Hypotheses testing for generalized order statistics with simple order restrictions on model parameters under the alternative

In applications of generalized order statistics as, for instance, reliability analysis of engineering systems, prior knowledge about the order of the underlying model parameters is often available and may therefore be incorporated in inferential procedures. Taking this information into account, we establish the likelihood ratio test, Rao's score test, and Wald's test for test problems arising from the question of appropriate model selection for ordered data, where simple order restrictions are put on the parameters under the alternative hypothesis. For simple and composite null hypothesis, explicit representations of the corresponding test statistics are obtained along with some properties and their asymptotic distributions. A simulation study is carried out to compare the order restricted tests in terms of their power. In the set-up considered, the adapted tests significantly improve the power of the associated omnibus versions for small sample sizes, especially when testing a composite null hypothesis.     

Second order asymptotic and non-asymptotic optimality properties of combined tests

Let p independent test statistics be available to test a null hypothesis concerned with the same parameter. The p are assumed to be similar tests. Asymptotic and non-asymptotic optimality properties of combined tests are studied. The asymptotic study centers around two notions. The first is Bahadur efficiency. The second is based on a notion of second order comparisons. The non-asymptotic study is concerned with admissibility questions. Most of the popular combining methods are considered along with a method not studied in the past. Among the results are the following: Assume each of the p statistics has the same Bahadur slope. Then the combined test based on the sum of normal transforms, is asymptotically best among all tests studied, by virtue of second order considerations. Most of the popular combined tests are inadmissible for testing the noncentrality parameter of chi-square, t, and F distributions. For chi-square a combined test is offered which is admissible, asymptotically optimal (first order), asymptotically optimal (second order) among all tests studied, and for which critical values are obtainable in special cases. Extensions of the basic model are given.     

The test for logical order in English language testing

In this paper we assess a marking scheme for the test for logical order in English language testing. This is achieved by considering the potential misclassification of students due to guessing. The distribution of scores under the hypothesis of random selection is derived and the asymptotic distribution is shown to be Poisson with a mean of unity. Recommendations are given for selecting the number of sentences in the test for logical order.     

A test for second order stochastic dominance

In this paper we propose a test for second order stochastic dominance (SSD), for the case where both distribution functions are unknown. This is a generalization of a test proposed by Deshpande and Singh (1985), who compare a new random prospect with a known distribution function. We then show that our test is based on comparing the mean minus one half of Gini's mean difference of the distributions, which is known to be a necessary condition for SSD, as developed in the economics literature (Yitzhaki, 1982).     

Test of Significance in order selection

To select a proper order is a critical step in fitting a model to a data set. It is well known that Mallows' C_p,as well as other asymptotically equivalent order selectors, often selects an excessively large order. To reduce the chance of overestimating the optimal order using a heavier penalty term has been suggested. In this paper, the issue about setting the penalty term is considered from a perspective similar to hypothesis testing. It is suggested setting the penalty according to the probabiiity of overestimating the true or optimal order. A test procedure based on the order selecters is proposed to assess the significance of the order selected. Some simulation results are presented to illustrate the properties of the order selectors and the test procedure. The application of the procedures to two real examples are also presented.     

On nonparametric test for the residual probability order

Residual probability order is a partial ordering analog of the total order stochastic precedence which can be fruitfully applied in many reliability problems. We introduce a U-statistic test for the null hypothesis that two random variables are equally distributed versus the alternative that one strictly dominates the other in residual probability order. The critical value is obtained by normal approximation and bootstrap procedures. The performance of the test statistic is evaluated using a simulation study. Finally, a numerical example illustrating the theory is also given.     

Testing the dilation order by using cumulative residual Tsallis entropy

In this paper, we introduce a new test for the dilation order based on cumulative residual Tsallis entropy of order α. The effect of the values of parameter α on the power of the test statistics is numerically investigated. The asymptotic distribution of the test statistic is given. The performance of the test statistic is evaluated using a simulation study. Finally, some numerical examples illustrating the theory are also given.     

Inference in a two-parameter generalized order statistics model

As a submodel of generalized order statistics with two unknown model parameters, m-generalized order statistics may serve as a simple model for ordered quantities in a given application. It is shown that the joint distribution of m-generalized order statistics has a representation as a regular exponential family in the model parameters, as it is the case for the comprising model. Utilizing this finding, a minimal sufficient and complete statistic is obtained along with distributional properties. Joint maximum likelihood estimation of the parameters is considered, and strong consistency and asymptotic efficiency of the estimator are established. A test is provided to decide whether a restriction to the submodel is reasonable.     

