Search

Instead of relying on null-hypothesis significance testing (NHST), researchers are consistently advised to use effect sizes (ESs) and confidence intervals (CIs) to convey research findings. However, typical ES measures for most linear models (e.g., multiple regression) assume data normality, a condition that is often violated in behavioral research. This may lead to inaccurate interpretation of ES. In multiple regression models, Cohen's f2, R2 and R adj 2 are employed by researchers, but no study has systematically evaluated their robustness in practice. Thus, this Monte Carlo simulation study evaluates the robustness of f2, R2 and R adj 2 and the associated CIs based on manipulated levels of sample sizes, magnitudes of ESs, numbers of predictors, and data violations (i.e., heavy-tailed, skewed, contaminated, lognormal, and heteroscedastic distributions of errors). This study offers guidelines regarding how robust these ESs are so that researchers can report the most appropriate ES in their research studies. [ABSTRACT FROM AUTHOR]

Small number of clusters combined with cluster level heterogeneity poses a great challenge for the data analysis. We have published a weighted Jackknife approach to address this issue applying weighted cluster means as the basic estimators. The current study proposes a new version of the weighted delete-one-cluster Jackknife analytic framework, which employs Ordinary Least Squares or Generalized Least Squares estimators as the fundamentals. Algorithms for computing estimated variances of the study estimators have also been derived. Wald test statistics can be further obtained, and the statistical comparison in the outcome means of two conditions is determined using the cluster permutation procedure. The simulation studies show that the proposed framework produces estimates with higher precision and improved power for statistical hypothesis testing compared to other methods. [ABSTRACT FROM AUTHOR]

The intraclass correlation coefficient ρ is used as a measure of the degree of resemblance among family members with respect to some biological or psychological attribute, such as blood pressure level or IQ. Many studies have been done in the area. However, a significant gap in the literature is that few statistics are available to test the significance ( H 0 : ρ = 0 against H A : ρ ≠ 0 ) of this important biological measure and no study is available as to the performance of these statistics with respect to level and power. In this paper we develop three large sample test statistics and compare these, in terms of level and power, using extensive simulation studies of these, the Wald test developed in this paper is found to have a very simple form and highest power while maintaining level well. [ABSTRACT FROM AUTHOR]

Data analysis techniques that rely on standard statistical tools and algorithms often encounter problems when dealing with data sets that have large sample sizes. In this study, two statistical tests done in conducting simple linear regression analysis were revisited. In particular, the study simulated the effects of large sample sizes and amount of contamination in the data due to non-sampling errors on the false positive rate of the Kolmogorov-Smirnov (K-S) test in testing for normality of error terms. The study also characterized the effects of varying sample size and amount of contamination in the data on the false negative rate of the t-test in testing the significance of a regression coefficient. Lastly, an optimality index was developed to determine the sample sizes and the values of the percent noise at which both the false positive rate of the K-S test and the false negative rate of the t-test are minimized. [ABSTRACT FROM AUTHOR]

This work proposes powerful tests using bootstrap methods for the tail index in the family of distribution functions with nondegenerate right tail. It is of interest to test whether the right tail of a distribution function is the same as or heavier than that of the Pareto distribution with index m0 for some m0, vs. alternatively the right tail is lighter. Based on the nonparametric tests of Jureckova and Picek (2001, A class of tests on the tail index, Extreme 4:2, 165–183) that use empirical distribution functions of sample maxima, the proposed bootstrap tests employ two-stage subsamplings: One is subsampling from the sample maxima to be applicable even to the heavy tailed index and the other is subsampling from collections of empirical distributions of the sample maxima. We show the first-order validity of the bootstrap for the nonparametric test and construct bootstrap test statistics. The limiting distribution of the bootstrap test statistics in the null hypothesis is established and the consistency of the bootstrap tests is verified. Applications of the proposed bootstrap tests are given to autoregressive time series models. A simulation study is conducted to see performance of the bootstrap tests and it illustrates that the proposed bootstrap tests outperform the existing ones. Real data of heavy tailed financial stock indices are applied to the proposed bootstrap tests. [ABSTRACT FROM AUTHOR]

