Failure Modes and Effects Analysis: An Evaluation of Group versus Individual Performance

Failure modes and effects analysis (FMEA) is one of the most frequently used tools in process and product design: it is used in quality and reliability planning, and event and failure mode analysis. It has a long history of use and is a formally prescribed procedure by a number of prominent standards organizations. In addition, it's popular use has evolved as a less formal and widely interpreted tool in the area of Lean/Six Sigma (LSS) process improvement. This paper investigates one of the most important issues related to FMEA practice—the quality of individual vs. group performance in ranking failure modes. In particular, we compare FMEA rankings generated by: (i) individuals, (ii) group consensus, and (iii) non‐collaborative aggregation of group input (a synthesized group ranking). We find that groups outperform individuals and that synthetic groups perform as well as group consensus. We explain the implications of this result on the coordination of the design of products and processes amongst distributed organizations. The increasing distribution of product design efforts, both in terms of geography and different organizations, presents an opportunity to improve coordination using distributed synthetic group‐based FMEA.     

A Common Situation Conducive to Bizarre Distribution Shapes

Conditions conducive to bizarre-shaped distributions are fairly common in certain areas of research where, for perfectly valid reasons, an important causal variable is left uncontrolled. Theoretical rationale and actual examples are given to show how such distributions are generated, to exemplify their shapes, and to indicate their prevalence in practice. An L-shaped type tends to occur when time scores are recorded for a task subject to infrequent but time-consuming errors; other types occur when measuring behavior influenced by social conformity, or under other circumstances. The L shape appears to be far more conducive to nonrobustness than are previously investigated shapes.     

Computer Algebra Systems: DERIVE®

Seeing Through Statistics.; Jessica M. Utts. Belmont, CA: Duxbury Press, 1996. xvi + 464 pp. $40.95. Reviewed by Sheila O'Leary Weaver

Statistics and Data Analysis, 2nd ed.; Andrew F. Siegel and Charles J. Morgan. New York: John Wiley, 1996. xiv + 635 pp. $68.95. Reviewed by Jennifer Hoeting

Statistics Principles and Methods (3rd ed.).; Richard A. Johnson and Gouri K. Bhattacharyya. New York: John Wiley, 1996. xv + 720 pp. $70.95. Reviewed by Linda C. Malone

Engineering Statistics: The Industrial Experience.; Bernard Ostle, Kenneth V. Turner, Jr., Charles R. Hicks, and Gayle W. Mcelrath. Belmont, CA: Duxbury Press, 1996. xiv + 568 pp. $62.95. Reviewed by Anne M. Dougherty

Introductory Statistics.; Sheldon M. ROSS. New York: McGraw-Hill, 1996. 807 pp. $63.59. Reviewed by Youn-Min Chou

Workshop Statistics.; Allan J. ROSSMAN. New York: Springer-Verlag, 1996. xx + 452 pp. $26.95. Reviewed by Jonathan D. Cryer

Exploring Probability and Statistics with Spreadsheets.; J. T. Callender and R. Jackson. Englewood Cliffs, NJ: Prentice-Hall, 19xx. viii + 236 pp. $36.00. Reviewed by K. L. D. Gunawardena

NCSS (6.0.21 for Windows); Available from NCSS Statistical Software, 329 North 1000 East, Kaysville, UT 84037. For five or more student copies the cost is $25 for each set of software disks and $30 for each set of manuals. Reviewed by Michael Nabel

DASL—The Data and Story Library; This collection of datasets and stories designed for use in statistics courses is available on the Internet at http://www.stat.cmu.edu/DASL/. Reviewed by Jeffrey A. Witmer

StatLib: An Archive for Statistical Software, Datasets, and Information: StatLib is a system for distributing statistical software, datasets, and information by electronic mail, FTP, gopher, and WWW Although StatLib is not primarily intended as a resource for teachers, much of its contents are of interest when teaching statistics. StatLib is free and available via the Internet. Reviewed by Charles Kooperberg     

IT Governance Characteristics,Electronic Medical Records Sophistication,and Financial Performance in U.S. Hospitals: An Empirical Investigation

