survival method has attractive properties, and our empirical results show that this method can be used in countries where reliable estimates of adult death rates are sorely needed.     

An Application of Bayesian Methods to Small Area Poverty Rate Estimates

Efforts to estimate various sociodemographic variables in small geographic areas are proving difficult with the replacement of the Census long-form with the American Community Survey (ACS). Researchers interested in subnational demographic processes have previously relied on Census 2000 long-form data products in order to answer research questions. ACS data products promise to begin providing up-to-date profiles of the nation’s population and economy; however, unit- and item-level nonresponse in the ACS have left researchers with gaps in subnational coverage resulting in unstable and unreliable estimates for basic demographic measures. Borrowing information from neighboring areas and across time with a spatiotemporal smoothing process based on Bayesian statistical methods, it is possible to generate more stable and accurate estimates of rates for geographic areas not represented in the ACS. This research evaluates this spatiotemporal smoothing process in its ability to derive estimates of poverty rates at the county level for the contiguous United States. These estimates are then compared to more traditional estimates produced by the US Census Bureau, and comparisons between the two methods of estimation are carried out to evaluate the practical application of this smoothing method. Our findings suggest that by using available data from the ACS only, we are able to recreate temporal and spatial patterns of poverty in US counties even in years where data are sparse. Results show that the Bayesian methodology strongly agrees with the estimates produced by the SAIPE program, even in years with little data. This methodology can be expanded to other demographic and socioeconomic data with ease.     

Changes in cohort wealth over a generation

Empirical computation of expected wealth is hampered by two problems: mortality risks vary in the population and over time; and observation of net estates for most cohorts is truncated, as some individuals in a cohort survive the calendar date on which observation is terminated. These two problems are solved in estimating cohort wealth for a sample of Wisconsin taxpayers. Hazard rate models of differential occupational mortality risks were estimated from the occupational information on the tax records. Values of net estate are simulated for individuals in each birth cohort who survived. Survivors have characteristics that imply greater wealth holdings than the deceased in every birth year covered by the study (1890-1924). Because of this, estimates of wealth-age relationships produced by the estate multiplier method for any given year will have a serious downward bias. Longitudinal data imply that dissaving does not occur after age 65.     

The effect of child support enforcement on child support payments

This paper examines the effect of state child support enforcement legislation on child support received from absent fathers by ever-married women due support in 1978 or 1981. The analysis is based upon data from the 1979 and 1982 March/April Match files of the Current Population Survey, two nationally-representative surveys of the eligible child support population, combined with a data set assembled by the authors on child support enforcement techniques available in each state. Based upon probit estimates and OLS estimates corrected for sample selection bias, we find that expedited processes and liens (against real and personal property), as well as wage withholding laws in effect for at least three years, increased the amount of child support received in 1981. In general, enforcement is more effective at increasing the amount received than the probability of receiving something, and is more effective for Blacks than for nonBlacks.     

Revising Infant Mortality Rates for the Early Twentieth Century United States

Accurate vital statistics are required to understand the evolution of racial disparities in infant health and the causes of rapid secular decline in infant mortality during the early twentieth century. Unfortunately, U.S. infant mortality rates prior to 1950 suffer from an upward bias stemming from a severe underregistration of births. At one extreme, African American births in southern states went unregistered at the rate of 15 % to 25 %. In this study, we construct improved estimates of births and infant mortality in the United States for 1915–1940 using recently released complete count decennial census microdata combined with the counts of infant deaths from published sources. We check the veracity of our estimates with a major birth registration study completed in conjunction with the 1940 decennial census and find that the largest adjustments occur in states with less-complete birth registration systems. An additional advantage of our census-based estimation method is the extension backward of the birth and infant mortality series for years prior to published estimates of registered births, enabling previously impossible comparisons and estimations. Finally, we show that underregistration can bias effect estimates even in a panel setting with specifications that include location fixed effects and place-specific linear time trends.     

Empirical bayes estimation of demographic schedules for small areas

In this article, we analyze empirical Bayes (EB) methods for estimating small-area rate schedules. We develop EB methods that treat schedules as vectors and use adaptive neighborhoods to keep estimates appropriately local. This method estimates demographic rates for local subpopulations by borrowing strength not only from similar individuals elsewhere but also from other groups in the same area and from regularities in schedules across locations. EB is substantially better than standard methods when rates have strong spatial and age patterns. We illustrate this method with estimates of age-specific fertility schedules for over 3,800 Brazilian municipalities.     

