Normal Approximation Approach Research Articles

Confidence Intervals Based on the Difference of Medians for Independent Log-Normal Distributions

In this paper, we study the inferences of the difference of medians for two independent log-normal distributions. These methods include traditional methods such as the parametric bootstrap approach, the normal approximation approach, the method of variance estimates recovery approach, and the generalized confidence interval approach. The simultaneous confidence intervals for the difference in the median for more than two independent log-normal distributions are also discussed. Our simulation studies evaluate the performances of the proposed confidence intervals in terms of coverage probabilities and average lengths. We find that the parametric bootstrap approach would be a suitable choice for smaller sample sizes for the two independent distributions and multiple independent distributions. However, the method of variance estimates recovery and normal approximation approaches are alternative competitors for constructing simultaneous confidence intervals, especially when the populations have large variance. We also include two practical applications demonstrating the use of the techniques on observed data, where one data set works for the PM2.5 mass concentrations in Bangkapi and Dindaeng in Thailand and the other data came from the study of nitrogen-bound bovine serum albumin produced by three groups of diabetic mice. Both applications show that the confidence intervals from the parametric bootstrap approach have the smallest length.

Multivariate Bayesian Arm-Based Network Meta-Analysis of Pharmacological Interventions for the Treatment of Acute Bipolar Mania in Adults.

In a network meta-analysis (NMA), multiple treatments can be compared simultaneously by aggregating pieces of evidence from direct as well as indirect treatment comparisons in different randomized controlled trials (RCTs). Conventional NMA are performed using a normal approximation approach and can be applied for arm-level binary outcome data as well. This study aimed to estimate the treatment effects within a Bayesian framework using a binomial likelihood for a multivariate NMA model. The dataset consists of 57 RCTs comparing the effect of ten pharmacological drugs and a placebo for acute bipolar mania in adults. The binary outcomes of interest were treatment response and all-cause dropouts measured three weeks from the baseline. Binomial distribution was adopted for the number of events and the probability of event occurrence modeled on the logit scale. Jeffrey's Beta prior was considered for the heterogeneity and inconsistency of standard deviation (SD) parameters. Cholesky and spherical decomposition strategies were adopted for the between-study variance-covariance matrix. Deviance information criterion (DIC) indices were computed to determine the model fit. All results pertaining to Markov chain Monte Carlo simulations and all analyses were carried out in WinBUGS software. The estimated common heterogeneity SDs were similar, and the DIC values did not provide any evidence for superiority between the two decomposition strategies. The correlation (95% credible interval) between the outcomes was estimated as -0.31 (-0.71, -0.02) and -0.37 (-0.73, -0.03) for the Cholesky and spherical decompositions, respectively. Gelman-Rubin convergence statistics were stable, and Monte Carlo errors for all the parameters were around 0.005. Overall, olanzapine, paliperidone, and quetiapine were both significantly more effective and acceptable than a placebo when both the study outcomes were considered simultaneously. The findings favoring olanzapine, paliperidone, and quetiapine possess an excellent concordance with the one adopted in clinical practice, and the Canadian Network for Mood and Anxiety Treatments and Royal Australian and New Zealand College of Psychiatrists guidelines recommend these as first-line drugs for treating bipolar disorder.

Empirical likelihood analysis for accelerated failure time model using length-biased data

Length-biased data arise when the probability of observing a subject in a sample is proportional to its corresponding value. The statistical tools that should be applied to analyse such sets of data differ from those used for random samples. Ignoring this issue in an investigation results in a biased inference. In this paper, we propose an empirical likelihood-based method to draw inference on covariate effects in an accelerated failure time model while the observations are subject to length-bias. The asymptotic distribution of the empirical log-likelihood ratio statistic is derived to be a weighted sum of independent chi-square distributions. Hence, we then extend the results by an exploration into the adjusted empirical likelihood. We derive an asymptotic standard chi-square distribution for the adjusted log-likelihood statistic. The limiting distributions are applied to obtain confidence regions for the regression parameters. A simulation study is carried out to evaluate and compare the performance of the proposed procedures with an existing method based on the normal approximation approach. Finally, the procedures are illustrated by modelling the regression parameter and estimating confidence intervals for a set of real data on widths of shrubs.

Jackknife empirical likelihood of error variance for partially linear varying-coefficient model with missing covariates

In this paper, we apply the profile least-square method and inverse probability weighted method to define estimation of the error variance in partially linear varying-coefficient model when the covariates are missing at random. At the same time, we construct a jackknife estimator and jackknife empirical likelihood (JEL) statistic of the error variance, respectively. It is proved that the proposed estimators are asymptotical normality and the JEL statistic admits a limiting standard chi-square distribution. A simulation study is conducted to compare the JEL method with the normal approximation approach in terms of coverage probabilities and average interval lengths, and a comparison of the proposed estimators is done based on sample means, biases and mean square errors under different settings. Subsequently, a real data set is analyzed for illustration of the proposed methods.