Multivariate testing and model-checking for generalized order statistics with applications

The exponential family structure of the joint distribution of generalized order statistics is utilized to establish multivariate tests on the model parameters. For simple and composite null hypotheses, the likelihood ratio test (LR test), Wald's test, and Rao's score test are derived and turn out to have simple representations. The asymptotic distribution of the corresponding test statistics under the null hypothesis is stated, and, in case of a simple null hypothesis, asymptotic optimality of the LR test is addressed. Applications of the tests are presented; in particular, we discuss their use in reliability, and to decide whether a Poisson process is homogeneous. Finally, a power study is performed to measure and compare the quality of the tests for both, simple and composite null hypotheses.     

Testing order restricted hypotheses with proportional hazards

Inferences for survival curves based on right censored continuous or grouped data are studied. Testing homogeneity with an ordered restricted alternative and testing the order restriction as the null hypothesis are considered. Under a proportional hazards model, the ordering on the survival curves corresponds to an ordering on the regression coefficients. Approximate likelihood methods are obtained by applying order restricted procedures to the estimates of the regression coefficients. Ordered analogues to the log rank test which are based on the score statistics are considered also. Chi-bar-squared distributions, which have been studied extensively, are shown to provide reasonable approximations to the null distributions of these tests statistics. Using Monte Carlo techniques, the powers of these two types of tests are compared with those that are available in the literature.     

Testing for order among K populations: theory and examples

Testing for stochastic order among K populations is a common and important problem in statistical practice. It arises in the analysis of both planned experiments and observational studies. The authors develop a new nonparametric test for order among K populations that can accommodate any stochastic ordering. The test is based on a maximally selected chi‐bar‐square statistic. The authors find its limiting distribution and use simulations to derive critical values. Three important examples are used to illustrate the applicability of the general method. The authors find that the new tests outperform the existing methods in many practical cases. The Canadian Journal of Statistics 38: 97–115; 2010 © 2009 Statistical Society of Canada     

Reversed hazard rate order of equilibrium distributions and a related aging notion

This paper deals with preservation of the reversed hazard rate order between equilibrium random variables under formations of some reliability structures. We further investigate a new aging notion based upon the reversed hazard rate order between a random life and its equilibrium version. A nonparametric method is developed to test the exponentiality against such a strict aging property, some numerical results are presented as well. This research was supported by the Program for New Century Excellent Talent in University, China.     

A misspecification test for the higher order co-moments of the factor model

The traditional estimation of higher order co-moments of non-normal random variables by the sample analog of the expectation faces a curse of dimensionality, as the number of parameters increases steeply when the dimension increases. Imposing a factor structure on the process solves this problem; however, it leads to the challenging task of selecting an appropriate factor model. This paper contributes by proposing a test that exploits the following feature: when the factor model is correctly specified, the higher order co-moments of the unexplained return variation are sparse. It recommends a general to specific approach for selecting the factor model by choosing the most parsimonious specification for which the sparsity assumption is satisfied. This approach uses a Wald or Gumbel test statistic for testing the joint statistical significance of the co-moments that are zero when the factor model is correctly specified. The asymptotic distribution of the test is derived. An extensive simulation study confirms the good finite sample properties of the approach. This paper illustrates the practical usefulness of factor selection on daily returns of random subsets of S&P 100 constituents.     

Tests for a simple tree order restriction with application to dose–response studies

Summary.  We propose 'Dunnett-type' test procedures to test for simple tree order restrictions on the means of p independent normal populations. The new tests are based on the estimation procedures that were introduced by Hwang and Peddada and later by Dunbar, Conaway and Peddada. The procedures proposed are also extended to test for 'two-sided' simple tree order restrictions. For non-normal data, nonparametric versions based on ranked data are also suggested. Using computer simulations, we compare the proposed test procedures with some existing test procedures in terms of size and power. Our simulation study suggests that the procedures compete well with the existing procedures for both one-sided and two-sided simple tree alternatives. In some instances, especially in the case of two-sided alternatives or for non-normally distributed data, the gains in power due to the procedures proposed can be substantial.     

