Latent mixture modeling for clustered data

Statistics and Computing - This article proposes a mixture modeling approach to estimating cluster-wise conditional distributions in clustered (grouped) data. We adapt the mixture-of-experts model...     

Inequality and education decisions in developing countries

In this paper we analyze the effect of inequality on school enrollment, preferred tax rate and expenditure per student in developing countries; when parents can choose between child labor, public schooling or private schooling. We present a model in which parents make schooling decisions for their children, weighing the utility benefit of having a child with formal public or private education versus the forgone income from child labor or household work. Parents vote over the preferred tax rate to finance freely provided public education. The utility benefit of an educated child is proportional to expenditure per student, so that there is congestion in public school. We find that when parents can send their children to work or to private school, high inequality leads to exit from public education at both ends of the income distribution. Thus high inequality reduces the support for public education, leading to a low tax rate and expenditure per student. Exit from public education results in both high child labor and a large fraction of students attending private school. In fact there is a threshold level of inequality above which there is no longer support for public education. In addition we explore the implications for the design of foreign aid. The results suggest that foreign aid policies should focus on promoting school attendance rather than increasing school resources, as the later policy might be offset by a reduction in the recipient country’s fiscal effort, with little impact on outcomes.      

Aging in rural Japan--limitations in the current social care policy

Owing to equal and increased opportunities for education and employment, today's trend in Japanese marriages is characterized by late and less frequent marriage. This paper discusses unavoidable diversity in rural families to point out the anticipated consequences of aging in rural areas and to discuss limitations in current public social care policies. Specifically, the averaged proportion of never-married and single persons at ages 45 to 49 and 50 to 54 in legally recognized depopulated cities, towns, and villages in Japan is calculated to illustrate the expected diversity in families in rural depopulated areas. It also illustrates the need for future studies to develop better social care policies for increasing numbers of single caregivers and single elders.     

Some families of group divisible designs

Generalizing methods of constructions of Hadamard group divisible designs due to Bush (1979), some new families of semi-regular or regular group divisible designs are produced. Furthermore, new nonisomorphic solutions for some known group divisible designs are given, together with useful group divisible designs not listed in Clatworthy (1973).     

Simulated removal trapping considered in connection with actual census data in small mammals

Summary A record of simulated removal trapping of 50 samples drawn fromN=50 under a design has been considered as to how it does contribute to practical censuses of small mammals. It follows that, for ordinary-density populations so far experienced, the removal census based on the theory may frequently lead to quite wrong estimates from single-plot data by the simplest method, and that the maximum likelihood method is doubtless better, but even this method may produce right estimates in about half of the cases, their standard error being of limited availability. Therefore pooling of data from many plots are desirous for such populations.     

Adaptation and validation of the Japanese version of the Gambling Urge Scale

In Japan, accurately assessing and targeting gambling urges is important for the treatment of gambling disorder because people are constantly exposed to external triggers that elicit gambling urges. However, a valid measure that assesses gambling urges has not been adequately established. Therefore, the purpose of this study was to examine the psychometric properties of the Japanese version of the Gambling Urge Scale (GUS-J), which is a well-known valid 6-item instrument in English-speaking countries. Two hundred and thirty-nine participants completed questionnaires, 86 of which met the criteria for a probable gambling disorder. As with the original GUS, the GUS-J was demonstrated to be a one-factor model that accounted for 63.82% of the total item variance, with all items demonstrating loadings of .40 or higher. With regard to reliability, the GUS-J showed good internal consistency (α = .88). Furthermore, validity based on evidence of relationships with other variables, especially convergent and concurrent validity of the GUS-J, were supported by several analyses that examined hypotheses based on previous findings. In conclusion, this study showed that the GUS-J has good psychometric properties as an assessment tool for gambling urges in Japanese people and it will be useful for future clinical practice and research.     

Variation in offspring size within a population of the web-building spiderAgelena limbata

Within a population of the web-building spiderAgelena limbata, the weight of the first instar nymphs ranged from 1.187 to 6.559 mg. Both intraclutch and interclutch variation were recorded. The mean weights were different among clutches and the coefficients of variation within a clutch ranged from 3.3 to 29.2%. Variation in the nymphal weight was certainly derived from variation in the egg weight because there was a high correlation between the two weights. Factors affecting interclutch variation in nymphal weight were examined by multiple regression analysis. Nymphal weight was positively correlated with the body size and food conditions of female parents, and negatively correlated with the clutch size. Among these three factors, the food conditions of female parents had the largest apparent effect on the interclutch variation. The results suggest that females with larger body size and more food produce larger offspring, and that there is a trade-off between offspring size and clutch size. Heavier nymphs had larger body size (carapace width) and may have larger energy reserves. Heavier nymphs survived experimental starvation for a significantly longer period.     

Evolution and physiological consequences of de-alation in crickets

Wing shedding or de-alation is a common phenomenon among crickets. Its significance and effects on other traits were examined based on the results from experiments using artifical or natural de-alation. Artificial de-alation at adult emergence induces rapid egg production and flight muscle histolysis in several species examined. However, natural de-alation does not always shorten the pre-ovipositional period because it does not occur immediately after adult emergence and because oviposition starts before de-alation. In some cases, naturally de-alated females produce more eggs than to intact females during early adult life, but peak ovipositing activity occurs before de-alation. Therefore, retention of the hindwings does not suppress high ovipositing activity in such cases. It appears that de-alation is a result rather than a causal factor in ending migration. Ovarian development and flight muscle histolysis, which can be stimulated by de-alation, are controlled by the jevenile hormone, but the mechanism inducing de-alation remains unknown. The possible factors leading to the evolution of de-alation are discussed.     

How is life history variation generated from the genetic resource allocation?

A simple quantitative genetic model is proposed to explain the observed genetic correlation structure of a bruchid beetleCallosobruchus chinensis in terms of two underlying variables: the resource acquisition and the resource allocation. Heritabilities and genetic correlations among age-specific, fecundities are regarded as consequences of genetic variations of the two variables. Genetic correlations are predominantly positive in both predictions and observations. Nonetheless, comparison between observed and predicted values in heritabilities, genetic correlations, and genetic principal components suggested significant genetic variances both of the resource allocation and the resource acquisition. The prediction of the model is discussed in relation, to experimental tests of trade-off in life history evolution.