New techniques in small area population estimates by demographic characteristics

The increasing demand for small area population estimates calls for both innovative ways of using existing data and new techniques suitable for small area estimates. This paper explores the methods for population estimates by age, sex, race, and Hispanic origin at the census tract level for Multnomah County, Oregon. New techniques include employing building permits to indirectly estimate migration and examining the changes in age/sex structure using the American Community Survey (ACS). A practical method for bridging the race categories is also developed. Finally, the paper discusses some reflections on small area estimates and the potentials of using ACS to track the changes of the demographic characteristics for the sub-county level.     

Effect of provider characteristics on choice of contraceptive provider: A two-equation full-information maximum-likelihood estimation

We use surveys of households and health-care facilities conducted in the same area at the same time to determine which characteristics of providers attract users of contraceptives. By using the full-information maximum-likelihood technique to jointly estimate choice of contraceptive method and choice of provider, we avoid self-selection bias. Results support the need for modeling quality and for jointly estimating the choice of contraceptive method and the choice of provider to avoid biased estimates of coefficients. The results suggest that for the Cebu, Philippines region, small local clinics that focus on family planning tend to be most favored by clients.     

Sampling Weights for Analyses of Couple Data: Example of the Demographic and Health Surveys

In some surveys, women and men are interviewed separately in selected households, allowing matching of partner information and analyses of couples. Although individual sampling weights exist for men and women, sampling weights specific for couples are rarely derived. We present a method of estimating appropriate weights for couples that extends methods currently used in the Demographic and Health Surveys (DHS) for individual weights. To see how results vary, we analyze 1912 estimates (means; proportions; linear regression; and simple and multinomial logistic regression coefficients, and their standard errors) with couple data in each of 11 DHS surveys in which the couple weight could be derived. We used two measures of bias: absolute percentage difference from the value estimated with the couple weight and ratio of the absolute difference to the standard error using the couple weight. The latter shows greater bias for means and proportions, whereas the former and a combination of both measures show greater bias for regression coefficients. Comparing results using couple weights with published results using women’s weights for a logistic regression of couple contraceptive use in Turkey, we found that 6 of 27 coefficients had a bias above 5 %. On the other hand, a simulation of varying response rates (27 simulations) showed that median percentage bias in a logistic regression was less than 3 % for 17 of 18 coefficients. Two proxy couple weights that can be calculated in all DHS surveys perform considerably better than either male or female weights. We recommend that a couple weight be calculated and made available with couple data from such surveys.     

A Turning Point in Gender Bias in Mortality? An Update on the Number of Missing Women

Amartya Sen started a debate about gender bias in mortality by estimating the number of “missing women,” which refers to the number of females of any age who have presumably died as a result of discriminatory treatment. Depending on the assumptions made, the combined estimates for countries exhibiting the presence of such gender bias varied between 60 and 107 million. As new population data have become available for these countries, this article examines whether the number of “missing women” has changed in the past decade. The combined estimate of the number of missing women has risen in absolute terms but has fallen slightly in relation to overall population. Considerable improvement is evident in West Asia, North Africa, and parts of South Asia, while only small improvements have occurred in India and a deterioration took place in China. Analyses of the underlying causes of gender bias in mortality suggest that improvements are largely related to improved female education and employment opportunities and rising overall incomes, while deterioration is mostly attributable to the rising incidence of sex‐selective abortions.     

Modelling Marital Fertility by Age and Duration: An Empirical Appraisal of the Page Model

Page's model of marital fertility by age and duration is fitted by maximum likelihood techniques to data from 38 of the 41 developing countries that participated in the World Fertility Survey. The results indicate that the model does an excellent job of capturing variations in fertility patterns, with only two parameters. Moreover, national-level estimates of the parameter representing the degree of control of marital fertility correlate reasonably well with the proportion using contraception. On the other hand, estimates of the parameter representing the level of natural fertility correlate well with the duration of breastfeeding and with a measure of contraceptive use for spacing, but also show substantial regional variation. The paper closes with comments on several extensions and applications of the model.