Search

Various compromises that have occurred between Bayesian and non-Bayesian methods are reviewed. (A citation is provided that discusses the inevitability of compromises within the Bayesian approach.) One example deals with the masses of elementary particles, but no knowledge of physics will be assumed. KEY WORDS: Bayesians, animals as informal; Fine structure constant, relativistic; Hierarchical Bayes; Inexactification; Maximum Likelihood, type II; P values, standardized., 1. HISTORICAL BACKGROUND Some forms of compromise between Bayesian and non-Bayesian statistics date back perhaps to Laplace, but the concept of such a compromise seems to have been not fully [...]

Chappell (Apeiron 6, 251-253, 1999; in short 6, 251) correctly implies that the last word on record, on a given topic, is often given too much weight. This is especially [...]

When I wrote my 'closing argument' (Ref. 1), I intended to abandon my debate with Xu, but I don't want people to think I'm bonkers. On page 121, column i, [...]

There is one comment by Xu Shaozhi, or Xu for short (Apeiron 6, no. 3-4 which I will abbreviate to 6, 3-4), with which I heartily agree,--intellectual differences about science [...]

Mooney (1999) starts with the two sentences: Newton's Law of Gravity [Gravitation] involves a contradiction. On the one hand it states that gravity is dependent upon the mass of bodies [...]

Xu or Xu & Xu and I have interchanged arguments about the self-consistency of the Lorentz transformation. But it might confuse the reader when we 'speak at the same time': [...]

In this communication, I will again nag Campbell to nobly admit to error without beating about the bush. He claimed (Campbell, 1996) that a 'stationary' rod of length L, pointing [...]

Introduction Readers of Apeiron know that Xu(s) believe(s) the usual 'Lorentz transformation,' LT, to be self-contradictory whereas I am confident that their arguments are fallacious. See the items mentioning Apeiron [...]

1. I now continue my debate with Xu & Xu. See Xu & Xu (1997a, b) or Xu & Xu (a, b) and Good (1997a). They believe that the special [...]

Shaozhi & Xiangqun (1997) produce three arguments to support their expression 'Fairwell Relativity'. I will explain here why their first argument is not entirely fair, and why their other two [...]

Campbell (1996) considers a rod L (or segment) on a long straight line (O', O'), and two observers at O' and O' moving relative to the rod with equal velocities [...]

Walton (1996) says 'The assumption that v' = -v ... is responsible for the paradox that, if we put ct' = ct(1 -v/c) and ct = ct'(1 + v/c); ct [...]

Phipps (1997) asks why do clock rates (in the special theory of relativity, STR) lack group properties? One could reply with the question why should they? But velocities form a [...]

Consider the Lorentz transformation in its familiar form in 2 + 1 dimensions, between the coordinates of an event in an inertial system to those in another such system, (1) [...]

The probability that a spousal murderer was also a batterer is 0.9. This probability is derived from assuming that a representative batterer is one who commits an average of one assault per year. The probability of being murdered is 1/10,000 in the US where the number of murders is 25,000 out of a population of 250,000,000. Around a quarter of the victims are women reducing the probability of being murdered to 1/20,000 for women. The derivation of the probability that a batterer is the murderer is discussed.

In reply to E. Parzen's comment about the difficulty of getting people to use the maximum penalized-likelihood (MPL) method, notice first that a modification of the method, written by Scott, Tapia, and Thompson, is available in the International Mathematical and Statistical Libraries (JMSL). Parzen said there might be difficulty in explaining the MPL method to clients and that they might be suspicious of a method that requires 271 or 3,000 parameters. Clients would have reason to be suspicious if the parameters were all of equal importance, but of course they are not. Parzen's new method of density estimation can be regarded as belonging to exploratory data analysis and in some respects resembles a method suggested by a referee. The author is not yet clear whether the new method is better. Parzen says that further research is required to determine how much smoothing should be done in his method, a problem that corresponds in their method to the choice of the hyperparameter. Presumably he could use classical significance tests in the way they used them. Since he does not evaluate the bumps it is natural that he would need more collaboration with his client.

