Smooth bootstrap-based confidence intervals for one binomial proportion and difference of two proportions

Constructing confidence intervals (CIs) for a binomial proportion and the difference between two binomial proportions is a fundamental and well-studied problem with respect to the analysis of binary data. In this note, we propose a new bootstrap procedure to estimate the CIs by resampling from a newly developed smooth quantile function in [11 Newcombe, R. G. 1998. Two-sided confidence intervals for the single proportion: Comparison of seven methods. Stat. Med., 17: 857–872. (doi:10.1002/(SICI)1097-0258(19980430)17:8<857::AID-SIM777>3.0.CO;2-E)[Crossref], [PubMed], [Web of Science ®] , [Google Scholar]] for discrete data. We perform a variety of simulation studies in order to illustrate the strong performance of our approach. The coverage probabilities of our CIs in the one-sample setting are superior than or comparable to other well-known approaches. The true utility of our new and novel approach is in the two-sample setting. For the difference of two proportions, our smooth bootstrap CIs provide better coverage probabilities almost uniformly over the interval (?1, 1), particularly in the tail region as compared than other published methods included in our simulation. We illustrate our methodology via an application to several different binary data sets.     

Non parametric confidence intervals for the extinction probability of a Galton–Watson branching process

It is demonstrated that the confidence intervals (CIs) for the probability of eventual extinction and other parameters of a Galton–Watson branching process based upon the maximum likelihood estimators can often have substantially lower coverage when compared to the desired nominal confidence coefficient, especially in small, more realistic sample sizes. The same conclusion holds for the traditional bootstrap CIs. We propose several adjustments to these CIs, which greatly improves coverage in most cases. We also make a correction in an asymptotic variance formula given in Stigler (1971 Stigler, S.M. (1971). The estimation of the probability of extinction and other parameters associated with branching processes. Biometrika 58(3):499–508.[Crossref], [Web of Science ®] , [Google Scholar]). The focus here is on implementation of the CIs which have good coverage, in a wide variety of cases. We also consider expected CI lengths. Some recommendations are made.     

Nonparametric Interval Estimation in One-Way Random-Effects Models

The structural method provided by Hannig et al. (2006 Hannig , J. , Iyer , H. , Patterson , P. (2006). Fiducial generalized confidence intervals. J. Amer. Statist. Assoc. 101:254–269.[Taylor & Francis Online], [Web of Science ®] , [Google Scholar]) has proved to be a useful tool for constructing confidence intervals. However, it is difficult to apply this method to nonparametric problems since the pivotal quantity required in using it exists only in some special parametric models. Based on an extended structural method, this article discusses nonparametric interval estimation for smooth functions of the variances in one-way random-effects models. We use the bootstrap distribution estimator of a statistic to construct an approximate pivotal equation, and prove that the confidence interval derived by the approximate pivotal equation has asymptotically correct coverage probability. Simulation results are presented and show that the normal fiducial interval is not robust against non normality and that the proposed confidence interval has better finite-sample behaviors than the naive interval based on normal approximation.     

Robust BCa–JaB method as a diagnostic tool for linear regression models

The Jackknife-after-bootstrap (JaB) technique originally developed by Efron [8 B. Efron, Jackknife-after-bootstrap standard errors and influence functions, J. R. Stat. Soc. 54 (1992), pp. 83–127. [Google Scholar]] has been proposed as an approach to improve the detection of influential observations in linear regression models by Martin and Roberts [12 M.A. Martin and S. Roberts, Jackknife-after-bootstrap regression influence diagnostics, J. Nonparametr. Stat. 22 (2010), pp. 257–269. doi: 10.1080/10485250903287906[Taylor &; Francis Online], [Web of Science ®] , [Google Scholar]] and Beyaztas and Alin [2 U. Beyaztas and A. Alin, Jackknife-after-bootstrap method for detection of influential observations in linear regression model, Comm. Statist. Simulation Comput. 42 (2013), pp. 1256–1267. doi: 10.1080/03610918.2012.661908[Taylor &; Francis Online], [Web of Science ®] , [Google Scholar]]. The method is based on the use of percentile-method confidence intervals to provide improved cut-off values for several single case-deletion influence measures. In order to improve JaB, we propose using robust versions of Efron [7 B. Efron, Better bootstrap confidence intervals, J. Amer. Statist. Assoc. 82 (1987), pp. 171–185. doi: 10.1080/01621459.1987.10478410[Taylor &; Francis Online], [Web of Science ®] , [Google Scholar]]’s bias-corrected and accelerated (BCa) bootstrap confidence intervals. In this study, the performances of robust BCa–JaB and conventional JaB methods are compared in the cases of DFFITS, Welsch's distance and modified Cook's distance influence diagnostics. Comparisons are based on both real data examples and through a simulation study. Our results reveal that under a variety of scenarios, our proposed method provides more accurate and reliable results, and it is more robust to masking effects.     

