A comparison of two approaches for power and sample size calculations in logistic regression models

Whittemore (1981) proposed an approach for calculating the sample size needed to test hypotheses with specified significance and power against a given alternative for logistic regression with small response probability. Based on the distribution of covariate, which could be either discrete or continuous, this approach first provides a simple closed-form approximation to the asymptotic covariance matrix of the maximum likelihood estimates, and then uses it to calculate the sample size needed to test a hypothesis about the parameter. Self et al. (1992) described a general approach for power and sample size calculations within the framework of generalized linear models, which include logistic regression as a special case. Their approach is based on an approximation to the distribution of the likelihood ratio statistic. Unlike the Whittemore approach, their approach is not limited to situations of small response probability. However, it is restricted to models with a finite number of covariate configurations. This study compares these two approaches to see how accurate they would be for the calculations of power and sample size in logistic regression models with various response probabilities and covariate distributions. The results indicate that the Whittemore approach has a slight advantage in achieving the nominal power only for one case with small response probability. It is outperformed for all other cases with larger response probabilities. In general, the approach proposed in Self et al. (1992) is recommended for all values of the response probability. However, its extension for logistic regression models with an infinite number of covariate configurations involves an arbitrary decision for categorization and leads to a discrete approximation. As shown in this paper, the examined discrete approximations appear to be sufficiently accurate for practical purpose.     

Some non-parametric tests for duration dependence: an application to UK business cycle data

A distinction between Fisher's implied data-generating process for Monte Carlo cycles and the more general Markov process leads to non-parametric tests for duration dependence. Tests are based on the method of moments, Tauchen's generalized method of moments (GMM) procedure, and a statistic whose null distribution probability limit is zero. Using finite-sample critical values obtained by Monte Carlo methods, our test results are remarkably consistent. The null distribution of the GMM test statistic for samples of the size considered is distinctly non-normal, so that asymptotic critical values give erroneous results. The tests are applied to UK business cycle data for 1854-1992. There is evidence for duration dependence in expansions but not in contractions.     

Precedence tests for equality of two distributions based on early failures of ranked set samples

In this article, two different types of precedence tests, each with two different test statistics, based on ranked set samples for testing the equality of two distributions are discussed. The exact null distributions of proposed test statistics are derived, critical values are tabulated for both set size and number of cycles up to 8, and the exact power functions of these two types of precedence tests under the Lehmann alternative are derived. Then, the power values of these two test procedures and their competitors based on simple random samples and based on ranked set samples are compared under the Lehmann alternative exactly and also under a location-shift alternative by means of Monte Carlo simulations. Finally, the impact of imperfect ranking is discussed and some concluding remarks are presented.     

A K-S Type Test for Exponentiality Based on Empirical Hankel Transforms

A uniqueness theorem for a recently introduced special Hankel transform of probability distributions on the non negative half-line motivates a K-S type test statistic based on empirical Hankel transforms for testing the hypothesis of exponentiality. This article deals with the asymptotic behavior of the new test.     

Testing a markov chain for independence

A bootstrap procedure is proposed for testing whether an observed Markov chain is actually an independent process, based on the observed transition probability matrix. The results of simulations showing the power and size of the bootstrap test are presented. The asymptotic distribution of the non-unit eigenvalues is given under the null hypothesis.     

Goodness of fit test for the generalized Rayleigh distribution with unknown parameters

The use of goodness-of-fit test based on Anderson–Darling (AD) statistic is discussed, with reference to the composite hypothesis that a sample of observations comes from a generalized Rayleigh distribution whose parameters are unspecified. Monte Carlo simulation studies were performed to calculate the critical values for AD test. These critical values are then used for testing whether a set of observations follows a generalized Rayleigh distribution when the scale and shape parameters are unspecified and are estimated from the sample. Functional relationship between the critical values of AD is also examined for each shape parameter (α), sample size (n) and significance level (γ). The power study is performed with the hypothesized generalized Rayleigh against alternate distributions.     

Some Results on Comparing the Quantiles of Dependent Groups

Recently, Lombard derived an extension of the Doksum–Sievers shift function to dependent groups. This article suggests using a particular numerical method for determining the critical value, reports on the ability of the method to control the probability of a Type I error when sample sizes are small, and it provides comparisons with methods aimed at comparing deciles. It is found that for continuous distributions, Lombard's method performs well and in particular has high power relative to the other two methods considered. But when tied values can occur, now it can have relatively poor power; a method based on the Harrell-Davis estimator is found to give more satisfactory results.     

