Comparing least-squares lines for testing level changes in interrupted time-series analysis

A likelihood ratio test of equality of coefficients of variation or relative errors for two samples of unequal size is presented. The asymptotic distribution is found and the distribution is tabled for small sizes via computer simulation. Kolmogrov-Smirnov confidence intervals are constructed on the percentiles included in the table. Power studies are also simulated and the results indicate that the test has good power for small coefficients of variation     

Small sample power of multivariate nonparametric tests for monotone trend

Bhattacharyya and Kioiz (1966) propose two multivariate nonparametric tests for monotone trend, one involving coordinate-wise Mann statistics and the other, coordinate-wise Spearman statistics. Dietz and Killeen (1981) propose a different test statistic based on coordinate-wise Mann statistics. The Pitman asymptotic relative efficiency of all three tests with respect to a normal theory competitor equals the cube root of the efficiency of a multivariate signed rank test with respect to Hotelling's T2. In this article, the small sample power of the nonparametric tests, the normal theory test, and a Bonferroni approach involving coordinate-wise univariate Mann or Spearman tests is examined in a simulation study. The Mann statistic of Dietz and Killeen and the Spearman statistic of Bhattacharyya and Klotz are found to perform well under both null and alternative hypotheses     

Bootstrap misspecification tests for ARCH based on the empirical process of squared residuals

We propose and study by means of simulations and graphical tools a class of goodness-of-fit tests for ARCH models. The tests are based on the empirical distribution function of squared residuals and smooth (parametric) bootstrap. We examine empirical size and power by means of a simulation study. While the tests have overall correct size, their power strongly depends on the type of alternative and is particularly high when the assumption of Gaussian innovations is violated. As an example, the tests are applied to returns on Foreign Exchange rates.     

Accurate probabilities for the sign test

Limitations in currently available methods for producing significance probabilities for the sign test are discussed. Two simple modifications to the continuity corrected normal approximation are derived and presented in a simple form. These modifications are shown to markedly reduce the relative error in approximating exact probabilities. The relative simplicity of these modifications and the importance of accurate p values suggest that one or both modifications could be useful improvements.     

Cost-optimal performance of special reliability tests using prior knowledge

In the present paper we formulate a general statistical decision problem for reliability investigations and its specification for reliability tests with test period termi¬nated by time and failures,

The risk function includes the experimental costs as well as the expected financial expen¬ses caused by the decisions, Interpreted as     

A consistent correlation-type goodness-of-fit test;with application to the two-parameter weibull distribution

In this article the properties of a general univariate JiT-sample rank tests for complete block designs are investigated. Especially, the asymptotic distribution of the test .statistic under H0 and under contiguous alternatives is derived. Some asymptotic relative'PITMAN efficiencies are computed.

AMSX 1980 subject classifications: Primary 62G10; secondary 62K10     

Sharing Health Services: Two Universities,a Seminary,and a Community

Abstract

“The Inverse Relation Between Fish Consumption and 20-Year Mortality From Coronary Heart Disease,” DAAN KROMHOUT, et al. The low death rate from coronary heart disease among the Greenland Eskimos has been ascribed to their high fish consumption. We therefore decided to investigate the relation between fish consumption and coronary heart disease in a group of men in the town of Zutphen, the Netherlands. Information about the fish consumption of 852 middle-aged men without coronary heart disease was collected in 1960 by a careful dietary history obtained from the participants and their wives. During 20 years of follow-up 78 men died from coronary heart disease. An inverse dose-response relation was observed between fish consumption in 1960 and death from coronary heart disease during 20 years of follow-up. This relation persisted after multiple logistic-regression analyses. Mortality from coronary heart disease was more than 50 per cent lower among those who consumed at least 30g of fish per day than among those who did not eat fish.

We conclude that the consumption of as little as one or two fish dishes per week may be of preventive value in relation to coronary heart disease. (New England Journal of Medicine 1985;312:1205–9.)

