Subset selection of the best treatment via many-one tests

The method of Gupta (1956, 1965) was developed to select a subset from k normal populations that contains the best populations with given probability. This paper shows a duality between the general goal of selecting a subset for the best population and many-one tests. A population should be regarded as ‘candidate’ for the best population and thus retained in the subset if the samples from the other populations are not significantly better. Based on this ‘idea’ a general selection procedure is proposed using many-one tests for the comparison of each population against the remaining ones.     

Constrained Monotonicity and the Measurement of Power

In this paper we will discuss constraints on the number of (non-dummy) players and on the distribution of votes such that local monotonicity is satisfied for the Public Good Index. These results are compared to properties which are related to constraints on the redistribution of votes (such as implied by global monotonicity). The discussion shows that monotonicity is not a straightforward criterion of classification for power measures. This revised version was published online in June 2006 with corrections to the Cover Date.     

Sample size determination of steel's nonparametric many-one test

We derive sample size formulas for the many-one test of Steel (1959) when the all-pairs power is preassigned. In this large sample approach we replace, similar to Noether (1987), the unknown variances and also the unknown correlation coefficients in the power expressions by their known values under the null hypotheses. We then obtain least favorable configurations for one-and two-sided comparisons. The reliability of our formulas is examined in computer simulations for different alternatives with various distributions.     

Subjective and Objective Parameters Determining 'Salience' in Long-term Dialect Accommodation

In this article, we will present empirical results of a longitudinal study on long-term dialect accommodation in a German dialect setting. An important model of explaining which linguistic structures undergo such convergence and which do not makes use of the notion of 'salience'. Dialect features which are perceived by the speakers as 'salient' are taken up and given up more easily and faster than those which are perceived as 'less salient'. The notion of salience has a tradition which goes back to the 1920s. We will discuss this research tradition, apply the criteria for salience that played a role in it to our results, and discuss the question of whether perceived (subjective) salience can be explained in objective (structural-phono logical or phonetic) terms.     

Can oil-led growth and structural change go hand in hand in Ghana?

Unlike in Asia, the manufacturing sector has not (yet) become a driver of structural change in Africa. One common explanation is that the natural resource-focus of many African economies leads to Dutch disease effects. To test this argument for the case of newly found oil in Ghana we develop a multi-sector intertemporal general equilibrium model with endogenous savings and investment behavior. Results show that in addition to the well-known short-term Dutch disease effects, long-term structural effects can indeed impede Asian-style economic transformation in Ghana (and other resource rich countries). We also demonstrate how oil wealth may go hand in hand with structural change in the future.     

Generalized linear models and transformations for unreplicated factorial experiments in the presence of dispersion effects

For the analysis of replicated designs, many different methods have been suggested. These allow for the estimation of functional dependencies between mean and variance as well as possible dispersion effects within the same model framework. However, in the situation of unreplicated designs, most methods known so far rely on the assumption of constant variances, or a functional relationship between mean and variance as the only source of heteroscedasticity. In this paper, we propose two methods for dealing with unreplicated data, when dispersion effects might also be of importance. One of these is an extension of the Box–Cox-method [Box, G.E.P., Cox, D.R., 1964. An analysis of transformations. Journal of the Royal Statistical Society B 26, 211–252], the other is based on double generalized linear models. Both these methods turn out to yield approximately equivalent results in the case of comparable assumptions, whereas the double generalized linear model is the more general one and allows further extensions. If this class of models is assumed, consistency, asymptotic efficiency and normality of the resulting estimates are shown.     

Mental Models in Risk Assessment: Informing People About Drugs

One way to communicate about the risks of drugs is through the use of package inserts. The problems associated with this medium of informing patients have been investigated by several researchers who found that people require information about drugs they are using, including extensive risk information, and that they are willing to take this information into account in their usage of drugs. But empirical results also show that people easily misinterpret the information given. A conceptual framework is proposed that might be used for better understanding the cognitive processes involved in such a type of risk assessment and communication. It is based on the idea that people develop, through experience, a mental model of how a drug works, which effects it might produce, that contraindications have to be considered, etc. This mental model is "run" when a specific package insert has been read and a specific question arises such as, for example, whether certain symptoms can be explained as normal or whether they require special attention and action. We argue that the mental model approach offers a useful perspective for examining how people understand package inserts, and consequently for improving their content and design. The approach promises to be equally useful for other aspects of risk analysis that are dependent upon human judgment and decision making, e.g., threat diagnosis and human reliability analysis.     

Assessing children's relationships with pet dogs: A multi-method approach

Relationships with pet dogs are thought to provide substantial benefits for children, but the study of these relationships has been hindered by a lack of validated measures. Approaches to assessing the quality of children's pet dog relationships have tended to focus on positive relationship qualities and to rely on self-report questionnaires. The aim of this study was to develop and test multiple measures that could be used to assess both positive and negative features of children's relationships with pet dogs. In a sample of 115 children ages 9–14 years who were pet dog owners, we assessed six qualities of pet dog relationships: Affection, Nurturance of Pet, Emotional Support from Pet, Companionship, Friction with Pet, and Pets as Substitutes for People. All qualities were assessed with child questionnaires, parent questionnaires, and child daily reports of interactions with pets. We found substantial convergence in reports from different observers and across different measurement approaches. Principal components analyses and correlations suggested overlap for many of the positive qualities, which tended to be distinct from negative relationship qualities. The study provides new tools which could be used to test further how relationships with pets contribute to children's development.