At the end of 2005 Biometric Society of Japan (BSJ) cerebrated a quarter of a century of its establishment, a 25-year period over which the number of BSJ members has increased to about 450 from 220 and it has engaged a wide range of research and education. In this report, I reviewed the major activities of BSJ and the achievements that BSJ members have made by investigating the temporal trend of published papers in such journals as Japanese Journal of Biometrics, Biometrics and Statistics in Medicine. I further investigated the Japanese achievements published in the Encyclopedia of Statistical Sciences and the Cambridge Dictionary of Statistics. Based on these investigations, I discussed the key conditions that, I think, could improve the present situation and infuence the future activities and status of Japanese Biostatisticians.
Contemporary evolutionary biology has used various statistical methods for collecting and analyzing data. Here methods for estimating phylogenetic trees are reviewed in the context of recent history of evolutionary biology, especially of systematics and phylogenetics. Estimating evolutionary history based on character data (molecular or morphological) poses a couple of epistemological problems all of which are common to historical sciences in general. Karl Popper's philosophy of science, in particular, his theory of falsification and corroboration has been espoused by many theoretical phylogeneticists. In this paper the long-standing controversy on philosophical aspects of phylogenetics and its implications for statistical methods in this discipline is discussed.
This paper describes several methods for estimating essential parameters in fishery resource management and conservation, which we developed so far. First, an unbiased estimator for evaluating total landings and a ratio estimator for measuring contribution of hatchery releases were derived based on a two-stage random sampling and applied to a masu salmon stock enhancement program. Second, the efficacy of an integrated likelihood approach to estimate Wright's FST in a metapopulation was demonstrated using numerical simulations and real data analysis. Third, an empirical Bayes procedure for estimating genetic parameters is explained, which can simulate posterior distributions for any parametric functions of allelic frequencies. As an example, the effective population size of the Northern pike was estimated. Finally, a method for simultaneous estimation of mixing proportion and genetic drift based on series of genetic tagging data was presented with application to the data from a stock enhancement program of the mud crab.
In this paper, I will review recent development in epidemiologic theories. It is emphasized that case-control studies are considered to be based on a underlying hypothesized cohort. Thus, various control selection options have been developed as the selection from the population at risk in that hypothesized cohort. As an example of the use of new epidemiologic designs, a case-control study of infant asthma is illustrated. It is a part of the SORA (Study on Respiratory Disease and Automobile Exhaust) project which is a complex of epidemiologic studies for infants, school children, and adults conducted by the Ministry of the Environment. In that study, a two-stage case-control design which is effcient both for rare exposure and rare disease is adopted.
Recently, risk assessment attracts increasingly the attention of academic researchers as well as of various stakeholders in our society. Biostatistical approaches are essential in quantitative risk assessment of human environmental factors. We review recent development of quantitative risk assessment, and discuss the need of further researches. Our interests will be placed on the contrast between an individual and a society. The two aspects of the contrast are: that between toxic responses of an individual and of individuals in a population, and that between judgements for private safety and for public safety.