Describes a Fortran algorithm for computing the Bayes factor F(sub 1) against the hypothesis of independence in a two-dimensional contingency table with r rows and s columns. Numerical theory; Structure; Restrictions; Accuracy and precision; Running time.

Comments on the articles published in The American Statistician journal. Confusion on the figures for the educational status of women in a fertility survey; Suggestion for the order of authorship; Bias on the coverage probability; Impact of statistical analysis on national policy; Ambiguity of statistical formula.

Reports developments on statistical computation in the United States as of May 1983. Generation of TESTRAND, a program for library testing of distributed random number sequences; Assessment on the program called LIMPED, used for estimating the parameters of qualitative and limited dependent variables; Efficacy of tabulation software for survey questionnaire data.

Presents several letters to the editor on concerning statistical issues in the United States. Preparations of a statistician for a career on federal government; Comment on confidence intervals using transparencies; Proposal on normal standard.

Comments on several articles featured in The American Statisticians journal. Applicability of facets on computation variables; Presentation of the Bayesian inference for a binary sequence; Complications in the Bayesian method.

Comments on the articles published in 'The American Statistician' journal. Standard error of difference between means; Policies for publishing statistic research results; Details on the alternative geometric representation of the correlation coefficient.

Presents letters to the editor of the August 1975 issue of the 'American Statistician' journal. Introduction of the other version of Minitab, a statistical computing system; Extensions and generalizations of the celebrated birthday problem; Solution for the Monty Hall problem.

Presents letters addressed to the editor regarding educational statistics. Reliability of the coefficient of consistence; Usability of compartment model as an introduction to the techniques of stochastic processes; Discussion on the simulation of order statistics.

Presents letters to the editor on statistics. Decomposition of moments by conditional moments; Estimation of maximum livelihood; Presentation of the degrees of freedom.

Presents comments on issues related to the use of computers in statistics. Description of a statistical computer program; Programs developed for statistical analysis; Discussion on the interface for statistical data within the computer.

Focuses on the inverse of a centrosymmetric matrix. Development of toeplitz matrices; Algorithm for factorizing a centrosymmetric determinant of odd order.

It is emphasized that the singular decomposition of a matrix has a variety of uses, especially in statistics although it is seldom mentioned in books on either matrices or statistics. Some applications are surveyed and some new ones are given. [ABSTRACT FROM AUTHOR]

Presents letters to the editor of thee American Statistician,' journal. Implementation of a step-wise clustering procedure in the FORTRAN computer program for the determination of optimal separation; Analysis of the distribution of sum of independent integer valued random variables; Solutions of truncated power series distributions.

Comments on the articles published by the American Statistician journal in the United States. Developments in the natural language computing; Availability of a regression algorithm program at the University of California in Los Angeles; Contributions of statisticians to tournament-using sports.

Focuses on the existing problems on statistics in sociology studies in the United States. Complications of sociological systems; Difficulty in organizing experiments; Applicability of statistical methods in dealing with crime data.

Comments on the articles in the 1971 issues of the journal 'The American Statistician'. Discussion on the issue of student participation in affairs of the university; Estimation of the variance of the normal distribution; Details of the general problem of a multi-linear regression.

Presents several letters to the editor of The American Statistician journal published on December 1971. Confusions regarding pairwise zero correlation; Problem of fitting a regression line passing through the origin; Comment on the Work of Wani and Kabe concerning derivation of likelihood ratio.

Presents letters on statistics addressed to the editor of 'The American Statistician' journal in the United States. Simulation studies for comparison of different estimators; Distribution of a sample variance obtained from a two-point population; Presentation of a novel approach to the factorial analysis of marketing experiments.

Presents several correspondence related to statistics in the United States. Derivation of the maximization of R[sup 2]; Proof on the sum of squared errors; Presentation of maximum likelihood estimates.

Examines the randomness of random numbers. Application of random numbers in random sampling and the randomized design of experiments; Capabilities of the Monte Carlo method; Definition of a random sequence.

Presents correspondence concerning statistics and statistical methods. Rationale behind the use of transformations in data analysis; Effect of extrapolation to large estimation error; Feasibility of the Fisher exact test.