As a result of a recent federal government mandate, an increasing number of hospitals have decided to adopt electronic medical record (EMR) systems. This initiative is expected to lead toward more efficient and higher quality health care; however, little is known about governance characteristics and organizational performance for EMR adopters. Our goal is to inform theory and practice by examining hospitals with a sophisticated EMR and comparing those hospitals to similar hospitals (with a less sophisticated EMR) to understand the association between information technology (IT) governance characteristics and the implications on financial performance. Leveraging elements of the upper echelon theory, we posit that hospitals in which the chief information officer (CIO) reports to the chief executive officer, CIO turnover is low, and an IT steering committee is present are more likely to have a sophisticated EMR. We argue that EMR sophistication leads to improved financial performance. Our results underscore the importance of continuity in the CIO position on successful EMR implementations. Results also show that hospital size and financial performance are strongly associated with EMR sophistication. In addition, we find that a sophisticated EMR appears to be a fundamental element in improving hospitals’ revenue cycle management. Moreover, we find that hospitals with a sophisticated EMR appear to be more profitable. Finally, we observe that total payroll expense adjusted by total discharges drops among the sophisticated hospitals, potentially due to an increase in employee productivity. These insights can serve as a basis for tempering expectations relative to the financial impact of EMR adoption.     

Unpacking Team Familiarity: The Effects of Geographic Location and Hierarchical Role

Examination of team productivity finds that team familiarity, i.e., individuals' prior shared work experience, can positively impact the efficiency and quality of team output. Despite the attention given to team familiarity and its contingencies, prior work has focused on whether team members have worked together, not on which team members have worked together, and under what conditions. In this paper, I parse overall team familiarity to consider effects of geographic location and the hierarchical roles of team members. Using data on all software‐development projects completed over 3 years at a large Indian firm in the global outsourced software services industry, I find that team familiarity gained when team members work together in the same location has a significantly more positive effect on team performance compared with team familiarity gained while members were collaborating in different locations. Additionally, I find that hierarchical team familiarity (a manager's experience with front‐line team members) and horizontal team familiarity (front‐line team members' experience gained with one another) have differential effects on project team performance. These findings provide insight into the relationship between team experience and team performance.     

制备L-阿拉伯糖的多元集成反应分离装备开发

针对用木糖母液和玉米皮水解液通过色谱分离来制取L 阿拉伯糖的传统工艺,提出了一种将其水解反应和色谱连续分离进行过程耦合调控的新工艺,并应用最新机电一体化控制技术,成功开发出制备L 阿拉伯糖的多元分流式反应分离集成装备。该类新型装备流程简洁、结构紧凑、自动化程度高、性能稳定可靠,具有较显著的节能减排效果,符合现代轻工机械最新发展方向,值得应用与推广。     

A developmental, biopsychosocial model for the treatment of children with gender identity disorder

This article provides a summary of the therapeutic model and approach used in the Gender Identity Service at the Centre for Addiction and Mental Health in Toronto. The authors describe their assessment protocol, describe their current multifactorial case formulation model, including a strong emphasis on developmental factors, and provide clinical examples of how the model is used in the treatment.     

Generalized gamma parameter estimation and moment evaluation

The generalized gamma distribution includes the exponential distribution, the gamma distribution, and the Weibull distribution as special cases. It also includes the log-normal distribution in the limit as one of its parameters goes to infinity. Prentice (1974) developed an estimation method that is effective even when the underlying distribution is nearly log-normal. He reparameterized the density function so that it achieved the limiting case in a smooth fashion relative to the new parameters. He also gave formulas for the second partial derivatives of the log-density function to be used in the nearly log-normal case. His formulas included infinite summations, and he did not estimate the error in approximating these summations.

We derive approximations for the log-density function and moments of the generalized gamma distribution that are smooth in the nearly log-normal case and involve only finite summations. Absolute error bounds for these approximations are included. The approximation for the first moment is applied to the problem of estimating the parameters of a generalized gamma distribution under the constraint that the distribution have mean one. This enables the development of a correspondence between the parameters in a mean one generalized gamma distribution and certain parameters in acoustic scattering theory.     

Testing for empty sets when misclassifications may be present

A test for empty sets is described. This new test is suitable for testing for mutual exclusivity, sets being nested, and many other configurations. It might be hypothesized that two medical conditions do not both occur in individuals. This can be expressed as an empty intersection of the two sets of people with each condition. An important feature of the test is the incorporation of misclassification rates into the analysis. The test utilizes the potential misclassifications for false designations in the sets that are hypothesized to be empty. The test is quite powerful. When the null hypothesis is rejected, follow-up tests on one or more sets that are included in the null hypothesis can be performed using the same new test. MatLab code is supplied.