Ecological fallacies and the analysis of areal census data

The author examines problems related to the fact that in many countries, census data are only reported for areal units and not at the individual level. Attention is paid to the question of ecological fallacy problems that arise from this situation. Data from a 10 percent sample of the United Kingdom population and individual census data from Italy are used to illustrate the problem. "It is concluded that ecological fallacy effects are endemic to areal census data, although their magnitude is perhaps not as large as might have been expected. The principal difficulty is that there is at present no way of predicting in advance the degree of severity likely to be associated with particular variables and particular techniques. Finally, a suggestion is made concerning how the potentially serious practical consequences can be reduced."

A note on the accuracy of count of Kuwaiti population. 1980

"This note shows that a reason for the annual rate of growth of Kuwaiti population between 1975 and 1980...being lower than the projected value...is the improved accuracy of the count of Kuwaiti population of Kuwait, resulting from the requirement of production of proof of Kuwaiti Nationality, insisted [on] at the 1980 Census of Kuwait."

The boundary value problem in spatial statistical analysis

"The primary objective of this paper is to investigate procedures for detecting and handling existing border biasing in spatial statistical analysis." Six conventional solutions to the boundary value problem are criticized, and three alternate statistical solutions are proposed.

Underrecording of infant homicide in the United States

Homicide rates for infants dropped suddenly between 1967 and 1969. The abrupt nature of this decline suggested the change was artifactual. Investigation suggests that two classification revisions instituted at this time were causes of this decline: changes in related codes set forth in the Eighth Revision of the International Classification of Diseases, Adapted, and revision of the standard certificate of death in 1968. Infant homicides may have been disproportionately underrecorded after 1968. (Am J Public Health 1983; 73:195-197.)

[Birth expectations of males and females]

This study is concerned with the extent to which the inclusion of males improves the accuracy of birth expectations data in fertility surveys. "Several studies were compared in order to find an answer to this question. Most surveys, in which both males and females are interviewed, show that very often spouses disagree on the future number of children, depending on the number of children they have already and particularly on the duration of marriage. On an aggregate level, however, the results for males and females show remarkably less disagreement because of averaging-out-mechanisms. Three Dutch studies of 1969, 1975 and 1980 confirm these findings." The general conclusion is that surveys that include responses from women only can be considered highly reliable. (summary in ENG)

[Private households according to socioeconomic classification in national accounting: the model of the Federal Statistical Office]

The author notes that "for calculating income distribution and expenditure by groups of households, annual averages are required concerning the number of private households by household groups and...the structure of the household members in a socio-economic classification." Problems of adjusting such household data so that they are compatible with data from other sources such as employment statistics are discussed, and a procedure for adjusting annual micro-census data on households in the Federal Republic of Germany is described. "Results for the year 1982 are presented, showing the main development trends as compared with the year 1972. Finally, the quality of forecasting with the adjustment procedure is studied." (summary in ENG)

[The estimation of the expectation of life in the case of truncation of the survivorship function at older ages: some considerations]

"A tentative approximation of the expectation of life at 60-65 years, for populations with defective demographic statistics, is explored and expounded on the basis of a recent Horiuchi and [Coale] paper." The method is applied to data for El Salvador, Mexico, Puerto Rico, and Peninsular Malaysia, and it is shown that the method can be used on actual data, although it requires some drastic rounding off. (summary in ENG, FRE)

Some sources of nonsampling errors in African census and survey work: a review

"The paper is a review of some of the more common sources of nonsampling errors in African census and survey work illustrated by some country experiences. [The author suggests] that development and sustainment of an ongoing Integrated System of Household Surveys under the auspices of the United Nations National Household Survey Capability Programme may go a long way toward monitoring and controlling these and other errors that arise in the said census and survey work." (summary in FRE)

The census undercount: effects on federal aid to cities

The impact of adjusting 1980 U.S. population census figures for the undercount on federal grant allocations to cities is examined. In contrast to claims by local officials that their cities have lost up to 200 dollars per uncounted person in federal grants, the authors suggest that an undercount adjustment "could be expected to add little more than [20 dollars] per uncounted person in a few large cities--with most large cities experiencing much smaller gains or even federal aid losses after an adjustment." Several reasons for this finding are discussed. "They relate to the significant limitations of available techniques for estimating the census undercount; the fact that population data are not used in all formula allocation systems; the varied ways in which population data can affect formula allocation systems when they are used; and [the] assumption that federal aid funds would not increase in proportion to the population added as part of an undercount adjustment procedure for federal grants."