MEASURING LEVELS OF PARTY IDENTIFICATION DOES QUESTION ORDER MATTER?

This article examines whether asking the vote question beforeparty identification alters the strength of partisanship andits relationship to vote choice. It employs the 1992 BritishElection Survey, which included a random split half-sample experiment,and the 1992–93 American Election Study Panel, where thequestion order for party identification and the vote were changed.The results show that altering the question ordering had verylittle effect in Britain and no significant effect in the UnitedStates. These results are consistent with the notion that partyidentification is one of the more enduring and stable componentsof mass political behavior in both presidential and parliamentarysystems.     

The effects of learning and forgetting on the optimal lot size quantity of intermittent production runs

This paper studies the effects of learning and forgetting on the production lot size problem with infinite and finite planning horizons. It is assumed that the determination of the economic manufactured quantity (EMQ) in the succeeding production run is dependent on: (1) the maximum inventory accumulated prior to interruption; (2) the length of the interruption period which incurs total forgetting; and (3) the level of experience in equivalent units remembered at the start-up of the next production run. The optimum operating inventory doctrines is obtained by trading off procurement cost per unit time and the inventory carrying cost per unit time, so that their sum will be a minimum. A numerical example is presented to demonstrate the application of learning and forgetting to the determination of the EMQ.     

Gender and migration: geographical mobility and the wife's sacrifice

Although it has become almost axiomatic in much contemporary sociology that geographical mobility for husband's career reasons is disruptive of wives' labour market careers, there have been few systematical empirical inquiries into the issue. This study reports findings from a survey of in-migrants in Aberdeen in 1986. Considerations relating to male partner's occupational career were the main factor accounting for the in-migration of married women, but single women were more likely to come for educational or employment reasons. Levels of grievance among in-migrant partnered women about the consequences of this move for their employment were low but there was some evidence that women in-migrant for such reasons saw the move as less helpful careerwise than other in-migrant women. More, however, saw the move as helpful or making no difference, rather than being a hindrance. Many respondents were able to find what they regarded as suitable new employment and others saw themselves at the time of the move primarily as mothers or housewives. The research findings lend support for the view that migrant wives' roles are orientated primarily around their husband's occupational concerns but this need not be detrimental to their own labour market activity since their skills, aptitudes and resources facilitate their adaptation to the new situation.     

Present positions and future prospects in management science

This paper offers some thoughts on the present state of management science as a discipline and a profession, and considers possible developmental strategies. The growth of management science, and the optimism this engendered, was largely premised on the successful use of scientific methodology and quantitative techniques to solve a relatively narrow range of management problems. In the 1970s, however, this positivist/quantitative ‘traditional management science’ became subject to increasing critical assault from those who wanted to broaden the impact of the discipline. Alternative management science approaches were born—soft systems thinking, organisational cybernetics, critical management science—and succeeded in establishing themselves. The existence of these alternative strands of work, alongside traditional management science, raises important questions about the future development of the discipline and profession. In this paper four developmental possibilities are set down and examined to see what future prospects they hold out for management science. It is argued that three of these—the isolationist, imperialist and pragmatist strategies—would lead the discipline into a dead end. A pluralist option, however, offers excellent opportunities for successful, future development.     

The simulation of industrial robot systems

This paper describes the GRASP computer aided design system for modelling and evaluating industrial robot workplaces. GRASP satisfies a range of simulation needs within the context of designing, implementing and operating industrial robotic systems. The GRASP software may be used to investigate robots operating by themselves or, more likely, as part of an integrated cell. Facilities within GRASP assist workplace layout, position and velocity evaluations, clash detection and co-ordination between items. A robot library exists and is being extended, and an embryo off-line programming facility has been used under restricted conditions. GRASP has been used to help solve a wide range of practical industrial robot problems and has proved itself technically, and also as an educational tool, by showing how a proposed system would operate. The paper describes the facilities within GRASP.     

The future for wool

Single dimensional projections are inadequate to determine what the future holds. A wide variety of economic, technological and social changes will influence the future environment for wool and other fibres. In these circumstances, and bearing in mind that, for planning purposes, strategic assumptions are often markedly superior to firm predictions, a set of alternative scenarios is a useful instrument. Consideration of some brief scenarios leads to the conclusion that wool's future depends, in no small measure, upon the actions of the wool-growing industry and governments of wool exporting countries.     

A multivariate regression model for continuous genetic traits

Summary.  In many commonly used models for multivariate traits, the likelihood is specified as a mixture of nested sums of products over the unobserved genotypes of all the family members, in which the familial covariance matrices vary in size and structure for different families, and their sizes can be immense for large family units. These issues pose computational difficulties in many applications. Bonney's compound regressive model for univariate traits simplifies the familial covariance structure and reduces the mixture of nested sums only to the parent–offspring level, thus enhancing computation significantly. This model has been extended to the multivariate case in the absence of unobserved genotypes. Here, we further extend this model to incorporate major genes, covariates and multiple loci. As is typical in practice, this causes new computational difficulties. We study the computational issues and explore the behaviour of this extended model.     

Compound adaptive GPU design for clinical trials

In clinical trials, several competing treatments are often carried out in the same trial period. The goal is to assess the performances of these different treatments according to some optimality criterion and minimize risks to the patients in the entire process of the study. For this, each coming patient is allocated sequentially to one of the treatments according to a mechanism defined by the optimality criterion. In practice, sometimes different optimality criteria, or the same criterion with different regimes, need to be considered to assess the treatments in the same study, so that each mechanism is also evaluated through the trail study. In this case, the question is how to allocate the treatments to the incoming patients so that the criteria/mechanisms of interest are assessed during the trail process, and the overall performance of the trial is optimized under the combined criteria or regimes. In this paper, we consider this problem by investigating a compound adaptive generalized Pólya urn design. Basic asymptotic properties of this design are also studied.