Wild Bootstrap of the Sample Mean in the Infinite Variance Case

It is well known that the standard independent, identically distributed (iid) bootstrap of the mean is inconsistent in a location model with infinite variance (α-stable) innovations. This occurs because the bootstrap distribution of a normalised sum of infinite variance random variables tends to a random distribution. Consistent bootstrap algorithms based on subsampling methods have been proposed but have the drawback that they deliver much wider confidence sets than those generated by the iid bootstrap owing to the fact that they eliminate the dependence of the bootstrap distribution on the sample extremes. In this paper we propose sufficient conditions that allow a simple modification of the bootstrap (Wu, 1986 Wu , C. F. J. ( 1986 ). Jackknife, bootstrap, and other resampling methods . Annals of Statistics 14 : 1261 – 1295 .[Crossref], [Web of Science ®] , [Google Scholar]) to be consistent (in a conditional sense) yet to also reproduce the narrower confidence sets of the iid bootstrap. Numerical results demonstrate that our proposed bootstrap method works very well in practice delivering coverage rates very close to the nominal level and significantly narrower confidence sets than other consistent methods.     

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.     

Confidence Intervals from Stochastic Approximation

We propose a nonparametric method of constructing confidence interval for a scalar parameter from stochastic approximation through the efficient Robbins–Monro procedure proposed by Joseph (2004 Joseph, V.R. (2004). Efficient Robbins–Monro procedure for binary data. Biometrika 91:461–470.[Crossref], [Web of Science ®] , [Google Scholar]). Unlike the bootstrap method where the number of resampling is fixed in advance, the proposed procedure iteratively searches the endpoints in an optimal way such that the convergence is fast and the coverage is obtained accurately. Simulation and real data application illustrate its superiority over the usual Robbins–Monro procedure and common bootstrap methods.     

Large sample confidence intervals for the skewness parameter of the skew-normal distribution based on Fisher's transformation

The skew-normal model is a class of distributions that extends the Gaussian family by including a skewness parameter. This model presents some inferential problems linked to the estimation of the skewness parameter. In particular its maximum likelihood estimator can be infinite especially for moderate sample sizes and is not clear how to calculate confidence intervals for this parameter. In this work, we show how these inferential problems can be solved if we are interested in the distribution of extreme statistics of two random variables with joint normal distribution. Such situations are not uncommon in applications, especially in medical and environmental contexts, where it can be relevant to estimate the distribution of extreme statistics. A theoretical result, found by Loperfido [7 Loperfido, N. 2002. Statistical implications of selectively reported inferential results. Statist. Probab. Lett., 56: 13–22. [Crossref], [Web of Science ®] , [Google Scholar]], proves that such extreme statistics have a skew-normal distribution with skewness parameter that can be expressed as a function of the correlation coefficient between the two initial variables. It is then possible, using some theoretical results involving the correlation coefficient, to find approximate confidence intervals for the parameter of skewness. These theoretical intervals are then compared with parametric bootstrap intervals by means of a simulation study. Two applications are given using real data.     

Shortest Expected Length Confidence Interval for the Power of the t-Test

Tarasińska (2005 Tarasińska , J. ( 2005 ). Confidence intervals for the power of student's t-test . Statistics & Probability Letters 73 ( 2 ): 125 – 130 .[Crossref], [Web of Science ®] , [Google Scholar]) considered a method to construct the shortest length confidence interval on the power of the t-test using a confidence interval for the population standard deviation in the non centrality parameter. Gilliland and Li (2008 Gilliland , D. , Li , M. ( 2008 ). A note on confidence intervals for the power of t-test . Statistics & Probability Letters 78 ( 5 ): 488 – 489 .[Crossref], [Web of Science ®] , [Google Scholar]) used simulations to show that this confidence interval has less than the nominal coverage, particularly in small samples. We propose to find the shortest expected length confidence interval for the power of the t-test by accounting for the variation in the sample standard deviation, and provide the necessary constants for its implementation for some selected sample and shift sizes. It is seen that the proposed interval is reasonably robust to the specification of the population standard deviation and maintains the nominal coverage.     

A Bootstrap Stationarity Test for Predictive Regression Invalidity

In order for predictive regression tests to deliver asymptotically valid inference, account has to be taken of the degree of persistence of the predictors under test. There is also a maintained assumption that any predictability in the variable of interest is purely attributable to the predictors under test. Violation of this assumption by the omission of relevant persistent predictors renders the predictive regression invalid, and potentially also spurious, as both the finite sample and asymptotic size of the predictability tests can be significantly inflated. In response, we propose a predictive regression invalidity test based on a stationarity testing approach. To allow for an unknown degree of persistence in the putative predictors, and for heteroscedasticity in the data, we implement our proposed test using a fixed regressor wild bootstrap procedure. We demonstrate the asymptotic validity of the proposed bootstrap test by proving that the limit distribution of the bootstrap statistic, conditional on the data, is the same as the limit null distribution of the statistic computed on the original data, conditional on the predictor. This corrects a long-standing error in the bootstrap literature whereby it is incorrectly argued that for strongly persistent regressors and test statistics akin to ours the validity of the fixed regressor bootstrap obtains through equivalence to an unconditional limit distribution. Our bootstrap results are therefore of interest in their own right and are likely to have applications beyond the present context. An illustration is given by reexamining the results relating to U.S. stock returns data in Campbell and Yogo (2006 Campbell, J. Y. and Yogo, M. (2006), “Efficient Tests of Stock Return Predictability,” Journal of Financial Economics, 81, 27–60.[Crossref], [Web of Science ®] , [Google Scholar]). Supplementary materials for this article are available online.     

