Noninformative priors for linear combinations of normal means with unequal variances

For normal populations with unequal variances, we develop matching priors and reference priors for a linear combination of the means. Here, we find three second-order matching priors: a highest posterior density (HPD) matching prior, a cumulative distribution function (CDF) matching prior, and a likelihood ratio (LR) matching prior. Furthermore, we show that the reference priors are all first-order matching priors, but that they do not satisfy the second-order matching criterion that establishes the symmetry and the unimodality of the posterior under the developed priors. The results of a simulation indicate that the second-order matching prior outperforms the reference priors in terms of matching the target coverage probabilities, in a frequentist sense. Finally, we compare the Bayesian credible intervals based on the developed priors with the confidence intervals derived from real data.     

Response‐adaptive designs for binary responses: How to offer patient benefit while being robust to time trends?

Response‐adaptive randomisation (RAR) can considerably improve the chances of a successful treatment outcome for patients in a clinical trial by skewing the allocation probability towards better performing treatments as data accumulates. There is considerable interest in using RAR designs in drug development for rare diseases, where traditional designs are not either feasible or ethically questionable. In this paper, we discuss and address a major criticism levelled at RAR: namely, type I error inflation due to an unknown time trend over the course of the trial. The most common cause of this phenomenon is changes in the characteristics of recruited patients—referred to as patient drift. This is a realistic concern for clinical trials in rare diseases due to their lengthly accrual rate. We compute the type I error inflation as a function of the time trend magnitude to determine in which contexts the problem is most exacerbated. We then assess the ability of different correction methods to preserve type I error in these contexts and their performance in terms of other operating characteristics, including patient benefit and power. We make recommendations as to which correction methods are most suitable in the rare disease context for several RAR rules, differentiating between the 2‐armed and the multi‐armed case. We further propose a RAR design for multi‐armed clinical trials, which is computationally efficient and robust to several time trends considered.     

Development of predictive signatures for treatment selection in precision medicine with survival outcomes

For survival endpoints in subgroup selection, a score conversion model is often used to convert the set of biomarkers for each patient into a univariate score and using the median of the univariate scores to divide the patients into biomarker‐positive and biomarker‐negative subgroups. However, this may lead to bias in patient subgroup identification regarding the 2 issues: (1) treatment is equally effective for all patients and/or there is no subgroup difference; (2) the median value of the univariate scores as a cutoff may be inappropriate if the sizes of the 2 subgroups are differ substantially. We utilize a univariate composite score method to convert the set of patient's candidate biomarkers to a univariate response score. We propose applying the likelihood ratio test (LRT) to assess homogeneity of the sampled patients to address the first issue. In the context of identification of the subgroup of responders in adaptive design to demonstrate improvement of treatment efficacy (adaptive power), we suggest that subgroup selection is carried out if the LRT is significant. For the second issue, we utilize a likelihood‐based change‐point algorithm to find an optimal cutoff. Our simulation study shows that type I error generally is controlled, while the overall adaptive power to detect treatment effects sacrifices approximately 4.5% for the simulation designs considered by performing the LRT; furthermore, the change‐point algorithm outperforms the median cutoff considerably when the subgroup sizes differ substantially.     

Evaluating the causal economic impacts of transport investments: evidence from the Madrid–Barcelona high speed rail corridor

This paper evaluates economic impacts arising from the introduction of high-speed rail (HSR) between Madrid and Barcelona. Using difference-in-differences estimation we estimate an average treatment effect for provinces with stops on the HSR line of 2.4% for economic output, 3.3% for numbers of firms, and 1.1% for labour productivity. We complement our DID results with a synthetic control analysis for Lleida and Tarragona, two provinces that we argue were assigned HSR stations largely due to their incidental location. We find that both the number of firms and labour productivity are substantially higher in these provinces than in their synthetic counterparts.     

