Carol A. Markowski

Conditions for the Effectiveness of a Preliminary Test of Variance

Abstract We evaluate the performance of the F test for the equality of two variances as a preliminary test to determine the appropriateness of the two-sample t test. When sampling is from a normal distribution, our results indicate that, for equal sample sizes, the t test is insensitive to variance heterogeneity and hence no preliminary test is necessary. For unequal sample sizes, the F test has small probability of detecting many alternatives for which the t test performs poorly, and so the F test is not an effective preliminary test. When sampling is from other distributions, our results confirm the sensitivity of the F test to the normality assumption.

Assessing the effects of manufacturing infrastructure preparation prior to enterprise information-systems implementation

Reports on the success or failure of enterprise information system (EIS) implementation have been decidedly mixed. In this study, we focus on manufacturing infrastructure preparation prior to EIS implementation and report the results of a survey of 152 US manufacturing companies that have implemented EIS. We have provided four major findings in this study: (1) the requirements from customers and trading partners are more powerful drivers motivating US manufacturing firms to implementing enterprise systems than internal business planning needs; (2) one manufacturing infrastructural issue often has implications for other infrastructural items in implementing technology, so various manufacturing infrastructural issues should be prepared simultaneously; (3) manufacturing infrastructure preparation prior to EIS implementation has significant positive effects on customer-focused performance, production/operations performance, and financial performance; and (4) better customer-focused performance contributes to better financial performance.

An experimental comparison of several approaches to the discriminant problem with both qualitative and quantitative variables

Abstract This paper reports the results of an experimental comparison between a linear programming approach and the well known statistical procedure by Fisher for solving discriminant analysis problems. This work represents an extension of the work of Bajgier and Hill as we incorporate qualitative variables into our design as well as enlarge the estimation sample from 30 to 90. Our results indicate that both methods are enhanced by the inclusion of the qualitative variables, but that the Fisher approach seems preferable. Lastly, we discuss the effect of the different experimental factors on the relative performance of the two methods.

Improved attribute acceptance sampling plans in the presence of misclassification error

Abstract This paper considers an attribute acceptance sampling problem in which inspection errors can occur. Unlike many common situations, the source of the inspection errors is the uncertainty associated with statistical sampling. Consider a lot that consists of N containers, with each container including a large number of units. It is desired to sample some of the containers and inspect a sample of units from these selected containers to determine proper disposition of the entire lot. Results presented in the paper demonstrate significant shortcomings in traditional sampling plans when applied in this context. Alternative sampling plans designed to address the risk of statistical classification error are presented. These plans estimate the rate of classification errors and set plan parameters to reduce the potential impact of such errors. Results are provided comparing traditional plans with the proposed alternatives. Limitations of the new plans are also discussed.

Duality and transformations in multiphase and sequential linear goal programming

Abstract We consider herein a specific type of goal programming model, namely the lexicogrpahic linear goal programming model. Although the two most common methods of solution, the sequential process and the multiphase process, produce the same solutions, the interiors of the final tableaus will differ. We present algorithms which allow one to transform the sequential tableau into the multiphase tableau and vice versa and, in doing so, demonstrate the respective mathematical duals and their relationships. These results are particularly important when performing sensitivity analysis.

An adaptive statistical method for the discriminant problem

Abstract This paper proposes an adaptive statistical method for the discriminant problem. The method selects Fishers linear or Smiths quadratic discriminant function or the nearest neighbor method for use on the holdout sample, depending upon which method minimizes the sum of overall accuracy and balance on the estimation sample. A simulation study examines the two group discriminant problem with variables generated from both bivariate normal and nonnormal distributions. The resulting misclassification rates indicate that the adaptive method is an effective alternative to existing statistical and linear programming methods.

Evaluation of an adaptive discriminant procedure

This paper reports the results of a simulation study comparing Fishers Linear Discriminant Function, Smiths Quadratic Discriminant Function, a nonparametric Nearest Neighbor approach, a linear programming approach, and a new adaptive statistical method for solving the discriminant problem. The study examines the two-group discriminant problem with four variables, of which two are continuous and two are discrete. The analysis is based on the rate of misclassification using each method. The results indicate that the adaptive method is an effective alternative to existing methods and that the adaptive philosophy of using the training sample to identify which of several methods should be applied to the validation sample merits further study.

Viewpoint: Practical Uses of Statistical Power in Business Research Studies

Abstract The concept of power in statistical hypothesis testing is typically introduced by emphasizing (a) the relationship with Type I and Type II errors and (b) the use of power for planning sample sizes. In this article, we propose use of power information subsequent to results of testing hypotheses through use of statistical tests in business research studies to assist in interpretation of existing data and for guidance in determining whether additional data might be necessary. We provide several examples and recommendations on how such posttest use of power might be integrated within business statistics courses.