A Multisample Rank Test for Location-Scale Parameters

One of the multisample problems is discussed in this article. A new multisample rank tests based on a k-sample Baumgartner statistic are proposed for testing the location-scale parameters. The exact critical values of proposed statistics are calculated. Simulations are used to investigate the power of proposed statistics for various population distributions.     

Locally best tests for exponential distributions with interval-censored data

In this article, we derive a locally best test for testing the mean of exponential distributions with interval-censored samples. This locally best test possesses certain optimality. It is of unbiasedness and equivalent to a likelihood ratio test in some circumstances, and it is also a Bayes test for some loss function. For the locally best test, the associated critical values and powers at a nominal level of significance are provided. For a large sample size case, asymptotic critical values and powers are also calculated and tabulated. Moreover, based on the locally best test, a curtailed test is proposed. This curtailed test is equivalent to the locally best test on the acceptance or rejection of the null hypothesis. A Monte Carlo simulation is carried out to illustrate the performance of the curtailed test compared with the locally best test. Numerical results show that the experimental duration time of the curtailed test is substantially smaller than that of the locally best test.     

A BETA-OPTIMAL TEST OF THE EQUICORRELATION COEFFICIENT

This paper considers the problem of testing for nonzero values of the equicorrelation coefficient of a standard symmetric multivariate normal distribution. Recently, SenGupta (1987) proposed a locally best test. We construct a beta-optimal test and present selected one and five percent critical values. An empirical power comparison of SenGupta's test with two versions of the beta-optimal test and the power envelope shows the relative strengths of the three tests. It also allows us to assess and confirm Efron's (1975) rule of when to question the use of a locally best test, at least for this testing problem. On the basis of these results, we argue that the two beta-optimal tests can be considered as approximately uniformly most powerful tests, at least at the five percent significance level.     

First hitting time of integral diffusions and applications

We study the first hitting time of integral functionals of time-homogeneous diffusions, and characterize their Laplace transforms through a stochastic time change. We obtain explicit expressions of the Laplace transforms for the geometric Brownian motion (GBM) and the mean-reverting GBM process. We also introduce a novel probability identity based on an independent exponential randomization and obtain explicit Laplace transforms of the price of arithmetic Asian options and other derivative prices that non-linearly depend on the integral diffusions. Numerical examples are given to demonstrate the accuracy and efficiency of the proposed method.     

Optimal inference for Simon's two-stage design with over or under enrollment at the second stage

Simon's two-stage designs are widely used in clinical trials to assess the activity of a new treatment. In practice, it is often the case that the second stage sample size is different from the planned one. For this reason, the critical value for the second stage is no longer valid for statistical inference. Existing approaches for making statistical inference are either based on asymptotic methods or not optimal. We propose an approach to maximize the power of a study while maintaining the type I error rate, where the type I error rate and power are calculated exactly from binomial distributions. The critical values of the proposed approach are numerically searched by an intelligent algorithm over the complete parameter space. It is guaranteed that the proposed approach is at least as powerful as the conditional power approach which is a valid but non-optimal approach. The power gain of the proposed approach can be substantial as compared to the conditional power approach. We apply the proposed approach to a real Phase II clinical trial.     

Detecting parameter shift in garch models

This paper applies recent theories of testing for parameter constancy to the conditional variance in a GARCH model. The supremum Lagrange multiplier test for conditional Gaussian GARCH models and its robustified variants are discussed. The asymptotic null distribution of the test statistics are derived from the weak convergence of the scores, and the critical values from the hitting probability of squared Bessel process.

Monte Carlo studies on the finite sample size and power performance of the supremum LM tests are conducted. Applications of these tests to S&P 500 indicate that the hypothesis of stable conditional variance parameters can be rejected.     

Detecting parameter shift in garch models

This paper applies recent theories of testing for parameter constancy to the conditional variance in a GARCH model. The supremum Lagrange multiplier test for conditional Gaussian GARCH models and its robustified variants are discussed. The asymptotic null distribution of the test statistics are derived from the weak convergence of the scores, and the critical values from the hitting probability of squared Bessel process.

Monte Carlo studies on the finite sample size and power performance of the supremum LM tests are conducted. Applications of these tests to S&P 500 indicate that the hypothesis of stable conditional variance parameters can be rejected.     

