Assessing post-census state poverty estimates

While the decennial census provides poverty figures for states and other subnational geographic units every ten years, their utility declines over the course of a decade. Consequently, there is growing interest in producing post-census estimates for a variety of indicators. This study extends recent efforts to estimate post-census poverty figures for states by producing such estimates using a multiple regression approach. The accuracy of the multiple regression estimates along with recently produced estimates from the Current Population Survey (CPS) are evaluated relative to the decennial census. The mean absolute percentage point error (MAPPE) using the ratio-correlation technique (1.56 percentage points) was somewhat higher than the MAPPE of 1989 CPS (1.37 percentage points) and an average of 1988–1990 CPS data (1.15 percentage points). However, a simple regression technique using data from 1979 to estimate poverty in 1989 produced a set of estimates where the MAPPE (1.37 percentage points) is nearly as accurate as the single-year CPS estimates. Estimates which average regression estimates and CPS-based estimates are more accurate than either regression or CPS estimates used alone. Several suggestions are offered for improving regression estimates.This article is a revised version of a paper presented at the annual meeting of the Population Association of America, 1–4 April 1993, Cincinnati, OH, USA.     

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.     

The estimation of adult mortality in Africa from data on orphanhood

Summary Until recently, very little information has been available about the levels and patterns of adult mortality in tropical Africa, but during the past decade several countries have included questions in censuses and surveys as to whether a person's father and mother are still alive. From the data so obtained, estimates of adult mortality have been prepared. This paper compares the results of three such exercises with alternative estimates of adult mortality derived from other sources. In the case of Chad, the orphanhood data obtained in the demographic sample survey of 1964 yielded estimates of mortality which agreed reasonably closely with those obtained from questions on deaths of household members occurring during the twelve months preceding the survey. The latter data however were themselves subject to substantial errors and had to be corrected using techniques based on stable population theory. For Kenya, the orphanhood questions were included in the 1969 census and the results were compared with the mortality estimates derived from inter-censal survival from 1962 to 1969. Once again, the data obtained from the latter were subject to error but in general appeared to be consistent with the orphanhood estimates. The third comparison was made from Malawi, where alternative mortality figures were available from the Malawi Population Change survey which was a 'dual record' type of operation, conducted in 1971/2. The agreement in this case was remarkably close, once the number of deaths had been corrected for omissions by both systems with allowance for positive correlation. It is concluded that as a simple and inexpensive technique of estimating adult mortality, the orphanhood approach has much to recommend it.     

Measuring Inequality of Opportunity for the Backward Communities: Regional Evidence from the Indian Labour Market

The affirmative action policy for socially and economically backward communities in employment has been a debated issue in India. In this context, this paper aims to analyze the level of inequality by distinguishing between ‘circumstance’ and ‘effort’ factors in the Roemer’s framework on equality of opportunity. We measure inequality of opportunities due to two circumstances: caste and religion. Our empirical analysis, at state-level, utilizes a recent household survey data, which provides information related to efforts as well as circumstances of workers. The paper estimated inequality in the labour market and then decomposed it to know the circumstances that cause income inequality. Our estimates indicated that inequality and inequality of opportunity is substantially higher in India. Specifically, the outcome of our analysis evidently indicated that the socially backward communities do have economically disadvantageous position in some of the Indian states. However, the degree of circumstances based on inequality varies to a great extent among the states. Therefore, we suggest that the country does not need a nation-level affirmative action policy instead a state-level policy could be more appropriate as the intensity of the problem differ significantly among the Indian states.     

