Data quality in demographic surveys: tests and experiments

The urban population projection is an important tool for planning and policy making. There are so many projection methods available, one of the standard methods is the United Nations method. This article aims to examine the best urban population projection method among different methods viz. Modified Exponential curve, Gompertz curve and Logistic growth curve by using the Mean Absolute Percentage Error (MAPE) and Mean Algebraic Percentage Error (MALPE). The estimation of urban population for sixty years from 1961-2011 is computed and later the urban population is projected for next fifty years by using growth models.
The data collected for this study is from different census reports for India, from www.indexmundi.com for China and from www.worldbank.org for the World. The results revels that there is a strong relationship between the forecast error and the bias with growth rate of urban population during the base period for all the three curves. The Gompertz curve has smaller forecast error compared to other curves and it shows underestimate for India and China. The study also reveals that the logistic growth has the smaller forecast error compared to other growth curves and it shows overestimate for World.

One of the Millenium Development Goals (MDG5) is to reduce the maternal mortality ratio in all developing countries by 75% between 1990 and 2015. Progress toward this goal has been documented by the World Health Organization (WHO) and the Institute for Health Metrics and Evaluation (IHME), using similar models to estimate levels and trends in virtually all countries. The Demographic and Health Surveys (DHS) are the principal source of data for the WHO and IHME models. Approximately 120 DHS surveys have included the maternal mortality module, using the direct sisterhood method; all of those surveys will be included. Estimates of adult mortality are also produced from these data, sometimes including information about brothers of the respondent. This will be the first such assessment since 1997, when only 14 surveys had included the module. Several strategies will be used, including comparisons of the reports from related members of the same household and comparisons across successive surveys in the same country. The paper will include recommendations for interpreting the estimates, will discuss potential adjustments (although DHS will continue to publish only unadjusted estimates) and will describe potential implications for the WHO and IHME models.

The quality of the information in household surveys is always of concern to researchers and to policy makers who depend on their results. Although considered the “gold-standard” of household surveys in developing countries, the Demographic and Health Surveys (DHS) are not immune to errors in data, either due to respondent mis-reporting, interviewer mis-recording or deliberate falsification. These errors are thought to be more common in surveys where basic data are not known well, such as that of the dating of events and ages.
In quite a few households, more than one respondent is interviewed and many of these consist of mother-daughter pairs. Indeed, in the 2011 Benin DHS, there were 1488 mother-daughter pairs. A relatively unused technique in assessing data quality is to compare the reports of mothers and daughters. Direct comparison of birth dates, age, and educational attainment are available. When a sibling history is asked, the mother’s birth history information can be compared with the daughter’s sibling history to determine omission of births by either mother or daughter by survival status.

The timely data availability is important for successful public policy development. Increasingly, public and private institutions are aware that data accessibility enables them to identify priority target groups and surveillance activities, as well as allowing researchers to identify risk groups. All democratic processes need access to data and information. In this regard, this paper aims to explore data accessibility on the web sites of the National Statistical Offices (NSOs) in Latin America countries. Based on international quality framework for statistical activities, this paper points out the strengths and weaknesses of accessibility as one of the formal dimensions of data quality. The idea is to examine the accessibility of micro data, metadata, territorial level, periodicity, and the information systems that provide insight into the evolution and recent behavior of population. Twenty NSOs were summated into the research. We highlight the data of demographic research and the improvement of data measurement. The paper starts with a brief description of the international conceptual frame of data quality and ends with the paper’s forward and a note about weaknesses in data access.

KEYWORDS: data accessibility, data quality, micro data, census, National Statistic Office, Latin America.

This paper reports on the results of a study conducted in the Agincourt Health and Demographic Surveillance Site in 2009. A special round,of fieldwork, using fieldworkers who had not been involved in fieldwork of the DSS, was implemented in which census-type questions were asked of respondents (and via them, of other household members). The study had three principal goals: First, to evaluate and measure the errors in reponses to those census-type questions when benchmarked against the longitudinally-collected portraits collected over approximately 15 years, differentially by respondents and respondents reporting on other household members. A second intention of the study was to compare the results from a variety of indirect estimation techniques with the directly-observed estimates reconstructed from the DSS data. Doing so would provide insight into how well the techniques work. The final goal of the project was to find ways of improving those techniques that did not reliably capture the underlying demographc trend so that they might be improved upon.

The current paper will report on the findings related to the first, and (probably) the second, goals outlined above.

