Testing hypotheses when there could be an initial effect

In many treatment-versus-control experiments, the observed random variables can be written as the product of a Bernoulli and a continuous random variable. The treatment can affect the distribution of the observations in two ways.

1. the probability that the observation is 0 could be altered.

2. the distribution of the nonzero observations could be changed.

We may also want to measure the combined effect of the treatment.

3. the expected value of control and treated units may differ.

A method is presented for testing for the presence of the combined effect when the general form of the distribution function of the continuous observations is known. For the case when this distribution function is from the family of gamma distributions, a previously proposed test criterion for the combined effect has poor power properties. In this paper, we discuss a test criterion that has improved power properties.     

On the use of the Box-Cox transformation in a linear regression with a lagged dependent variable

In this paper, we show a maximum likelihood estimation procedure in the Box-Cox model when a lagged dependent variable is included among explanatory variables and the first observation of the dependent variable is random. It is shown in a numerical example that a test of a coefficientof the lagged dependent variable is sensitive to whether the first observation of the dependentvariable is random or not.     

CORRECTING FOR MEASUREMENT ERROR IN AN EXPOSURE-RESPONSE RELATIONSHIP BASED ON DICHOTOMISING A CONTINUOUS DEPENDENT VARIABLE

Random error in a continuous outcome variable does not affect its regression on a predictor. However, when a continuous outcome variable is dichotomised, random measurement error results in a flatter exposure-response relationship with a higher intercept. Although this consequence is similar to the effect of misclassification in a binary outcome variable, it cannot be corrected using techniques appropriate for binary data. Conditional distributions of the measurements of the continuous outcome variable can be corrected if the reliability coefficient of the measurements can be estimated. An unbiased estimate of the exposure-response relationship is then easily calculated. This procedure is demonstrated using data on the relationship between smoking and the development of airway obstruction.     

Characterizations of exponentiality within the HNBUE class and related tests

A harmonic new better than used in expectation (HNBUE) variable is a random variable which is dominated by an exponential distribution in the convex stochastic order. We use a recently obtained condition on stochastic equality under convex domination to derive characterizations of the exponential distribution and bounds for HNBUE variables based on the mean values of the order statistics of the variable. We apply the results to generate discrepancy measures to test if a random variable is exponential against the alternative that is HNBUE, but not exponential.     

A modified Kolmogorov-Smirnov test for the inverse gaussian density with unknown parameters

The Kolmogorov-Smirnov (KS) test is an empirical distribution function (EDF) based goodness-of-fit test that requires the underlying hypothesized density to be continuous and completely specified. When the parameters are unknown and must be estimated from the data, standard tables of the KS test statistic are not valid. Approximate upper tail percentage points of the KS statistic for the inverse Gaussian (IG) distribution with unknown parameters are tabled in this paper.

A study of the power of the KS test for the IG distribution indicates that the test is able todiscriminate between the IG distribution and distributions such as the uniform and exponentialdistributions that are very different in shape, but is relatively unable to discriminate between the IG distribution and distributions that are similar in shape such as the lognormal and Weibull distributions. In modeling settings the former distinction is typically more important to make than the latter distinction.     

Computational procedures for deriving sampling distributions of attribute-contaminated random variables

Statistical distributions generated from any J- or U-shaped random variables are cumbersome to derive if not completely indefinable and thus are unavailable analytically because of the singularities at the tails of the basic random variable. This paper presents a computational method for providing a numerical convolution derived from a basic U-shaped random variable composed of a continuous part mixed with (or contaminated by) a discrete part at the tails. The J-shaped sampling distribution case is implied as a special case. Though the computations are based on a background Normal Distribution, it can be generalized on any other distribution.Such distributions will open up an area of sampling distributions of mixed random variables that are not elaborately covered in textbooks dealing with the theory of distributions.     

Uniform convergence of an empirical cdf based on a relatively small number of order statistics

An empirical distribution function F_m, defined on a subset of order statistics of a random sample of size n taken from the distribution of a random variable with continuous distribution function F, is shown to converge uniformly with probability one to F. Small sample distributions of the one and two sided deviations and the asymptotic normality of the standardized F_m are established. The relative efficiency of F_m as compared to the classical empirical distribution function is calculated and tabled. for n = 10, 20, 50, 100, 200.     

An algorithm to test divisibility of continuous distributions,applied to the uniform distribution

We present an algorithm to test if a continuous random variable is n-divisible. Through the algorithm, testing for divisibility is put in the format of numerical analysis amenable to computer processing. This is illustrated with the uniform distribution which is known to be non-divisible. New evidence and new proofs arising out of the insight developed with this style are given. This algorithm, along with the analytical methods should provide much needed versatility to tackle problems in this area. The fact that Bondesson [1987] works on the divisibility of the Half Cauchy Distribution gives evidence that there are many distributions which can use the combination of the analytical and the algorithmic tools to settle the divisibility issue.     

Power of the Neyman Smooth Test for Evaluating Multivariate Forecast Densities

We compare and investigate Neyman's smooth test, its components, and the Kolmogorov-Smirnov (KS) goodness-of-fit test for testing the uniformity of multivariate forecast densities. Simulations indicate that the KS test lacks power when the forecast distributions are misspecified, especially for correlated sequences of random variables. Neyman's smooth test and its components work well in samples of size typically available, although there sometimes are size distortions. The components provide directed diagnosis regarding the kind of departure from the null. For illustration, the tests are applied to forecast densities obtained from a bivariate threshold model fitted to high-frequency financial data.     