How much Confidence can we have in EU-SILC? Complex Sample Designs and the Standard Error of the Europe 2020 Poverty Indicators

If estimates are based on samples, they should be accompanied by appropriate standard errors and confidence intervals. This is true for scientific research in general, and is even more important if estimates are used to inform and evaluate policy measures such as those aimed at attaining the Europe 2020 poverty reduction target. In this article I pay explicit attention to the calculation of standard errors and confidence intervals, with an application to the European Union Statistics on Income and Living Conditions (EU-SILC). The estimation of accurate standard errors requires among others good documentation and proper sample design variables in the dataset. However, this information is not always available. Therefore, I complement the existing documentation on the sample design of EU-SILC and test the effect on estimated standard errors of various simplifying assumptions with regard to the sample design. It is shown that accounting for clustering within households is of paramount importance. Although this results in many cases in a good approximation of the standard error, taking as much as possible account of the entire sample design generally leads to more accurate estimates, even if sample design variables are partially lacking. The effect is illustrated for the official Europe 2020 indicators of poverty and social exclusion and for all European countries included in the EU-SILC 2008 dataset. The findings are not only relevant for EU-SILC users, but also for users of other surveys on income and living conditions which lack accurate sample design variables.     

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 method for combining sample survey data and symptomatic indicators to obtain population estimates for local areas

Past efforts of statistical demographers to compute postcensal population estimates for local units have been hampered by the fact that they have had to rely completely on symptomatic information. In this paper, a new method of postcensal estimation is presented in which the symptomatic information is combined with sample data by means of a regression format. Combining symptomatic information on births, deaths, and school enrollment with sample data from the Current Population Survey, county estimates of population growth were computed by means of the new method for the postcensal period after 1960. These estimates were tested for accuracy by comparison with a set of special censuses which were conducted between 1964 and 1967 in 75 counties. The results of this test are promising, but not conclusive. A more conclusive test is currently underway using 1970 data. While the method has been tested as a means for estimating population growth, it is to be emphasized that it can be used to compute postcensal estimates for any variable for which the necessary symptomatic information and sample data are available.     

A state-based regression model for estimating substate life expectancy

Life expectancy is an important indicator of the level of mortality in a population. However, the conventional way of calculating life expectancy--constructing a life table--has rigorous data requirements. As a consequence, life expectancy data are not usually available for substate areas. In this article, a regression model for estimating life expectancy is constructed, using state-level data, and is tested against two sets of 1980 life expectancy data: (1) a nationwide sample of metropolitan areas and (2) selected cities, their suburbs, and rural counties in Ohio. An additional test shows the sensitivity of the model's accuracy to errors in one of its input data elements. The results suggest that the model should be given serious consideration for generating life expectancy estimates for substate areas.     

Importance Weighting in Client Satisfaction Measures: Lessons from the Life Satisfaction Literature

The linearisation approach to approximating variance of complex non-linear statistics is a well-established procedure. The basis of this approach is to reduce non-linear statistics to a linear form, justified on the basis of asymptotic properties of large populations and samples. For diverse cross-sectional measures of inequality such linearised forms are available, though the derivations involved can be complex. Replication methods based on repeated resampling of the parent sample provide an alternative approach to variance estimation of complex statistics from complex samples. These procedures can be computationally demanding but tend to be straightforward technically. Perhaps the simplest and the best established among these is the Jackknife Repeated Replication (JRR) method. Recently the JRR method has been shown to produce comparable variance for cross-sectional poverty measures (Verma and Betti in J Appl Stat 38(8):1549–1576, 2011); and it has also been extended to estimate the variance of longitudinal poverty measures for which Taylor approximation is not currently available, or at least cannot be easily derived. This paper extends the JRR methodology further to the estimation of variance of differences and averages of inequality measures. It illustrates the application of JRR methodology using data from four waves of the EU-SILC for Spain. For cross-sectional measures design effect can be decomposed into the effect of clustering and stratification, and that of weighting under both methodologies. For differences and averages of these poverty measures JRR method is applied to compute variance and three separate components of the design effect—effect of clustering and stratification, effect of weighting, and an additional effect due to correlation of different cross-sections from panel data—combining these the overall design effect can be estimated.     

