Douglas A. Zahn

Modifications of and Revised Critical Values for the Half-Normal Plot

The original version of the half-normal plot [C. Daniel, Technometrics, t, (1959), 311–34] is reviewed and two modified versions are introduced. The modified versions incorporate corrections of a major flaw in the original version relating to critical values and a minor flaw relating to plotting positions. The critical values given in this article control the probability error rate, the probability of at least one false positive in the analysis of an experiment, and are considerably different from those given by Daniel. One of the modified versions is more powerful than Daniels original procedure. Anothe improvement made in the modified versions is that only the smallest 70%, of the contrasts not declared significant, are used to constuct the final estimate of the error variance. This results in an appreciable redaction in the mean square error of this estimate. Examples are plesented to illustrate the use of the various half-normal plot versions.

An Empirical Study of the Half-Normal Plot

Monte Carlo studies of a modification of the original version of the half-normal plot (Daniel, Technometrics, 1 (1959), 311–341) and four new versions are reported. Data representative of the 15 contrasts from a 2p–q , p – q = 4, factorial experirnenl. are generated. Design parameters in the Main Simiulation Study are the probability error rate, the number of real contrasts, and the size of the real contrasts. The critical values used by the various versions control the probability error rate. These critical values are considerably different, than those given by Daniel. The Monte Carlo strtdies indicate that the detection rate, i.e., the proportion of real contrasts declared significant, is larger for one of the new versions than for the original version. The detection rate of all versions decreases drastically when the number of real contrasts present increases from one to two to four. Nomination procedures for analyzing single replication 24 factorial experiments have a smaller detection rate than the h...

An Expanded Approach to Educating Statistical Consultants

Abstract This article describes an expanded approach to the education of statistical consultants that is being implemented at Florida State University. Our program begins with a preconsulting course in which students first study general problem-solving techniques. Next these techniques are used in solving applied statistics “textbook” problems. Then problem formulation and research methods are studied. Finally the structure of consulting sessions is considered, along with interpersonal issues that arise in consulting. In the supervised consulting course, students work with actual clients in videotaped sessions, attend supervision sessions in which the tapes are reviewed, present case conferences, attend new material sessions, and complete a project. We conclude this article with preliminary evaluation data on the program and observations regarding what we have learned.

Nonstatistical Aspects of Statistical Consulting

Abstract Statistical consulting is a complex activity that requires statistical and nonstatistical skills. Together, these skills determine the ultimate success of the consultation. This article focuses on the nonstatistical aspects of the initial meeting between statistical consultant and client and presents a model that views consulting sessions as consisting of four parts: (a) identification of relevant aspects of the problem situation, (b) definition of the clients goals, (c) determination of the actions to be taken, (d) discussion of various aspects of the consulting relationship and who will do what when. Each part gives rise to specific issues and requires that the consultant have various nonstatistical skills to deal with these issues. The effectiveness of the consultation can be enhanced by learning and practicing this model and the constituent skills.

Promoting Statistics: On Becoming Valued and Utilized

A sense of dissatisfaction exists in the statistics profession stemming from a consensus that statisticians in all working environments are undervalued and undemtilized. This dissatisfaction is reflected in the last 10 ASA presidential addresses and in numerous conversations we have had with statisticians at professional meetings and workshops. Statisticians have reacted to their dissatisfaction with several proposals: develop a stronger professional identity by establishing professional certification, stamp out the misuse of statistics in widely available PC software packages, and mount marketing campaigns to remake the image of the statistician from number cmncher to mainstream societal decision maker. All of these are misguided. Looking outside the profession for solutions to problems in the areas of value and utilization will not be productive. In the words of Walt Kelly’s Pogo, “We have met the enemy and he is us.” The sources of these problems lie within the

Exact χ2 Criterion Tables with Cell Expectations One: An Application to Coleman's Measure of Consensus

Abstract The usual χ2 goodness-of-fit criterion is considered when all cells have expectations of one and the cell boundaries are not estimated from the data. This situation arises in the use of a measure of consensus proposed by Coleman [1]. Under these circumstances the criterion reduces to X2 = Σt=1 N n i 2−N, where n i is the number of observations in cell i and N is the total number of observations in the sample. The exact distribution of X2 is given for N = 1(1)25(5)40, 50. For N≥25, the χ2 approximation to the exact distribution of X2 errs by less than .007 or 20.3 percent for right-tail areas. For left-tail areas the errors are less than .0165 or 89.9 percent when N≥25.

Current challenges in statistics: large-lecture courses

Each semester approximately 80% of the students taught by the Florida State University Department of Statistics are enrolled in STA 3014: Fundamental Business Statistics. During the academic year this course is taught in large lecture sections of 250 students each. It is either the only statistics course or one of two statistics courses taken in their undergraduate career for probably 90% of these students. A similar situation exists in many statistics departments around the nation. These large introductory courses offer us the opportunity to introduce the power of statistics to a large fraction of our future business leaders. In the past it appears that this opportunity has often been missed. In fact, some suggest that these courses help contribute to the general publics negative attitude toward statistics courses, the discipline of statistics, and statisticians. Hence, I propose that one of the current challenges in statistics is the challenge of improving the quality of these courses so that statistic...

TEACHING STATISTICAL CONSULTING: STATISTICAL AND NON-STATISTICAL ASPECTS

Publisher Summary This chapter presents an interdisciplinary approach to teaching statistical consulting. The program described in this chapter presents an alternative way to achieve many of the goals of the internship program. The program embodies many features of an internship program. In addition, the program offers the option to cover some additional ground not usually considered in consulting internships. More students can be taught using the course program than using an internship program. The basic nonstatistical and statistical concepts are covered as a group in an orderly fashion rather than dealing only with the subset that would occur in selected consulting sessions. In this way, the group can learn from the literature and from each other. Students may see more consulting styles in the course than in internships. The supervision procedures described in this chapter could profitably be added to an internship program. It is a good idea for instructors to increase the number of hours of consultation that students do in the presence of a faculty member before seeing their own clients, though this form of supervision can get expensive, time-wise.

THE EVOLUTION OF SUPERVISED STATISTICAL CONSULTING AT FLORIDA STATE UNIVERSITY: A RESPONSE TO THE REVIEW OF JAMES R. BOEN

Publisher Summary This chapter discusses the evolution of supervised statistical consulting at Florida State University. The Statistical Consulting Center was established in the Department of Statistics in 1969, though informal consultation had been provided since the early 1960s. The Center was charged with providing free statistical consultation to faculty and graduate students at Florida State. In 1970, to train the graduate students in the techniques of statistical consultation, the supervised consulting course was started, initially focusing almost entirely on statistical aspects of consulting. In 1980, a common goal of training more effective statistical consultants started, relating to publications and research in psychology and statistics, which could be achieved while working toward the common goal. Videotaping was started in actual consulting sessions in April 1980 in the departmental conference room using equipment borrowed from the FSU Media Center. Students in the consulting class in April 1980 were given the assignment to videotape an initial consulting session with a new client.

Empirical analysis of dramatic structure

This study investigated perceptions of dramatic structure. Subjects were randomly assigned to groups in which they only read, only saw, or read and then saw Edward Albees The Zoo Story. The script was divided into motivational units and the subjects in each group completed a questionnaire ranking and rating the units with respect to dramatic tension. With no a priori definitions of dramatic tension or structure, the subjects responded very similarly when asked to express their analyses of the play they saw and/or read. The audience judgments of The Zoo Story structure as a whole very closely approximated the Matthews “well‐made modern play” model.