On computing standard errors for marginal structural Cox models

In recent decades, marginal structural models have gained popularity for proper adjustment of time-dependent confounders in longitudinal studies through time-dependent weighting. When the marginal model is a Cox model, using current standard statistical software packages was thought to be problematic because they were not developed to compute standard errors in the presence of time-dependent weights. We address this practical modelling issue by extending the standard calculations for Cox models with case weights to time-dependent weights and show that the coxph procedure in R can readily compute asymptotic robust standard errors. Through a simulation study, we show that the robust standard errors are rather conservative, though corresponding confidence intervals have good coverage. A second contribution of this paper is to introduce a Cox score bootstrap procedure to compute the standard errors. We show that this method is efficient and tends to outperform the non-parametric bootstrap in small samples.     

The Block-Block Bootstrap for Time Series

This article proves that the block-block bootstrap of Andrews (2004 Andrews , D. W. K. ( 2004 ). The block-block bootstrap: improved asymptotic refinements . Econometrica 72 ( 3 ): 673 – 700 .[Crossref], [Web of Science ®] , [Google Scholar]) can be helpful to provide asymptotic refinements for the GMM estimator when autocorrelation structures of moment functions are unknown (i.e., incorporating the HAC covariance matrix) and when we allow for statistics that are inefficient. The asymptotic refinements of this block-block bootstrap in the time series context are shown to exist with the use of less restricted kernels than in the block bootstrap in Inoue and Shintani (2006 Inoue , A. , Shintani , M. ( 2006 ). Bootstrapping GMM estimators for time series . J. Econometrics 113 : 531 – 555 .[Crossref] , [Google Scholar]), since they do not require to have a characteristic exponent larger than 2. The procedure allows to apply in practice kernels that guarantee that the HAC covariance matrix estimator is positive semidefinite, and to get asymptotic refinements at the same time.     

Exact confidence intervals for randomized response strategies

For surveys with sensitive questions, randomized response sampling strategies are often used to increase the response rate and encourage participants to provide the truth of the question while participants' privacy and confidentiality are protected. The proportion of responding ‘yes’ to the sensitive question is the parameter of interest. Asymptotic confidence intervals for this proportion are calculated from the limiting distribution of the test statistic, and are traditionally used in practice for statistical inference. It is well known that these intervals do not guarantee the coverage probability. For this reason, we apply the exact approach, adjusting the critical value as in [10 J. Frey and A. Pérez, Exact binomial confidence intervals for randomized response, Amer. Statist.66 (2012), pp. 8–15. Available at http://dx.doi.org/10.1080/00031305.2012.663680.[Taylor &; Francis Online], [Web of Science ®] , [Google Scholar]], to construct the exact confidence interval of the proportion based on the likelihood ratio test and three Wilson-type tests. Two randomized response sampling strategies are studied: the Warner model and the unrelated model. The exact interval based on the likelihood ratio test has shorter average length than others when the probability of the sensitive question is low. Exact Wilson intervals have good performance in other cases. A real example from a survey study is utilized to illustrate the application of these exact intervals.     

EXPLICIT NONPARAMETRIC CONFIDENCE INTERVALS FOR THE VARIANCE WITH GUARANTEED COVERAGE

ABSTRACT

In this paper, we provide a method for constructing confidence intervals for the variance which exhibits guaranteed coverage probability for any sample size, uniformly over a wide class of probability distributions. In contrast, standard methods achieve guaranteed coverage only in the limit for a fixed distribution or for any sample size over a very restrictive (parametric) class of probability distributions. Of course, it is impossible to construct effective confidence intervals for the variance without some restriction, due to a result of Bahadur and Savage.[1] Bahadur, R. and Savage, L. 1956. The Nonexistence of Certain Statistical Procedures in Nonparametric Problems. Annals of Mathematical Statistics, 25: 1115–1122.  [Google Scholar] However, it is possible if the observations lie in a fixed compact set. We also consider the case of lower confidence bounds without any support restriction. Our method is based on the behavior of the variance over distributions that lie within a Kolmogorov–Smirnov confidence band for the underlying distribution. The method is a generalization of an idea of Anderson,[2] Anderson, T. 1967. Confidence Limits for the Expected Value of an Arbitrary Bounded Random Variable with a Continuous Distribution Function. Bull. ISI, 43: 249–251.  [Google Scholar] who considered only the case of the mean; it applies to very general parameters, and particularly the variance. While typically it is not clear how to compute these intervals explicitly, for the special case of the variance we provide an algorithm to do so. Asymptotically, the length of the intervals is of order n ^?/2 (in probability), so that, while providing guaranteed coverage, they are not overly conservative. A small simulation study examines the finite sample behavior of the proposed intervals.