Using modified cohort change and child-woman ratios in the Hamilton–Perry forecasting method

The Hamilton–Perry method, which uses cohort change ratios (CCR) and child-woman ratios (CWR), has gained acceptance as research has demonstrated its practical value and accuracy in forecasting population composition. Assessments of this method have been based on the usual assumption that CCRs and CWRs developed over the base period are held constant over the forecast horizon. We propose several approaches for modifying CCRs and CWRs over the forecast horizon. These alternatives are averaging and trending these ratios and a synthetic method that bases local CCRs and CWRs changes on state-level changes in CCRs and CWRs. We evaluate the errors for these alternatives against the errors holding the CCRs and CWRs constant for counties in Washington State and for census tracts in New Mexico. The evaluation suggests that averaging or trending CCRs and CWRs are not worthwhile strategies, but the synthetic method reduces errors compared to holding the ratios constant over the horizon.     

Frequency of geographic mobility: Findings from the National Health Interview Survey

Age stands out as an important factor in multiple moves over a relatively short period of time. In addition, in the NHIS data whites show a higher tendency to be multiple movers than do blacks. It was also found that persons defined as interstate migrants moved more often than either intrastate migrants or local movers, but it is impossible to conclude anything more definitive about that relationship. In conclusion, mobility defined as a move rather than as a mover is clearly higher than estimates derived from five-year census data or the CPS, which only use the idea of a mover. On the whole, though, over three-year periods most persons do not move at all or they move and then return. Of those who do change residence, the majority move only once. By far the bulk of moves is accounted for by a relatively small segment of Americans. Of 34,000 moves reported in the NHIS, more than one-half were made by a small number of persons--4,400--who not only constituted scarcely one-fourth of all movers but were less than 1 out of every 11 Americans.     

The impact of incomplete geocoding on small area population estimates

Small-area population estimates are often made using geocoded address data in conjunction with the housing-unit method. Previous research, however, suggests that these data are subject to systematic incompleteness that biases estimates of race, ethnicity, and other important demographic characteristics. This incompleteness is driven largely by an inability to complete georeference address-based datasets. Given these challenges, small-area demographers need further, and to date largely unavailable, information on the amount of error typically introduced by using incompletely geocoded data to estimate population. More specifically, we argue that applied demographers should like to know if these errors are statistically significant, spatially patterned, or systematically related to specific population characteristics. This paper evaluates the impact of incomplete geocoding on accuracy in small-area population estimates, using a Vintage 2000 set of block-group estimates of the household population for the Albuquerque, NM metro area. Precise estimates of the impact of incomplete geocoding on the accuracy of estimates are made, associations with specific demographic characteristics are considered, and a simple potential remediation based on Horvitz-Thompson theory is presented. The implications of these results for the practice of applied demography are reviewed.     

Self-reports of child maltreatment in the U.S.: a key social indicator

A key social indicator of the well being of a society is the health and welfare of their children. Child maltreatment is a major problem in the U.S. and the world and the reporting of maltreatment has been the subject of much research and debate. However, little is known about self-reports of child maltreatment. Children face many obstacles that may prevent them from reporting their own maltreatment to authorities. Despite these obstacles, a small percentage of all child maltreatment reports made to child protective services (CPS) in the US are made by the victim. The purpose of this paper is to describe an analysis of reports made to CPS by the child-victim of maltreatment over a three-year period. Results indicated significant differences related to child race, gender and substantiation rate among self-reports and reports made by others. The data utilized in this publication were made available by the National Data Archive on Child Abuse and Neglect, Cornell University, Ithaca NY; and have been used by permission. Data from the National Child Abuse and Neglect Data System were supplied by state child protective service agencies to the Children's Bureau or the Administration of Children, Youth and Families, U.S. Department of Health and Human Services. Funding for NCANDS was provided by the Children's Bureau, U.S. Department of Health and Human Services. The National Child Abuse and Neglect Data System was implemented by Walter R. McDonald & Associates, Inc. Neither the participating state agencies; Walter R. McDonald & Associates, Inc.; the Children's Bureau, Administration on Children, Youth and Families, U.S. Department of Health and Human Services, the Archive, Cornell University, or its agents or employees bear any responsibility for the analyses, opinions, or interpretations presented here.     

