Percentage points and power calculations for outlier tests in a regression situation

The usefulness of an extra sum of squares statistics Q_K for detecting K outliers has been discussed previously in the context of two-way tables. (See Gentleman and Wilk, 1975a, 1975b; John and Draper 1978; and Draper and John, 1980,) That work is extended here to straight line regression situations arising from, and motivated by, a specific set of research data. Percentage points for the appropriate test statistics are obtained by simulation, and approximations for these percentage points are suggested. Power calculations made for various designs and outlier situations are briefly summarized.     

Ecological Community Structure Analyses in the Formulation,Implementation, and Enforcement of Law and Policy

Most of what is currently known about ecosystems derives from two sources: comprehensive studies by classical ecologists and economically motivated studies by resource specialists. The former have traditionally concentrated on natural or long-undisturbed ecosystems, the latter on those few species in managed or polluted ecosystems that are harvested, hunted, or fished. A recent genre of legislation, drafted with insight provided by the scientific community, recognizes the dearth of ecological information on managed or polluted communities (which are becoming far more ubiquitous and important) and dramatically raises the requirements for comprehensive data and data analyses to guide managerial and pollution abatement policies. This paper describes several important current efforts to use ecological community structure analyses to implement and enforce such legislation and some effects of such analyses on environmental policy.     

Translation Formulae for Non-repeatable Events

China's one-per-hundred population survey, conducted in mid-1987, provides the first nation-level data with which to study recent fertility change in China. Using a recently developed extension of the ‘own-children’ method of fertility estimation, period parity progression ratios are computed from the survey data. Comparison with similar statistics computed from the 1982 one-per-thousand fertility survey provides a rigorous check on the quality of the results. The level of fertility so measured rose by 13 per cent between 1985 and 1987, compared with an increase of eight per cent in conventional total fertility ratios. Nearly 90 per cent of the increase was due to rising levels of progression from first to second birth. There can be little doubt that this, in turn, was due to a relaxation in the one-child family policy. Overall levels of progression to births of higher orders have been declining since 1982, but the evidence suggests that this is so only because of stringent government efforts to control births of third and higher orders.     

Letter to the editor the asymptotic distribution of the process capability index cpmk

In theorem 3.1 of Chen and Hsu (1995) an expressin for an upper confidence bound for C_pmk in the case when µ = M and so C_pmk = C_pm, is given as a multiple of C_pm. However, if the value µ = M known then it would be more reasonable to use the estimator obtained by replacing X_n by M.     

Actual rejection levels in a certain stepwise test

Continuing me work of Draper, Guttman and Kanemasu (1971) we obtain the appropriate distribution and evaluate some actual probability levels tor entry of variables in stepwise regression when the denominator of the F-statistic is a biased estimate of the residual variance and there are two possible entry candidates.     

The proportional hazards model: applications in epidemiology

Cox's (1972) Proportioal hazards failure time model, already widely used in the analysis of clinical trials, also provides an elegant formalization of the epidemiologic concept of relative risk. When used to compare the disease experience of a study cohort with that of an external control population, it generalizes the notions of the standardized morbidity ratio (SMR) and the proportional morbidity ratio (PMR). For studies in which matched sets of cases and controls are sampled retrospectively from the population at risk, the model provides a flexible tool for the regression analysis of multiple risk factors.