Proportional hazards model for discrete data: Some new developments

ABSTRACT

Many times, a product lifetime can be described through a non negative integer valued random variable. In this article, we propose a proportional hazards model for discrete data analogous to the version for continuous data. Some ageing properties of the model are discussed. Stochastic comparison of pair of random variables that follow the model are also made. A new test based on U-statistics is developed for testing that the proportionality parameter in the proposed model is 1. The asymptotic properties of the proposed test are studied. We present some numerical results to asses the performance of the test procedure.     

Expectation of a uniform random variable with uniform observation errors after selection of the highest observations

In selection processes of a random variable with random observation errors, the crucial variable is the conditional expectation of the target variable given the sum of the observations. An example is the selection of the most talented young researchers for tenure track. This paper derives an explicit expression for this conditional expectation for the case that both the target variable and the observation errors have a uniform, but different, distribution.     

An Efficient Simulation Method for the Computation of a Class of Conditional Expectations

Suppose that the random vector X and the random variable Y are jointly continuous. Also suppose that an observation x of X can be easily simulated and that the probability density function of Y conditional on X = x is known. The paper presents an efficient simulation-based algorithm for estimating E{ g ( X , Y ) | h ( X , Y ) = r } where g and h are real-valued functions. This algorithm is applicable to time series problems in which X = ( X ₁, . . . , X _n−1) and Y = X_n where { x_t } is a discrete time stochastic process for which ( X₁ , . . . , X_n ) is a continuous random vector. A numerical example from time series analysis illustrates the algorithim, for prediction for an ARCH(1) process.     

Products of random variables and star-shaped ordering

It is shown that if V₁ and V₂ are two positive random variables such that V₁ is “star-shaped” with respect to V₂, then for any random variable X with a distribution F(x) such that F(ax) has the monotone likelihood-ratio property, XV₁ is star-shaped with respect to XV₂. This result is then used to prove that the stable laws are star-shaped ordered.     

Generalizations of a contamination model for continuous type random variables

In 1965, Stanley Warner (Warner, 1965) introduced a model for contaminating discrete type random variables. He presented this scheme as being potentially useful in survevs where sensitive in-formation is being gathered. Since that time much research has been conducted and many papers written on the development of these discrete type randomized response models. More recently, atten-tion has been focused on the application of randomized response type models for preservation of confidentiality in existing data files (Boruch 1971 and 1972, Ranney 1975, Felligi 1974, and Inge-marsson 1975). In 1974, Poole (Poole, 1974) introduced a randomized response model for a positive continuous type random variable which was basically a continuous variable analog of the discrete variable Warner model. In this paper the results of the 1974 paper are extended to a lt-dimensional continuous type random variable in k-dimensional Euclidean space.     

The sign test in Current Control

Current Control of a continuous random variable is investigated without assuming a normal distribution. The sign test is used as test procedure. The parameters of the corresponding control chart are to be determined in such a way that in the long run the expected loss per produced item is minimal. Approximation formulas for the parameters of such an optimum control chart are derived and discussed.     

Covariate Order Tests for Covariate Effect

A new approach for constructing tests for association between a random right censored life time variable and a covariate is proposed. The basic idea is to first arrange the observations in increasing order of the covariate and then base the test on a certain point process defined by the observation times. Tests constructed by this approach are robust against outliers in the covariate values or misspecification of the covariate scale since they only use the ordering of the covariate. Of particular interest is a test based on the Anderson-Darling statistic. This test has good power properties both against monotonic and nonmonotonic dependencies between the covariate and the life time variable.     

Goodness of Fit Test and Latent Distribution Estimation in the Mixed Rasch Model

The Cramér-von Mises test methodology is applied to build a goodness-of fit test for the mixed Rasch model. The Mixed Rasch Model is a probability model of a multivariate discrete random variable driven by an unknown latent continuous variable. The problem of estimation of the unknown fixed difficulty parameters and the latent density function is also considered. The theoretical results are illustrated through simulations and an application to real Quality of Life data.     

Asymptotocally nonparametric tests with censored paired data

A class of asymptotically nonparametric test with contains a test proposed by Wei(1980), is considered for testing the equality of two continuous distribution funcitons when paired observations are subject to arbitrary right censorship. It is shown that under the null hypothesis each test statistic converges in distribution to the standard normal random variable. Furthermore. the Monte Carlo simulation results indicate that some tests in this class are more powerful than Wei's test. A generalization to incomplete censored paired data is also included.     

The inspection paradox with random time

When considering a delayed renewal process one may be interested in both, the renewal function and the expected length of the interarrival time that contains some fixed time t. In general, it is difficult to obtain explicit expressions for specific underlying distributions. Replacing t by a random variable T and using prior information about T, that is, assuming that T has some continuous NBU (NWU) distribution function G, bounds of the quantities are derived as well as representations, if T is exponentially distributed. As an implication an equation of Wald type is shown. The results can be applied to the analysis of control charts in quality control. Moreover, related bounds of a sample mean based on a random sample size are given and an elementary renewal reward theorem is stated.     

Some orthogonal expansions for the logistic distribution

We expand a continuous random variable as a sum of a sequence of un-correlated random variables. These variables are principal components of a Bernoulli process, as well as principal dimensions in continuous metric scaling on a particular distance function. We obtain expansions for the uniform, exponential and logistic distributions. A goodness-of-fit application is given.