Rank regression in order restricted randomised designs

This paper uses order restricted randomised design (ORRD) to create a judgment ranked blocking factor based on available subjective information in a small set of experimental units (EUs). The design then performs a carefully designed randomisation scheme with certain restriction to assign the treatment levels to EUs across these subjective judgment blocks. Such an assignment induces positive dependence among within-set units, and the restrictions on the randomisation translate this positive dependence into a variance reduction technique. We provide a unified theory to analyse the data sets collected from an ORRD. The analysis uses the general framework of rank regression methodology in linear models, with some modification to our randomisation scheme, to estimate regression parameter and to test general linear hypotheses. It is shown that the estimators and test statistics have limiting normal and chi-square distributions regardless the quality of ranking information. A simulation study shows that the asymptotic results remain valid even for relatively small sample sizes. The proposed tests are applied to a clinical trial data set.     

A new distribution-free control scheme based on order statistics

We establish a class of nonparametric Shewhart-type control charts based on a reference sample drawn from the process. The proposed nonparametric control chart takes advantage of the location of two different order statistics of the reference and test sample respectively. The decision rule of the new monitoring scheme is filled out by the number of test observations that are located between the control limits. The general setup of the new class of control charts is presented in detail, while the operating characteristic function is studied for both in- and out-of-control processes. Closed formulae for the evaluation of the alarm rate and the average run length are concluded for plausible shift in the underlying distribution to Lehmann alternatives. Several numerical results, displayed for the new family of nonparametric control charts, depict that the proposed control scheme attains competitive performance.     

Two-sample nonparametric stochastic order inference with an application in plant physiology

In this paper, a new nonparametric methodology is developed for testing whether the changing pattern of a response variable over multiple ordered sub-populations from one treatment group differs with the one from another treatment group. The question is formalized into a nonparametric two-sample comparison problem for the stochastic order among subsamples, through U-statistics with accommodations for zero-inflated distributions. A novel bootstrap procedure is proposed to obtain the critical values with given type I error. Following the procedure, bootstrapped p-values are obtained through simulated samples. It is proven that the distribution of the test statistics is independent from the underlying distributions of the subsamples, when certain sufficient statistics provided. Furthermore, this study also develops a feasible framework for power studies to determine sample sizes, which is necessary in real-world applications. Simulation results suggest that the test is consistent. The methodology is illustrated using a biological experiment with a split-plot design, and significant differences in changing patterns of seed weight between treatments are found with relative small subsample sizes.     

A modified chi-square test for testing equality of two multinomial populations against an order restricted alternative

A modified chi-square test for testing the equality of two multinomial populations against an order restricted alternative in one sample and two sample cases is constructed. The relation between the concepts of dependence by cM-square and stochastic ordering is established, The asymptotic distribution of the test statistic is the chi-bar-square type discussed by Robertson, Wright and Dykstra (1988). Simulations are used to compare the power of this test with the power of the likelihood ratio test of stochastic ordering of the two multinomial populations.     

Comparing two treatments in terms of the likelihood ratio order

In this paper new families of test-statistics are introduced and studied for the problem of comparing two treatments in terms of the likelihood ratio order. The considered families are based on φ-divergence measures and arise as natural extensions of the classical likelihood ratio test and Pearson test-statistics. It is proven that their asymptotic distribution is a common chi-bar random variable. An illustrative example is presented and the performance of these statistics is analysed through a simulation study. Through a simulation study it is shown that, for most of the proposed scenarios adjusted to be small or moderate, some members of this new family of test-statistic display clearly better performance with respect to the power in comparison to the classical likelihood ratio and the Pearson's chi-square test while the exact size remains closed to the nominal size. In view of the exact powers and significance levels, the study also shows that the Wilcoxon test-statistic is not as good as the two classical test-statistics.     

Admissibility of exact conditional tests of stochastic order

Consider the problem of testing for independence against stochastic order in a 2 × J contingency table, with two treatments and J ordered categories. By conditioning on the margins, the null hypothesis becomes simple. Careful selection of the conditional alternative hypothesis then allows the problem to be formulated as one of a class of problems for which the minimal complete class of admissible tests is known. The exact versions of many common tests, such as t-tests and the Smirnov test, are shown to be inadmissible, and thus the non-randomized versions are overly conservative. The proportional hazards and proportional odds tests are shown to be admissible for a given data set and size. A new result allows a proof of the admissibility of convex hull and adaptive tests for all data sets and sizes.