Testing for a treatment effect in a heterogeneous population: a modified sign-test statistic and a leapfrog statistic

This paper proposes two non-parametric statistics that test for a treatment effect in a heterogeneous population. In the model considered, data on two examinations for both a control and a treatment group are needed to perform the test. The model allows for individual (fixed) effects that may be correlated with the choice of treatment. In addition, the model allows for an unspecified, monotonic transformation of the response variable. The techniques are illustrated by testing whether high levels of unemploymentbenefit eligibility affect the consumption patterns of unemployed American workers.     

Testing for a Unit Root in a Time Series With a Changing Mean

This study considers testing for a unit root in a time series characterized by a structural change in its mean level. My approach follows the “intervention analysis” of Box and Tiao (1975) in the sense that I consider the change as being exogenous and as occurring at a known date. Standard unit-root tests are shown to be biased toward nonrejection of the hypothesis of a unit root when the full sample is used. Since tests using split sample regressions usually have low power, I design test statistics that allow the presence of a change in the mean of the series under both the null and alternative hypotheses. The limiting distribution of the statistics is derived and tabulated under the null hypothesis of a unit root. My analysis is illustrated by considering the behavior of various univariate time series for which the unit-root hypothesis has been advanced in the literature. This study complements that of Perron (1989), which considered time series with trends.     

Statistical games against a prophet-proof of a minimax conjecture

“Prophet theory” quantifies the price a statistician has to pay for his lack of information in stochastic sequences. In a recent paper, Schmitz (1991) gave a game-theoretical interpretation of this situation and he formulated in particular a minimax conjecture for the difference case. In this note we prove that conjecture and, moreover, present minimax ran domized stopping times (minimax procedures for the statistician).     

Managing a Comprehensive Cost-per-Use Project in a Large Academic Library

Abstract

This article reports the findings of a study to evaluate title-level cost per use (CPU) for most recurring resources at The University of Alabama Libraries. The study relied on cost and usage data from FY2016. The purpose of the project was to identify titles for cancellation and to establish CPU for most titles ordered individually and those within packages for use in future contract negotiations. The analysis began with a list of all recurring titles having a cost of $1,000 or more. The review involved 1,035 titles with an initial objective to highlight all titles with a CPU of $200 or more. Titles exceeding this CPU threshold received additional review, with several identified for cancellation. This article presents the findings and provides an overview of the process used to collect CPU on a large scale. The article outlines some of the challenges faced by libraries when trying to align data that are stored in various silos and the limitations for various platforms and software used when collecting CPU data. The findings from this study formed the basis for a LibGuide developed and used to raise awareness of the need for cancellations. The data also assisted with future contract negotiations.     

Target detection from a classical and a Bayesian viewpoint

Two approaches have been used for designing spatial surveys to detect a target. The classical approach controls the probability of missing a target that exists; a Bayesian approach controls the probability that a target exists given that none was seen. In both cases, information about the likely size of the target can reduce sampling requirements. In this paper, previous results are summarized and then used to assess the risk that Roman remains could be present at sites scheduled for development in Greater London.     

Testing the disturbance variance after a pre-test for a linear hypothesis on coefficients in a linear regression

In this paper, we examine the sampling performance of a two-stage test which consists of a pre-test for a linear hypothesis on regression coeffiecients followed by a main-test for a disturbance variance in a linear regression. It is shown that the actual size of the two-stage test can be well-controlled around the normal size if the suggested sizes presented in this paper are used in the pre-test. It is also shown that the two-stage test when the suggested sizes are used in the preferable to the usual test for the disturbance variable which incorporates no pre-test in terms of the power.     

Estimation of size and mean of a sensitive quantitative variable for a sub-group of a population

In the present paper, a new problem in survey sampling has been considered and an attempt has bcen made to find its solution. The procedures available in literature deal with the estimation of mean of a sensitive quantitative variable for the whole population. Sometimes the investigator may be interested in estimating the mean of a sensitive quantitative variable for a sub-group of the population. At the same time, he/she may also be interested in estimating the size of this sub-group. In this paper, we have obtained estimators of these parameters. Approximate bias and variance expressions are also obtained for these estimators.     

Testing a main effect in a three factor mixed model

In the present paper, the author has considered and compared the power of three test procedures, based on two preliminary tests of significance (PTS), for testing a main effect in a three factor factorial experiment. Davenport and Webster (1973) type test statistics are used in the final test. Recommendation for a suitable size of PTS has been made so that the power of a test procedure is adequately high.     

Crossing probabilities for a square root boundary by a bessel process

Probabilities for Wiener and Bessel processes to cross a square root boundary are calculated and the Mellin transform of the distribution of the associated stopping time is given.The transform is inverted and a table of critical values for the crossing probability as a function of the observation time is included. Also an analytic expression for the crossing probability is computed for the case of constant drift.     

Estimating a distribution for each category of a sensitive variate

The use of a Randomized Response (RR) design makes it possible to estimate the distribution of a sensitive variate. In this paper, the estimation of the distribution of a non-sensitive variate for each category of a sensitive variate is considered for the case where data on the sensitive variate is obtained by use of an RR procedure. Simple estimators are developed without making any distributional assumptions about the non-sensitive variate. However, if distributional assumptions are made, it is shown that the EM algorithm may be used to compute Maximum Likelihood estimates. Computational comparisons of the estimators, using simulation, indicate that the simple estimators perform well, particularly for large sample sizes.     