We propose a kernel density based estimation for nonlinear regression models when neither the response variable nor the covariates can be directly observed, but are measured with multiplicative distortion measurement errors. For the hypothesis testing of parametric components, restricted estimators under the null hypothesis and test statistics are proposed. The asymptotic properties for the estimators and test statistics are established. A bootstrap procedure is proposed to calculate critical values. Lastly, we make some comparisons between the proposed estimator and nonlinear least squares estimator through the simulation. Our simulation studies indicate that the kernel density based estimator provides a better alternative under some settings, particularly for multimodal distribution of the error term. [ABSTRACT FROM AUTHOR]

Multivariate isotonic regression theory plays an important role in the field of testing statistical hypotheses under order restriction for several vector valued parameters and this problem may arise in several situations. Sasabuchi et al. derived a likelihood ratio type test for homogeneity of order restricted mean vectors under a multivariate normality assumption with unknown and common covariance matrices; however, its null distribution depends on the unknown covariance matrix. In order to overcome this difficulty, in the present paper we propose a test that conditions on the sufficient statistic for the covariance matrix. We remove the dependence of the null distribution of test statistic proposed by Sasabuchi et al. on the unknown value of covariance matrix and present a condition on a sufficient statistic for covariance matrix under the null hypothesis. A method based on Markov Chain Monte Carlo sampling is used to calculate the p - value of the test. We evaluate the operating characteristics of the proposed test through simulation studies. Our extensive simulation studies demonstrate that the power of this new test is generally higher than the test proposed by Sasabuchi et al. and also by Sasabuchi. A real data example is presented. [ABSTRACT FROM AUTHOR]

In a high-dimensional linear regression model, this article is concerned with testing statistical significance of a subset of regression coefficients. The conventional partial F-test is not applicable in high-dimensional situations. Several methods for testing whether any of the discarded covariates is significant conditional on relative importance of predictors have been proposed in the recent literature, but they are adversely affected by the overestimation of the variance. To overcome this issue, we propose a novel nonparametric testing procedure to avoid this problem and enhance the empirical power. In addition, the new test is very effective when error distribution deviates from the normal scenario and can integrate all the individual information of the discarded covariates. Under the high-dimensional null and alternative hypotheses, we derive the asymptotic distribution of the proposed test statistic, which allows power evaluation of the test. Numerical studies are carried out to examine the numerical performance of the test. The results show that the test proposed here behaves well in terms of sizes and power and significantly outperforms the existing choices in a range of settings. [ABSTRACT FROM AUTHOR]

We consider the estimation and hypothesis test for single-index model by proposing a profile mean squares relative error estimation method, coupling with a leave-one-component-out method. The asymptotic normality for the PMSRE estimator is studied. For the hypothesis test of the single-index parameter, a restricted estimator under the null hypothesis and test statistics are proposed. The asymptotic properties for the restricted estimator and test statistic are established. Lastly, we make some comparisons between the proposed estimator and the profile least squares estimator and the profile least product relative estimator through the simulation. Our simulation studies indicate that the profile mean squares relative error estimator provides a better alternative under some settings, particularly for multimodal or asymmetric distributions of the error term. [ABSTRACT FROM AUTHOR]

In this article, we mainly discuss hypothesis tests to assess whether nonparametric function satisfies monotonicity or convexity in semiparametric partially linear model. For this purpose, we propose a new test statistic for monotonicity or convexity of nonparametric function based on Bernstein polynomials. Furthermore, we employ likelihood ratio statistic to test the significance of regression parameters of the model. We discuss the asymptotic properties of the tests for both nonparametric function and regression parameters. A simulation study is conducted to evaluate the finite sample performance compared with the other methods for shape tests. The method is illustrated by the fuel efficiency study. [ABSTRACT FROM AUTHOR]