Higher-order normal approximation approach for highly reliable system assessment

In this study, the issue of system reliability assessment (SRA) based on component failure data is considered. In industrial statistics, the delta method has become a popular approach for confidence interval approximation. However, for high reliability systems, usually the assessment is confronted with very limited component sample size, variant multi-parameter lifetime models, and complex system structure. Along with strict requirement on assessment accuracy and computational efficiency, existing approaches barely work under these circumstances. In this article, a normal approximation approach is proposed for determining the lower confidence limit of system reliability using components’ time-to-failure data. The polynomial adjustment method is adopted to construct higher-order approximate confidence limit. The main contribution of this work is constructing an integrated methodology for SRA. Specifically, a reliability-based Winterbottom-extended Cornish-Fisher (R-WCF) expansion method for log-location-scale family is elaborated. The proposed methodology exceeds the efficient limitation of Cramer Rao’s theory. Numerical studies are conducted to illustrate the effectiveness of the proposed approach, and results show that the R-WCF approach is more efficient than the delta method for highly reliable system assessment, especially with ultra-small sample size. Supplementary materials are available for this article. Go to the publisher’s online edition of IISE Transactions.

Sample Sizes for Estimating Differences in Proportions—Can We Keep Things Simple?

Sample sizes for studies that aim to estimate differences in proportions are often calculated using a confidence interval approach. In particular, the methods advocated by Day (1988) and Bristol (1989), based on asymptotic normal approximations, are commonly used. We compare these to the Wilson score approach discussed by Newcombe (1998) and show, except for extreme values, that the Day method gives results close to those of Wilson score method, while that of Bristol is consistently higher. We argue that the asymptotic normal approximation approach of Day serves as a better guide for quick calculations of sample sizes.

Bias-corrected smoothed score function for single-index models

We in this paper investigate smoothed score function based confidence regions for parameters in single-index models. Because a plug-in estimator of nonparametric link function causes the bias of smoothed score function to be non-negligible, the limit of the score function is asymptotically normal with a non-zero mean due to the slow convergence rate of nonparametric estimation. A bias-corrected smoothed score function is recommended for achieving centered normal limit without under-smoothing or high order kernel, and then the confidence region can be constructed by chi-square distribution. Simulation studies are carried out to assess the performance of bias-corrected local likelihood, and to compare with normal approximation approach.

Interim treatment selection using the normal approximation approach in clinical trials

We consider a study starting with two treatment groups and a control group with a planned interim analysis. The inferior treatment group will be dropped after the interim analysis, and only the winning treatment and the control will continue to the end of the study. This 'Two-Stage Winner Design' is based on the concepts of multiple comparison, adaptive design, and winner selection. In a study with such a design, there is less multiplicity, but more adaptability if the interim selection is performed at an early stage. If the interim selection is performed close to the end of the study, the situation becomes the conventional multiple comparison where Dunnett's method may be applied. The unconditional distribution of the final test statistic from the 'winner' treatment is no longer normal, the exact distribution of which is provided in this paper, but numerical integration is needed for its calculation. To avoid complex computations, we propose a normal approximation approach to calculate the type I error, the power, the point estimate, and the confidence intervals. Due to the well understood and attractive properties of the normal distribution, the 'Winner Design' can be easily planned and adequately executed, which is demonstrated by an example. We also provide detailed discussion on how the proposed design should be practically implemented by optimizing the timing of the interim look and the probability of winner selection.

Interval Estimation for True Raw and Scale Scores Under the Binomial Error Model

Assuming errors of measurement are distributed binomially, this article reviews various procedures for constructing an interval for an individual’s true number-correct score; presents two general interval estimation procedures for an individual’s true scale score (i.e., normal approximation and endpoints conversion methods); compares various interval estimation procedures through a computer simulation study; and provides some practical guidelines for use of the interval estimation procedures. To examine the effects of different types of scale scores, three nonlinearly transformed scale scores are employed. The conditional confidence intervals using conditional standard errors of measurement are recommended over the traditional confidence intervals using the overall standard error of measurement. For raw scores, the score confidence intervals, in general, tend to provide actual coverage probabilities that are closest to the nominal level. Results for scale score intervals seem to favor the endpoints conversion method using the true-score conversions over the normal approximation approach.