A a comparison of the markov chain and the integral equation approaches for evaluating the run length distribution of quality control charts

Two methods that are often used to evaluate the run length distribution of quality control charts are the Markov chain and integral equation approaches. Both methods have been used to evaluate the cumulative sum (CUSUM) charts and the exponentially weighted moving average (EWMA) control charts. The Markov chain approach involves "discretiz-ing" the possible values which can be plotted. Using properties of finite Markov chains, expressions for the distribution of the run length, and for the average run length (ARL), can be obtained. For the CUSUM and EWMA charts there exist integral equations whose solution gives the ARL. Approximate methods can then be used to solve the integral equation. In this article we show that if the product midpoint rule is used to approximate the integral in the integral equation, then both approaches yield the same approximations for the ARL. In addition we show that the recursive expressions for the probability functions are the same for the two approaches. These results establish the integral equation approach as preferable whenever an integral equation can be found     

基于回归的时变偏度和时变峰度识别检验

资产收益率时变高阶矩建模的首要前提是资产收益率的偏度和峰度具有时变性,即资产收益率存在类似于异方差性的异偏度和异峰度特征。目前文献中的时变偏度和时变峰度识别检验存在适用性较差且检验功效较低等不足。本文提出基于回归的检验方法识别资产收益率偏度和峰度的时变性。该检验一方面利用概率积分变换缓解了拉格朗日乘数检验对资产收益率序列的高阶矩存在性的限制,另一方面考虑了检验统计量中参数估计的不确定性对其统计性质的影响,具有良好的渐近统计性质且适用性更广。蒙特卡洛模拟表明该检验具有良好的有限样本性质,具有合适的检验水平和较高的检验功效。最后,将基于回归的检验运用于上证综指和深圳成指收益率的时变建模研究中。     

A Moment Preserving Finitization Across the Power Series Family of Probability Distributions

Finitization transforms a discrete distribution into a distribution with smaller support of specified size. In special cases finitization preserves moments (moments of the order n finitization coincide with those of the parent distribution). We create a moment preserving finitization method for power series distributions by introducing an alternative representation and showing how to finitize members of this new class in a manner that preserves moments of the parent distribution. We provide results on convolutions and a reproductive property for power series distributions that have been finitized in this manner, and show how these finitized distributions accelerate variate generation in simulation.     

Parameter estimates and tests of fit for infinite mixther distrirutions

Al though mixtures form a rich class of probability models, they often present difficulties for statistical inference. Likelihood functions are sometimes unbounded at certain values of the parameters, and densities often have no closed form. These features complicate hoth maximum-likelihood estimation and tests of fit based on the empirical distribution function. New inferential methods using sample characteristic functions (Cfs) and moment generating functions (MGFs) seem well-suited to mixtures. since these transforms often take simple form/ This paper reports a simulation study of the properties of estimators and tests of fit based on CFs, MGFs, and sample moments when applied to three specific families of thick tailed mixture distributios.     

On finite sample properties of the tests of robinson (1994) for fractional integration

This article proposes a bootstrap version of the tests of Robinson (1994) for testing unit and/or fractional roots. The finite-sample behaviour of the tests, based on these bootstrap critical values is compared with those based on asymptotic and on finite-sample results and with a number of leading unit-root tests. The Monte-Carlo simulations indicate that the bootstrap version of the tests of Robinson (1994) outperforms the other tests, including the one using finite-sample critical values. The improvement in the size and the power is particularly important under AR(1) alternatives. A small empirical application is also carried out with inflation for a panel of 16 European countries. The results show that the differences across countries depend on the critical values used: whereas the I (1) property of inflation is unclear with the asymptotic tests in some countries, the bootstrap version of Robinson's (1994) tests cannot reject the presence of a unit-root in inflation.     

Bootstrap Bartlett correction in inflated beta regression

The inflated beta regression model aims to enable the modeling of responses in the intervals (0, 1], [0, 1), or [0, 1]. In this model, hypothesis testing is often performed based on the likelihood ratio statistic. The critical values are obtained from asymptotic approximations, which may lead to distortions of size in small samples. In this sense, this article proposes the bootstrap Bartlett correction to the statistic of likelihood ratio in the inflated beta regression model. The proposed adjustment only requires a simple Monte Carlo simulation. Through extensive Monte Carlo simulations the finite sample performance (size and power) of the proposed corrected test is compared to the usual likelihood ratio test and the Skovgaard adjustment already proposed in the literature. The numerical results evidence that inference based on the proposed correction is much more reliable than that based on the usual likelihood ratio statistics and the Skovgaard adjustment. At the end of the work, an application to real data is also presented.