Ethical Numbers: Ethics Training in U.S. Graduate Statistics Programs, 2013–2014

As important members of research teams, statisticians bear an ethical responsibility to analyze, interpret, and report data honestly and objectively. One way of reinforcing ethical responsibilities is through required courses covering a variety of ethics-related topics at the graduate level. We assessed ethics requirements for graduate-level statistics training programs in the United States for the 2013–2014 academic year using the websites of 88 universities, examining 103 biostatistics programs, and 136 statistics degree programs. We categorized programs’ ethics training requirements as required or not required. Thirty-one (35.1%) universities required an ethics course for at least some degree students. Sixty-two (25.5%) degree programs required an ethics course for at least some students. The majority (77.4%) of required courses were worth 0 or 1 credit. Of the 177 programs without an ethics requirement, 19 (10.7%) listed an ethics elective. Although a single ethics course is insufficient for instilling an ethical approach to science, degree programs that model expectations through coursework point to the value of ethics in science. More training programs should prepare statisticians to consider the ethical dimensions of their work through required coursework. Supplementary materials for this article are available online.     

Climate regime shift detection with a trans‐dimensional,sequential Monte Carlo,variational Bayes method

We present an application study which exemplifies a cutting edge statistical approach for detecting climate regime shifts. The algorithm uses Bayesian computational techniques that make time‐efficient analysis of large volumes of climate data possible. Output includes probabilistic estimates of the number and duration of regimes, the number and probability distribution of hidden states, and the probability of a regime shift in any year of the time series. Analysis of the Pacific Decadal Oscillation (PDO) index is provided as an example. Two states are detected: one is associated with positive values of the PDO and presents lower interannual variability, while the other corresponds to negative values of the PDO and greater variability. We compare this approach with existing alternatives from the literature and highlight the potential for ours to unlock features hidden in climate data.     

The effects of data input errors on subsequent statistical inference

Data input errors can potentially affect statistical inferences, but little research has been published to date on this topic. In the present paper, we report the effect of data input errors on the statistical inferences drawn about population parameters in an empirical study involving 280 students from two Polish universities, namely the Warsaw University of Life Sciences – SGGW and the University of Information Technology and Management in Rzeszow. We found that 28% of the students committed at least one data error. While some of these errors were small and did not have any real effect, a few of them had substantial effects on the statistical inferences drawn about the population parameters.     

Heteroscedasticity and Distributional Assumptions in the Censored Regression Model

Data censoring causes ordinary least-square estimators of linear models to be biased and inconsistent. The Tobit estimator yields consistent estimators in the presence of data censoring if the errors are normally distributed. However, nonnormality or heteroscedasticity results in the Tobit estimators being inconsistent. Various estimators have been proposed for circumventing the normality assumption. Some of these estimators include censored least absolute deviations (CLAD), symmetrically censored least-square (SCLS), and partially adaptive estimators. CLAD and SCLS will be consistent in the presence of heteroscedasticity; however, SCLS performs poorly in the presence of asymmetric errors. This article extends the partially adaptive estimation approach to accommodate possible heteroscedasticity as well as nonnormality. A simulation study is used to investigate the estimators’ relative performance in these settings. The partially adaptive censored regression estimators have little efficiency loss for censored normal errors and appear to outperform the Tobit and semiparametric estimators for nonnormal error distributions and be less sensitive to the presence of heteroscedasticity. An empirical example is considered, which supports these results.     

Assessing direct and indirect seasonal decomposition in state space

The problem of whether seasonal decomposition should be used prior to or after aggregation of time series is quite old. We tackle the problem by using a state-space representation and the variance/covariance structure of a simplified one-component model. The variances of the estimated components in a two-series system are compared for direct and indirect approaches and also to a multivariate method. The covariance structure between the two time series is important for the relative efficiency. Indirect estimation is always best when the series are independent, but when the series or the measurement errors are negatively correlated, direct estimation may be much better in the above sense. Some covariance structures indicate that direct estimation should be used while others indicate that an indirect approach is more efficient. Signal-to-noise ratios and relative variances are used for inference.