Reverse Regression and Employment Discrimination

The importance of interval forecasts is reviewed. Several general approaches to calculating such forecasts are described and compared. They include the use of theoretical formulas based on a fitted probability model (with or without a correction for parameter uncertainty), various “approximate” formulas (which should be avoided), and empirically based, simulation, and resampling procedures. The latter are useful when theoretical formulas are not available or there are doubts about some model assumptions. The distinction between a forecasting method and a forecasting model is expounded. For large groups of series, a forecasting method may be chosen in a fairly ad hoc way. With appropriate checks, it may be possible to base interval forecasts on the model for which the method is optimal. It is certainly unsound to use a model for which the method is not optimal, but, strangely, this is sometimes done. Some general comments are made as to why prediction intervals tend to be too narrow in practice to encompass the required proportion of future observations. An example demonstrates the overriding importance of careful model specification. In particular, when data are “nearly nonstationary,” the difference between fitting a stationary and a nonstationary model is critical.     

Tests for a difference in timing of physiological response between two brain regions measured by using functional magnetic resonance imaging

Summary.  Functional magnetic resonance imaging (FMRI) measures the physiological response of the human brain to experimentally controlled stimulation. In a periodically designed experiment it is of interest to test for a difference in the timing (phase shift) of the response between two anatomically distinct brain regions. We suggest two tests for an interregional difference in phase shift: one based on asymptotic theory and one based on bootstrapping. Whilst the two procedures differ in some of their assumptions, both tests rely on employing the large number of voxels (three-dimensional pixels) in non-activated brain regions to take account of spatial autocorrelation between voxelwise phase shift observations within the activated regions of interest. As an example we apply both tests, and their counterparts assuming spatial independence, to FMRI phase shift data that were acquired from a normal young woman during performance of a periodically designed covert verbal fluency task. We conclude that it is necessary to take account of spatial autocovariance between voxelwise FMRI time series parameter estimates such as the phase shift, and that the most promising way of achieving this is by modelling the spatial autocorrelation structure from a suitably defined base region of the image slice.     

Estimating Process Capability Index C PM Using a Bootstrap Sequential Sampling Procedure

Construction of a confidence interval for process capability index C _PM is often based on a normal approximation with fixed sample size. In this article, we describe a different approach in constructing a fixed-width confidence interval for process capability index C _PM with a preassigned accuracy by using a combination of bootstrap and sequential sampling schemes. The optimal sample size required to achieve a preassigned confidence level is obtained using both two-stage and modified two-stage sequential procedures. The procedure developed is also validated using an extensive simulation study.     

Determining Sample Size Using Fast and Slow Thinking

ABSTRACT

Using SPSS's bootstrapping procedures, this article demonstrates an approach to determining sample size that combines fast (heuristics or rules-of-thumb) and slow (statistical power analysis) thinking to balance statistical power, precision, and practicality. Sample size is determined for six commonly used statistical procedures: independent groups t-test, one-way ANOVA, one-way MANOVA, Pearson's r correlation, linear regression, and logistic regression. Overall, findings suggest that both approaches may under or over-estimate sample size. Both approaches yielded similar parameter and confidence interval estimates, but varied, sometimes by a factor of two, in their sample size requirements. It is hoped that this study's procedure and results will provide beginning reference points for sample size determination, and encourage researchers continue to search for resolutions for often difficult sample-size decisions.     

Testing the proportional odds model for interval-censored data

This paper discusses the goodness-of-fit test for the proportional odds model for K-sample interval-censored failure time data, which frequently occur in, for example, periodic follow-up survival studies. The proportional odds model has a feature that allows the ratio of two hazard functions to be monotonic and converge to one and provides an important tool for the modeling of survival data. To test the model, a procedure is proposed, which is a generalization of the method given in Dauxois and Kirmani [Dauxois JY, Kirmani SNUA (2003) Biometrika 90:913–922]. The asymptotic distribution of the procedure is established and its properties are evaluated by simulation studies     