“Teaching Touch at Medical School,” JULES OLDER. The association between touch and healing is ancient and worldwide. Skilled hands are among the physician's most important diagnostic and therapeutic tools. Yet a survey of medical schools in the English-speaking world revealed that most offer no touch training in their curricula. Of 169 medical schools, only 12 give any formal instruction in the uses and meaning of therapeutic touch in medicine. The teaching modes of these 12 schools are discussed, and the program of the most touch-conscious school, the University of Otago, is described. (Journal of the American Medical Association 1984;252:931–3.)

“Tubal Infertility and the Intrauterine Device,” DANIEL W. CRAMER, et al. To study the association between intrauterine devices (IUDs) and pelvic inflammatory disease, we compared contraceptive histories in 4185 white women—283 nulliparous women with primary tubal infertility, 69 women with secondary tubal infertility, and 3833 women admitted for delivery at seven collaborating hospitals from 1981 to 1983. The relative risk of tubal infertility associated with IUD use was calculated by means of multivariate logistic regression to control for confounding factors, including region, year of menarche, religion, education, smoking, and reported number of sexual partners. The adjusted risk of primary tubal infertility associated with any IUD use before a first live birth was 2.0 (95 per cent confidence limits, 1.5 to 2.6) relative to nonuse. Users of the Dalkon Shield had an adjusted risk of 3.3 (1.7 to 6.1), users of the Lippes Loop or Saf-T-Coil had a risk of 2.9 (1.7 to 5.2), and users of copper IUDs had a risk of 1.6 (1.1 to 2.4). Women who reported having only one sexual partner had no increased risk of primary tubal infertility associated with IUD use. The adjusted risk of secondary tubal infertility associated with use of a copper IUD after a first live birth was not statistically significant (1.5; 95 per cent confidence limits, 0.8 to 3.0), whereas the risk from similar use of noncopper devices was significant (2.8; 1.3 to 5.9). We conclude that tubal infertility is associated with IUD use, but less so with copper IUDs. (New England Journal of Medicine 1985;312:941–7.)     

On similarity and independence properties of composite edf goodness-of-fit tests

Many test statistics for classical simple goodness-of-fit hypothesis testing problems are distancemeasures between the distribution function of the null hypothesis distributipn and the empirical distribution function sometimes called EDF tests. If a composite parametric null hypothesis is considered in place of the simple null hypothesis, then a test statistic can be obtained from each EDF test by replacing the known distribution function of the simple problem by the Rao-Blackwell estimating distribution function. In this note we use known results to show that these Rao-Blackwell-EDF test statistics have distributions that do not depend upon parameter values, and hence that these tests are independent of a complete sufficient statistic for the parameters.     

Chi-square tests for multivariate normality with application to common stock prices

The theory of chi-square tests with data-dependent cells is applied to provide tests of fit to the family of p-variate normal distributions. The cells are bounded by hyperellipses (x-[Xbar])'S^-1 (x-[Xbar]) = c_i centered at the sample mean [Xbar] and having shape deter-mined by the sample covariance matrix S. The Pearson statistic with these cells is affine-invariant, has a null distribution not depending on the true mean and covariance, and has asymptotic critical points between those of x² (M-1) and x² (M-2) when M cells are employed. The test is insensitive to lack of symmetry, but peakedness, broad shoulders and heavy tails are easily discerned in the cell counts. Multivariate normality of logarithms of relative prices of common stocks, a common assumption in finan-cial markets theory, is studied using the statistic described here and a large data base.     

The bivariate logarithmic series distribution

The bivariate logarithmic series distribution was introduced by Subrahmaniam (1966) as a Fisher-limit to the bivariate negative binomial distribution. The present paper considers the properties of the distribution along with various models giving rise to it. Problems of estimation and the goodness-of-fit are examined. Methods for simulating the distribution are developed and illusuated.