Comparing Several Regression Models with Unequal Variances

We consider the problem of comparing k regression models, when the variances are not assumed to be equal. For this problem, the classical F test can lead to misleading results, and there is no simple test which adequately controls the size when the sample sizes are small. For k = 2, the most widely used test is the “weighted F test,” also known as the “asymptotic Chow test.” But this test does not work well for small samples, and various modifications have been proposed in the literature. For k > 2, few tests are available and only the parametric-bootstrap (PB) test of Tian et al. (2009) Tian, L., Ma, C., Vexler, A. (2009). A parametric bootstrap test for comparing heteroscedastic regression models. Communications in Statistics—Simulation and Computation, 38, 1026–1036.[Taylor &; Francis Online], [Web of Science ®] , [Google Scholar] controls the size fairly adequately. In this article, we propose three fairly simple F tests which can easily be applied in the general case, k ? 2, and avoid the complications of the PB test. Our simulations indicate that these tests have satisfactory performance. Also, our simulations confirm that the power properties of our proposed tests are similar to the PB test. Therefore, our proposed tests provide simple alternatives to the PB test, which can easily be used by practitioners who may not be familiar with the PB.     

Analysis of Variance by Randomization when Variances are Unequal

If there are significant factor and interaction effects with analysis of variance using ran-domization inference, they can be detected by tests that compare the F -statistics for the real data with the distributions of these statistics obtained by randomly allocating either the original observations or the residuals to the various factor combinations. Such tests involve the assumption that the effect of factors or interactions is to shift the observations for a factor combination by a fixed amount, without changing the amount of variation at that combination. In reality the expected amount of variation at each factor combination, as measured by the variance, may not be constant, which may upset the properties of the tests for the effects of factors and interactions. This paper discusses several possible methods for adjusting the randomization procedure to allow for this type of problem, including generalizations of methods that have been proposed for comparing the means of several samples when there is unequal variance but no factor structure. A simulation study shows that the best of the methods examined is one for which the randomized sets of data are designed to approximate the distributions of F -statistics when unequal variance is present.     

Testing the equality of quantiles for several normal populations

Many procedures exist for testing equality of means or medians to compare several independent distributions. However, the mean or median do not determine the entire distribution. In this article, we propose a new small-sample modification of the likelihood ratio test for testing the equality of the quantiles of several normal distributions. The merits of the proposed test are numerically compared with the existing tests—a generalized p-value method and likelihood ratio test—with respect to their sizes and powers. The simulation results demonstrate that proposed method is satisfactory; its actual size is very close to the nominal level. We illustrate these approaches using two real examples.     

A Parametric Bootstrap Approach for One-Way ANOVA Under Unequal Variances with Unbalanced Data

This research is to provide a solution of one-way ANOVA without using transformation when variances are heteroscedastic and group sizes are unequal. Parametric bootstrap test (Krishnamoorthy et al., 2007 Krishnamoorthy, K., Lu, F., Mathew, T. (2007). A parametric bootstrap approach for anova with unequal variances: Fixed and random models. Computational Statistics and Data Analysis 51:5731–5742.[Crossref], [Web of Science ®] , [Google Scholar]) has been shown to be competitive with many other methods when testing the equality of group means. We extend the parametric bootstrap algorithm to a multiple comparison procedure. Simulation results show that the parametric bootstrap approach works well for one-way ANOVA.     

The adaptive calibration of testing p-values

Abstract

In statistical hypothesis testing, a p-value is expected to be distributed as the uniform distribution on the interval (0, 1) under the null hypothesis. However, some p-values, such as the generalized p-value and the posterior predictive p-value, cannot be assured of this property. In this paper, we propose an adaptive p-value calibration approach, and show that the calibrated p-value is asymptotically distributed as the uniform distribution. For Behrens–Fisher problem and goodness-of-fit test under a normal model, the calibrated p-values are constructed and their behavior is evaluated numerically. Simulations show that the calibrated p-values are superior than original ones.     