A non-stationary integer-valued autoregressive model

It is frequent to encounter a time series of counts which are small in value and show a trend having relatively large fluctuation. To handle such a non-stationary integer-valued time series with a large dispersion, we introduce a new process called integer-valued autoregressive process of order p with signed binomial thinning (INARS(p)). This INARS(p) uniquely exists and is stationary under the same stationary condition as in the AR(p) process. We provide the properties of the INARS(p) as well as the asymptotic normality of the estimates of the model parameters. This new process includes previous integer-valued autoregressive processes as special cases. To preserve integer-valued nature of the INARS(p) and to avoid difficulty in deriving the distributional properties of the forecasts, we propose a bootstrap approach for deriving forecasts and confidence intervals. We apply the INARS(p) to the frequency of new patients diagnosed with acquired immunodeficiency syndrome (AIDS) in Baltimore, Maryland, U.S. during the period of 108 months from January 1993 to December 2001.     

Residual plots for linear regression models with censored outcome data: A refined method for visualizing residual uncertainty

Residual plots are a standard tool for assessing model fit. When some outcome data are censored, standard residual plots become less appropriate. Here, we develop a new procedure for producing residual plots for linear regression models where some or all of the outcome data are censored. We implement two approaches for incorporating parameter uncertainty. We illustrate our methodology by examining the model fit for an analysis of bacterial load data from a trial for chronic obstructive pulmonary disease. Simulated datasets show that the method can be used when the outcome data consist of a variety of types of censoring.     

Lower Confidence Limits on the Generalized Exponential Distribution Percentiles Under Progressive Type-I Interval Censoring

In industrial life test and survival analysis, the percentile estimation is always a practical issue with lower confidence bound required for maintenance purpose. Sampling distributions for the maximum likelihood estimators of percentiles are usually unknown. Bootstrap procedures are common ways to estimate the unknown sampling distributions. Five parametric bootstrap procedures are proposed to estimate the confidence lower bounds on maximum likelihood estimators for the generalized exponential (GE) distribution percentiles under progressive type-I interval censoring. An intensive simulation is conducted to evaluate the performances of proposed procedures. Finally, an example of 112 patients with plasma cell myeloma is given for illustration.     

Characterizations of asymptotic distributions of continuous-time Pólya processes

Abstract

We propose an elementary but effective approach to studying a general class of Poissonized tenable and balanced urns on two colors. We characterize the asymptotic behavior of the process via a partial differential equation that governs the process, coupled with the method of moments applied in a bootstrapped manner. We show that the limiting distribution of the process underlying the Bagchi-Pal urn is gamma. We also look into the tenable and balanced processes associated with randomized replacement matrix. Similar results carry over to the process, with minor modifications in the methods of proof, done mutatis mutandis.     

A New Method for Statistical Multidimensional Unfolding

ABSTRACT

Consider the problem of estimating the positions of a set of targets in a multidimensional Euclidean space from distances reported by a number of observers when the observers do not know their own positions in the space. Each observer reports the distance from the observer to each target plus a random error. This statistical problem is the basic model for the various forms of what is called multidimensional unfolding in the psychometric literature. Multidimensional unfolding methodology as developed in the field of cognitive psychology is basically a statistical estimation problem where the data structure is a set of measures that are monotonic functions of Euclidean distances between a number of observers and targets in a multidimensional space. The new method presented in this article deals with estimating the target locations and the observer positions when the observations are functions of the squared distances between observers and targets observed with an additive random error in a two-dimensional space. The method provides robust estimates of the target locations in a multidimensional space for the parametric structure of the data generating model presented in the article. The method also yields estimates of the orientation of the coordinate system and the mean and variances of the observer locations. The mean and the variances are not estimated by standard unfolding methods which yield targets maps that are invariant to a rotation of the coordinate system. The data is transformed so that the nonlinearity due to the squared observer locations is removed. The sampling properties of the estimates are derived from the asymptotic variances of the additive errors of a maximum likelihood factor analysis of the sample covariance matrix of the transformed data augmented with bootstrapping. The robustness of the new method is tested using artificial data. The method is applied to a 2001 survey data set from Turkey to provide a real data example.