Maternal mortality remained as one of the indicators of health and development goal. Many countries lack the estimate of maternal mortality, and in some countries, estimates are not very reliable. In case of India, Sample Registration System (SRS) estimates of maternal mortality based on three year pooled data are considered reliable in comparison to other sample survey. It may be noted that, in many sample surveys, rare events are not studies as it required large sample size. However, if the operation of house-listing can be used meaningfully to identify a rare event like maternal mortality and there is a high probability that all the members of household will remember this event. It may be noted that, around 5 to 20 percent of project expenditure is devoted for this operation, and this operation should be utilized to identify maternal deaths. Number of deaths covered during house-listing operation of any large scale survey will be certainly more than SRS. This will also help to get reliable estimates at the state level especially for those high focused states, where immediate attention is required for maternal care programme in India. The causes of deaths obtained from the nationally representative sample will be much more useful than localized oral autopsy survey.

Background: Almost half of the preschool children are stunted and nearly 60 million children are underweight in India (IIPS & ORG Macro, 2006). The present study investigates the role of child feeding practices and its correlation with child nutritional indicators among Indian children based on data collected during 2005-2006 (NFHS-3, INDIA) using Ruel and Menon’s method with some modification in Indian context on children of 3-36 months of age. Methodology and findings: Data collected on breastfeeding, and complimentary feeding practices were used to construct age specific child feeding index with explanatory variables like, wealth index, birth order, mother’s educational status, and religion using binary and multivariate analysis. Conclusion: There are wide diversity in feeding practices among Indian children by different regions and SES. The proportion of children with better feeding practices are having better growth pattern. The richer, male children from southern region with high educated mothers are less likely to be stunted than their counterpart. There is wide diversity in giving animal products and meal frequencies across different age group. The stunting is strongly associated with child feeding indices especially in children of 6-9 months. The stunting is very high among children with poor feeding practices

Using the 2000 Census of the Philippines Microdata for the Autonomous Region of Muslim Mindanao, we examined the age-sex distribution at the barangay level, the smallest geographical and political unit in the Philippines, to determine data quality. Despite the sluggish economy in this region, and peace and order problem ARMM registered the highest intercensal growth rate in the country. Preliminary analysis showed that such growth was highly irregular and the barangay level age-sex distributions were atypical. In most areas there were more 18 year olds than those in the younger ages. Some areas had no infants, boys, girls, adult men, and elderly population. Analysis of age ratio between an older age group versus younger age group (such as 35-39 vs. 30-34) and/or sex ratio help identify problematic census data.

In sub-Saharan and south Asian countries with incomplete vital registration, adult mortality rates are estimated by asking respondents of household-based surveys (e.g., DHS) about the survival of their maternal siblings. These siblings' survival histories (SSH) are inexpensive to collect but are affected by biases, including sample selection and recall biases. Statistical adjustments have been proposed to correct for sample selection biases, but respondent errors (e.g., forgetting of siblings) are considered inevitable. We test whether a new survey instrument - the siblings enhanced life calendar (SELC) – improves the quality of mortality data relative to standard instruments currently used in DHS surveys. The SELC is based on a life history calendar, a tool widely used in other areas of survey research (e.g., migration histories, sexual behavior research) but not to improve the recall of adult mortality data in developing countries. It also incorporates recall cues developed through cognitive interviews. We report results of a randomized trial of the SELC conducted in a rural population of Senegal (Niakhar) that has been under demographic surveillance (DSS) for the past 50 years. To compare error rates in each arm of the trial, we link DSS data (our “gold standard”) to SSH and SELC data at the individual level.

National Sample Survey (NSS) provides national and sub-national level information on morbidities and health care at regular intervals since its inception on 1953-54. In these surveys, information for all members of sample household was gathered either from head or a key informant of the household. This paper aims to explore the effect of proxy-reporting on population estimates of morbidity prevalence and to identify patterns of biases in these estimates due to proxy responses based on recent 60th round NSS data on morbidity and health care. In this round, information for 68 percent sample population was collected from proxy respondents. Proxies underreported morbidities with an overall morbidity prevalence of 81 per 1000 adult population compared with self-reported morbidity prevalence of 148. Proxy-responses in NSS on morbidity and health care introduce systematic biases, affecting national and regional estimates of morbidity prevalence. Suitable adjustment for proxy-responses should be made while estimating realistic population risks from NSS data.