In social sciences, a plethora of studies utilize nominal data to establish the relationship between the variables. This, in turn, requires the correct use of correlation technique. The choice of correlation technique depends upon the underlying assumptions and power of the test of significance. The objective of the research is to explore the best measure of association for nominal data in terms of size, power and bias in estimation. Monte Carlo simulations reveal that the Phi and Pearson correlation statistics performs equally well in terms of size, power, and bias for naturally dichotomous variables. When both variables are artificially dichotomized, the Tetrachoric statistic has an edge in terms of bias to Pearson correlation statistic. If one variable is continuous and other is artificially dichotomized, the Biserial correlation measure turns out to be less biased as compared to Pearson statistic although both statistics exhibit similar power and size properties. If one variable is continuous and other is naturally dichotomized, it is hard to choose between the Point Biserial and Pearson correlation measures. Finally, if one variable is naturally dichotomous and other is artificially dichotomized, correlation coefficient V is compared with Pearson, Phi and Tetrachoric correlation techniques in terms of bias in estimate. The results indicate that the Tetrachoric statistic considerably overestimates the correlation value against non-normal distributions. Pearson and Phi correlation slightly underestimate the correlation value. In contrast, the correlation statistic V perform well. [ABSTRACT FROM AUTHOR]

In this article the optimal statistical hypotheses testing to make decision for arbitrarily varying object are considered. The statistician would like to control the overall error rate relative to draw statistically valid conclusions from each test. The familywise error rate (FWE) metric and the hypothesis test procedure while controlling both the type I and II FWEs are generalized by an arbitrarily varying object. The characteristics of logarithmically asymptotically optimal (LAO) hypotheses testing are studied. The purpose of research is to express the optimal functional relation among the reliabilities of LAO hypotheses testing and to judge with FWE metric. [ABSTRACT FROM AUTHOR]

Suppose we have k samples X 1 , 1 , ... , X 1 , n 1 , ... , X k , 1 , ... , X k , n k with different sample sizes n 1 , ... , n k and unknown underlying distribution functions F 1 , ... , F k as observations plus k families of distribution functions { G 1 (· , ϑ) ; ϑ ∈ Θ } , ... , { G k (· , ϑ) ; ϑ ∈ Θ } , each indexed by elements ϑ from the same parameter set Θ, we consider the new goodness-of-fit problem whether or not (F 1 , ... , F k) belongs to the parametric family { (G 1 (· , ϑ) , ... , G k (· , ϑ)) ; ϑ ∈ Θ }. New test statistics are presented and a parametric bootstrap procedure for the approximation of the unknown null distributions is discussed. Under regularity assumptions, it is proved that the approximation works asymptotically, and the limiting distributions of the test statistics in the null hypothesis case are determined. Simulation studies investigate the quality of the new approach for small and moderate sample sizes. Applications to real-data sets illustrate how the idea can be used for verifying model assumptions. [ABSTRACT FROM AUTHOR]

A semi-parametric spatial-temporal model is estimated using a hybrid of forward search algorithm and nonparametric regression in the context of backfitting an additive model. The model can account for a structural change represented by a function of time. Robust estimates of the parametric component are produced and the predicted model also exhibit good predictive ability. Nonparametric bootstrap is used to facilitate significance test for structural change even without prior knowledge of its nature. The test is properly-sized and powerful as observed from the simulation studies. [ABSTRACT FROM AUTHOR]

In this note we propose a new semi-parametric bootstrap procedure for hypothesis tests about a statistical function and termed bootstrap warping. This procedure was motivated by empirical likelihood and bootstrap tilting techniques. The procedure is computationally efficient and has a fixed number of parameters. We show that the warping procedure has good type I error control and has monotone power as a function of sample size and shift alternatives. [ABSTRACT FROM AUTHOR]