Gary L. Tietjen

Some Grubbs-Type Statistics for the Detection of Several Outliers

Several widely used tests for outlying observations are reviewed. Problems of repeated application and “masking” are described. Suggested as appropriate to over-come these problems are two new statistics: Lk which is based on the k largest (observed) values and Ek which is based on the k largest (in absolute value) residuals. Tables of approximate critical values for these statistics are given for 0.01, .025,0.05, and 0.10 levels of significance and for sample size n = 3 (1) 20 (5) 50.

A New Family of Probability Distributions with Applications to Monte Carlo Studies

Abstract A new probability distribution is presented that offers considerable potential for providing stochastic inputs to Monte Carlo simulation studies. The distribution includes the exponential power family as a special case. An efficient computational strategy is proposed for random variate generation. An example for testing the hypothesis of unit variance illustrates the advantages of the proposed distribution.

A melanoma case-control study at the Los Alamos National Laboratory.

We conducted a melanoma case-control study at the Los Alamos National Laboratory to investigate whether related occupational exposures or personal characteristics of employees could be identified. This study was prompted by a recent report from the Lawrence Livermore National Laboratory that melanoma was much more frequent than expected among employees and that persons suffering from melanoma more often worked as chemists. Our investigation did not uncover an association with plutonium body burden, cumulative external radiation exposure, or employment as a chemist or a physicist. The major finding was that cases were more educated than controls. Melanoma risk was 2.11 among college-educated employees and increased to 3.17 among those with graduate degrees (Mantel-extension linear trend probability = 0.038). This finding is consistent with the often reported increased melanoma incidence among persons of higher social class. It points to personal characteristics, particular to persons of higher educational attainment, as risk factors for melanoma at the Los Alamos National Laboratory.

EXACT AND APPROXIMATE TESTS FOR UNBALANCED RANDOM EFFECTS DESIGNS

SUMMARY In the three stage nested design with random effects, let ?A2 be the variance of factor A (the top stage). It is often desirable to test the hypothesis HA: GA2 = 0. In the unbalanced case, the conventional F-test for HA does not in general have the expected null distribution, and the expected mean squares are not in general equal under HA . Tietjen and Moore [1968] proposed that a Satterthwaite-like procedur-e be used to construct a denominator (and D.F.) which would have, under HA , the same expected mean square as the numerator. It was pointed out by W. H. Kruskal, however, that such a procedure had no justification since the mean squares were not independent and did not have chi-square distributions. This paper is an attempt to revisit the question by investigatiing the properties of this approximation and those of the conventional F-test. It is shown that the conventional F-test does considerably better as an approximation under imbalance than does the Satterthwaite test.

Two-sided tolerance limits for balanced random-effects ANOVA models

In this article, we derive two-sided approximate β-content tolerance limits for multiway balanced random-effects models. We provide factors, obtained from numerical integration, that can be used to obtain β-content tolerance intervals. We describe methods for extending the results to nested models and discuss the use of the tabled tolerance factors for exact intervals for simple random samples when we have an independent estimate of the variance. We demonstrate the procedure with an experimental design used to evaluate the firing time precision in a high-explosives system.

Mortality among workers at the Pantex weapons facility

We compared total and cause-specific mortality for workers at the Pantex nuclear weapons assembly facility employed between 1951 and 31 December 1978 with expected mortality based on U.S. death rates. We observed significantly fewer deaths than expected from all causes of death, all cancers, digestive cancers, lung cancer, arteriosclerotic heart disease, and digestive diseases. There were no causes of death which occurred significantly more frequently than expected. Analyses of worker mortality by duration of employment, time since first employment, and radiation exposure greater than 1.00 rem produced similar results. We found no evidence that mortality from any cause of death was increased as a result of employment at Pantex.

Double Sampling Plans where the Acceptance Criterion is the Variance

In certain instances a lot of items is judged to be acceptable if the variance of some measurable characteristic of the lot items is less than or equal to & Operating characteristics of single size sampling plans are given in [l] and [2]. This paper gives a set of double-sampling plans whose operating characteristics match those of a set of single sample plans.

Migratory Behaviors of Alveolar Macrophages During the Alveolar Clearance of Light to Heavy Burdens of Particles

We investigated the unstimulated and stimulated migratory activities of lavaged alveolar macrophages (AMs) in vitro over the course of alveolar clearance of three different mass lung burdens of microspheres. Our intent was to uncover potentially important relationships between the migratory behaviors of the AM and the retention kinetics of particles. Groups of adult, male Fischer-344 rats were intratracheally instilled with approximately 86 micrograms (low burden, LB), approximately 1 mg (medium burden, MB), or approximately 3.7 mg (high burden, HB) of polystyrene microspheres (2.13 microns diameter), or with carrier vehicle (phosphate buffered saline, PBS) alone. The lung retention kinetics of the particles were determined over an approximately 170 day period. On days 14, approximately 57, and approximately 85, lavaged AMs were assessed for their abilities to migrate through 5-microns pore membranes in response to inactivated rat serum (unstimulated condition) and yeast-activated rat serum (stimulated condition). The retention characteristics of the three burdens could be satisfactorily described by two-component, negative exponential equations. The kinetics of retention of the LB and MB were similar, although some evidence indicated the MB slightly retarded the lung clearance process. Deposition of the HB resulted in more marked prolongations of both the early, more rapid, and the slower, longer term components of alveolar clearance. The unstimulated migration indices of AMs from the particle-instilled lungs were generally not significantly different from those of AMs from PBS-instilled lungs except for a significant increase in the migration indices of LB AMs at the last assay time. The stimulated migration indices of MB and HB AMs were significantly decreased on assay days 14 and approximately 57. On day approximately 85, however, the migration indices of LB, MB, and HB AMs were all increased above the PBS AMs. Comparisons of the frequency distributions of particles in the unstimulated and stimulated AM that migrated to those in corresponding parent AM populations consistently indicated a preferential migration of particle-free AMs and of AMs with lesser loads of microspheres. The overall results of this study suggest that the unstimulated and stimulated migratory activities of particle-laden AMs are depressed in vitro. Our results also suggest that the migratory activities of generally particle-free AMs may be enhanced over a prolonged period of time following the deposition of particles in the lung.

Tables for Use of the Maximum F-Ratio in Multiple Comparison Procedures

Abstract In 1956 David devised a procedure for making multiple comparisons among a set of variances. The procedure, which was analogous to Duncans Multiple Range Test, was based on Hartleys F-max test. David gave tables for α = .01 for testing k variances where k = 2(1) 12. The present article extends the tables to α = .05 and α = .10 for k=1(1)15, 20, 30, 40(20)100.

Brain tumors at a nuclear facility.

In response to an observed excess risk of brain tumor deaths among workers at the Rocky Flats Nuclear Facility (Colorado), a case-control study of all (n = 16) primary brain tumor deaths occurring among white males employed during 1952 through 1977 was conducted to investigate their relationship with occupational radiation/nonradiation exposures. For each case, four controls were individually matched on year of birth and period of employment. Although limited by a small number of cases, our study showed no statistically significant association between brain tumor death and exposure to internally deposited plutonium, external radiation, or other occupational risk factors.