Behrens-Fisher问题的正态逼近

 本文提出用基于得分检验的正态逼近方法来解决Behrens-Fisher问题,即比较方差比未知时两正态总体的均值。模拟结果显示：在所有的研究情况下,我们的方法都能很好地控制第一类错误,检验功效也不差;而最常用的Welch近似t检验在样本量不等时大多数情况都不能控制第一类错误。     

Confidence Intervals of Effect Size for Randomized Comparative Parallel-Group Studies with Unequal Variances

Recent years have seen a heightened interest in estimating effect size—a common measure of effect magnitude in biomedical research—because of its direct clinical relevance. In this article, three interval estimates of effect size for randomized comparative parallel-group studies with unequal variances are discussed. Two real-life examples illustrate that confidence intervals obtained by three methods are quite different, especially when the sample sizes are small. Simulation results show that confidence intervals generated by the modified signed log-likelihood ratio method yield essentially the exact coverage probabilities, whereas the other two methods, even though they are more popular methods, yield less satisfactory results.     

Robust weighted one-way ANOVA: Improved approximation and efficiency

A robust test for the one-way ANOVA model under heteroscedasticity is developed in this paper. The data are assumed to be symmetrically distributed, apart from some outliers, although the assumption of normality may be violated. The test statistic to be used is a weighted sum of squares similar to the Welch [1951. On the comparison of several mean values: an alternative approach. Biometrika 38, 330-336.] test statistic, but any of a variety of robust measures of location and scale for the populations of interest may be used instead of the usual mean and standard deviation. Under the commonly occurring condition that the robust measures of location and scale are asymptotically normal, we derive approximations to the distribution of the test statistic under the null hypothesis and to its distribution under alternative hypotheses. An expression for relative efficiency is derived, thus allowing comparison of the efficiency of the test as a function of the choice of the location and scale estimators used in the test statistic. As an illustration of the theory presented here, we apply it to three commonly used robust location–scale estimator pairs: the trimmed mean with the Winsorized standard deviation; the Huber Proposal 2 estimator pair; and the Hampel robust location estimator with the median absolute deviation.     

Generalized Inferences on the Common Mean in MANOVA Models

This article presents procedures for testing hypothesis and interval estimation of the common mean vector in MANOVA models when the covariance matrices are unknown and unequal. The methods are based on the concepts of generalized p-value and generalized confidence interval. Some important statistical properties of the exact test and confidence region are given. For two multivariate normal populations, a minor modification to the combined tests given by Zhou and Mathew (1994a Zhou , L. P. , Mathew , T. ( 1994a ). Combining independent tests in multivariate linear models . J. Multivariate Anal. 51 : 265 – 276 . [Google Scholar]) is proposed. Some simulation results to compare the performance of the proposed tests with others are reported. The simulation results indicate that new tests have significant gain in the power.     

Assessing Occupational Exposure via the Unbalanced One-Way Random Effects Model

In this article, an unbalanced one-way random effects model is considered for the log-transformed shift-long exposure measurements. Exact test and confidence interval for the proportion of workers whose mean exposure exceeds the occupational exposure limit are developed based on the concepts of generalized p-value and generalized confidence interval. Some simulation results to compare the performance of the proposed test with that of the existing method are reported. The simulation results indicate that the proposed method appears to have significant gain in the size and power.     

A Revisit to the Behrens–Fisher Problem: Comparison of Five Test Methods

We revisit the well-known Behrens–Fisher problem and apply a newly developed ‘Computational Approach Test’ (CAT) to test the equality of two population means where the populations are assumed to be normal with unknown and possibly unequal variances. An advantage of the CAT is that it does not require the explicit knowledge of the sampling distribution of the test statistic. The CAT is then compared with three widely accepted tests—Welch–Satterthwaite test (WST), Cochran–Cox test (CCT), ‘Generalized p-value’ test (GPT)—and a recently suggested test based on the jackknife procedure, called Singh–Saxena–Srivastava test (SSST). Further, model robustness of these five tests are studied when the data actually came from t-distributions, but wrongly perceived as normal ones. Our detailed study based on a comprehensive simulation indicate some interesting results including the facts that the GPT is quite conservative, and the SSST is not as good as it has been claimed in the literature. To the best of our knowledge, the trends observed in our study have not been reported earlier in the existing literature.     