Resolving a Multi-Million Dollar Contract Dispute With a Latin Square

The City of New York negotiated a dispute over the performance of new garbage trucks purchased from a vehicle manufacturer. The dispute concerned the fulfillment of a specification in the purchase contract that the trucks load a minimum full-load of 12.5 tons of household refuse. On behalf of the City, but in cooperation with the manufacturer, the City's Department of Sanitation and consulting statisticians tested fulfillment of the contract specification, employing a Latin Square design for routing trucks. We present the classical analysis using a linear model and analysis of variance. We also show how fixed, mixed, and random effect models are useful in analyzing the results of the test. Finally, we take a Bayesian perspective to demonstrate how the information from the data overcomes the difference between the prior densities of the city and the manufacturer for the load capacities of the trucks to result in much closer posterior densities. This procedure might prove useful in similar negotiations. Supplementary material including the data and R code for computations in the article are available online.     

Inference for comparing a multinomial distribution with a known standard

We propose a new type of stochastic ordering which imposes a monotone tendency in differences between one multinomial probability and a known standard one. An estimation procedure is proposed for the constrained maximum likelihood estimate, and then the asymptotic null distribution is derived for the likelihood ratio test statistic for testing equality of two multinomial distributions against the new stochastic ordering. An alternative test is also discussed based on Neyman modified minimum chi-square estimator. These tests are illustrated with a set of heart disease data.     

Predicting a Future Median Life through a Power Transformation

A simple and unified prediction interval (PI) for the median of a future lifetime can be obtained through a power transformation. This interval usually possesses the correct coverage, at least asymptotically, when the transformation is known. However, when the transformation is unknown and is estimated from the data, a correction is required. A simple correction factor is derived based on large sample theory. Simulation shows that the unified PI after correction performs well. When compared with the existing frequentist PI's, it shows an equivalent or a better performance in terms of coverage probability and average length of the interval. Its nonparametric aspect and the ease of usage make it very attractive to practitioners. Real data examples are provided for illustration.     

Bayesian Analysis of a Queueing System with a Long-Tailed Arrival Process

Internet traffic data is characterized by some unusual statistical properties, in particular, the presence of heavy-tailed variables. A typical model for heavy-tailed distributions is the Pareto distribution although this is not adequate in many cases. In this article, we consider a mixture of two-parameter Pareto distributions as a model for heavy-tailed data and use a Bayesian approach based on the birth-death Markov chain Monte Carlo algorithm to fit this model. We estimate some measures of interest related to the queueing system k-Par/M/1 where k-Par denotes a mixture of k Pareto distributions. Heavy-tailed variables are difficult to model in such queueing systems because of the lack of a simple expression for the Laplace Transform (LT). We use a procedure based on recent LT approximating results for the Pareto/M/1 system. We illustrate our approach with both simulated and real data.     

Incorporating historical controls using a random-effects model with a normal prior

Employing a variance stabilization transformation, a random-effects model with a normal prior is proposed for incorporating historical controls in the estimation of dose-response relationships. The advantages and disadvantages of maximum likelihood and empirical Bayes estimators are discussed. A two-stage method is introduced for the prudent use of historical controls.     

Detection of a pair of outliers in a sample from a Gumbel distribution with known scale parameter

Test procedures on outlier detection problems for Gumbel distribution are rarely available. Hence, a test statistic is proposed here for detection of a pair of upper and lower outliers from a Gumbel distribution with known scale parameter. The critical values of the statistic are obtained and some examples are also given to highlight the use of the statistic. The advantage of the proposed statistic is that the scale parameter, though assumed to be known is not explicitly involved in the determination of the critical values.     

Detection of a changepoint,a mean-shift accompanied with a trend change,in short time-series with autocorrelation

In this study, a changepoint model, which can detect either a mean shift or a trend change when accounting for autocorrelation in short time-series, was investigated with simulations and a new method is proposed. The changepoint hypotheses were tested using a likelihood ratio test. The test statistic does not follow a known distribution and depends on the length of the time-series and the autocorrelation. The results imply that it is not possible to detect autocorrelation and that the estimate of the autocorrelation parameter is biased. It is therefore recommended to use critical values from the empirical distribution for a fixed autocorrelation.     

Error rate for testing a contrast after a significant F test

In statistical applications an experimenter often tests a particular contrast after a significant F test for the equality of means. This paper evaluates the overall error rate for this testing .procedure.     

On a perturbed MAP risk model under a threshold dividend strategy

In this paper, we consider a perturbed risk model where the claims arrive according to a Markovian arrival process (MAP) under a threshold dividend strategy. We derive the integro-differential equations for the Gerber–Shiu expected discounted penalty function and the moments of total dividend payments until ruin, obtain the analytical solutions to these equations, and give numerical examples to illustrate our main results. We also get a matrix renewal equation for the Gerber–Shiu function, and present some asymptotic formulas for the Gerber–Shiu function when the claim size distributions are heavy-tailed.     

Admissibility of a test procedure in a conditionally specified random model

Admissibility of a test procedure is a desirable property though not a compelling one; while lack of it often results in discarding a given test procedure. In this paper we have proved necessary conditions for admissibility for a test procedure.