A flexible two-dimensional mortality model for use in indirect estimation

Mortality estimates for many populations are derived using model life tables, which describe typical age patterns of human mortality. We propose a new system of model life tables as a means of improving the quality and transparency of such estimates. A flexible two-dimensional model was fitted to a collection of life tables from the Human Mortality Database. The model can be used to estimate full life tables given one or two pieces of information: child mortality only, or child and adult mortality. Using life tables from a variety of sources, we have compared the performance of new and old methods. The new model outperforms the Coale-Demeny and UN model life tables. Estimation errors are similar to those produced by the modified Brass logit procedure. The proposed model is better suited to the practical needs of mortality estimation, since both input parameters are continuous yet the second one is optional.     

MAPE-R: a rescaled measure of accuracy for cross-sectional subnational population forecasts

Accurately measuring a population and its attributes at past, present, and future points in time has been of great interest to demographers. Within discussions of forecast accuracy, demographers have often been criticized for their inaccurate prognostications of the future. Discussions of methods and data are usually at the centre of these criticisms, along with suggestions for providing an idea of forecast uncertainty. The measures used to evaluate the accuracy of forecasts also have received attention and while accuracy is not the only criterion advocated for evaluating demographic forecasts, it is generally acknowledged to be the most important. In this paper, we continue the discussion of measures of forecast accuracy by concentrating on a rescaled version of a measure that is arguably the one used most often in evaluating cross-sectional, subnational forecasts, Mean Absolute Percent Error (MAPE). The rescaled version, MAPE-R, has not had the benefit of a major empirical test, which is the central focus of this paper. We do this by comparing 10-year population forecasts for U.S. counties to 2000 census counts. We find that the MAPE-R offers a significantly more meaningful representation of average error than MAPE in the presence of substantial outlying errors, and we provide guidelines for its implementation.     

Estimation of interprovincial migration for Canada from place of birth by residence data, 1951–1961

In Canada, unlike many other countries, birth-residence data by age and sex are available in each of the decennial censuses from 1931 to 1961 which permit the estimation of intercensal net migration for the provinces and regions. After a brief discussion of the basic measures of migration from birth-residence data the paper focusses on the problems and procedures in estimating interprovincial net migration, 1951–1961 for Canada using “the place of birth survival ratio method, ” and it evaluates the estimates thus obtained. The evaluation of the estimates, taking into consideration the inherent limitations of the method and its merits compared with period migration estimates by the census survival ratio method and life table survival ratio method, suggests that the net migration estimates for the Canadian born by the place of birth survival ratio method are probably more reliable than those by the other two methods. One striking finding was that the net migration curves by age obtained from the census survival ratio and place of birth survival ratio estimates were smoother than the curve obtained with the use of the more accurate life table survival ratios. Furthermore, whatever the relative accuracy of net migration may be, the birth-residence approach is capable of furnishing more details about the net migration of the native born than by the standard survival-ratio methods. For the population under age 10 intercensal estimates were directly derived from the place of birth and residence distributions by age.     

A comparison of three maximum likelihood models for stage-frequency data

Summary A comparison has been made between the estimates obtained from maximum likelihood estimation of gamma, inverse normal, and normal distribution models for stage-frequency data. Results have been compared for six of sets of test data, and from many sets of simulated data. It is concluded that (1) some estimates may differ substantially between the models, (2) estimates from the correct model have little bias, and estimated standard errors are generally close to theoretical values, (3) there are problems in determining degrees of freedom for chi-squared goodness of fit tests, so that it is best to compare test statistics with simulated distributions, and (4) goodness of fit tests may not discriminate well between the three models.     

The Canadian sample for labour force and other population data

An area sample is now being taken quarterly in Canada for labour force and a wide variety of other population data. Since reliable lists of households cannot be obtained without substantial expense the amount of field work is reduced to a minimum by the use of maps which may be thought of as an implicit listing. Random selection with known probabilities is applied in several stages, the overall result being a 1 % sample of the population. The survey data provide not only estimates of unemployment, movement of population, etc., but also give the errors of estimate. During the 3 years of its operation, information has been obtained for government and public use on housing conditions, rents, farm fires and accidents, sickness, fish consumption and many other matters.