Inferences on the Coefficients of Variation in a Multivariate Normal Population

In this article, the problem of testing the equality of coefficients of variation in a multivariate normal population is considered, and an asymptotic approach and a generalized p-value approach based on the concepts of generalized test variable are proposed. Monte Carlo simulation studies show that the proposed generalized p-value test has good empirical sizes, and it is better than the asymptotic approach. In addition, the problem of hypothesis testing and confidence interval for the common coefficient variation of a multivariate normal population are considered, and a generalized p-value and a generalized confidence interval are proposed. Using Monte Carlo simulation, we find that the coverage probabilities and expected lengths of this generalized confidence interval are satisfactory, and the empirical sizes of the generalized p-value are close to nominal level. We illustrate our approaches using a real data.     

A Test for the Equality of Parameters for Separate Regression Models in the Presence of Heteroskedasticity

Testing for the equality of regression coefficients across two regressions is a problem considered by analysts in a variety of fields. If the variances of the errors of the two regressions are not equal, then it is known that the standard large sample F-test used to test the equality of the coefficients is compromised by the fact that its actual size can differ substantially from the stated level of significance in small samples. This article addresses this problem and borrows from the literature on the Behrens-Fisher problem to provide some simple modifications of the large sample test which allows one to better control the probability of committing a Type I error. Empirical evidence is presented which indicates that the suggested modifications provide tests which are superior to well-known alternative tests over a wide range of the parameter space.     

Robust test for means when population variances are unequal

We consider the problem of testing the equality of two population means when the population variances are not necessarily equal. We propose a Welch-type statistic, say T^* _c, based on Tiku!s ‘1967, 1980’ modified maximum likelihood estimators, and show that this statistic is robust to symmetric and moderately skew distributions. We investigate the power properties of the statistic T^* _c; T^* _c clearly seems to be more powerful than Yuen's ‘1974’ Welch-type robust statistic based on the trimmed sample means and the matching sample variances. We show that the analogous statistics based on the ‘adaptive’ robust estimators give misleading Type I errors. We generalize the results to testing linear contrasts among k population means     

A Note on an Unbiased Test of Homogeneity of Variances

A necessary and sufficient condition for unbiasedness of the test of homogeneity of variances in normal samples is derived in a convenient form. In the case of two samples, it is shown that Bartlett's test is the only unbiased test of homogeneity of variances. A simple alternative proof of the unbiasedness of Bartlett's test in the general case is also provided.     

Testing for a separation between two normal distributions

Three tests are proposed for testing for a specified degree of overlap between two normal distributions, The hypotheses considered are an extension of the Behrens-Fisher problem, A simulation study of the performance of the tests is presented.     

Testing the equality of several log-normal means based on a computational approach

In this study, we propose a new test based on a computational approach to test the equality of several log-normal means. We compare this test with some existing methods in terms of the type-I error rate and power using Monte Carlo simulations for varying values of number of groups and sample sizes. The simulation results indicate that the proposed test could be suggested as a good alternative for testing the equality of several log-normal means.     

Inferences on the Means of Two Log-Normal Distributions: A Computational Approach Test

In this article, we propose a novel approach for testing the equality of two log-normal populations using a computational approach test (CAT) that does not require explicit knowledge of the sampling distribution of the test statistic. Simulation studies demonstrate that the proposed approach can perform hypothesis testing with satisfying actual size even at small sample sizes. Overall, it is superior to other existing methods. Also, a CAT is proposed for testing about reliability of two log-normal populations when the means are the same. Simulations show that the actual size of this new approach is close to nominal level and better than the score test. At the end, the proposed methods are illustrated using two examples.