Policies for economic growth

Summary The old growth theory of the 1950s led to certain conclusions about the sorts of economic policies that would promote economic growth, and also about their limitations. The new growth theory of the 1980s makes much stronger assumptions and leads to correspondingly stronger conclusions about the scope of growth-promoting policy. This article argues that: (1) empirical work so far has neither confirmed nor denied the strong assumptions underlying the new theory; (2) the theory is worth pursuing because of its intrinsic interest and the possibilities it opens up; (3) whatever the final verdict on the new theory, both theory and evidence support the belief that significant long-run gains, even if not permanent changes in the growth rate, can be achieved by increased investment in the broadest sense, including human capital, technological knowledge, and industrial plant and equipment.Fifth Tinbergen Lecture delivered on October 4, 1991, in Amersfoort for the Royal Netherlands Economic Association     

The role and place of research and intellectual discourse in the reproduction of social relations of racial domination in South Africa

The South African research community which undertakes all research activity in the social and natural sciences, with and without state and corporate sponsorship, draws its membership mainly from the dominant social group. In this country, the dominant group is both economically and racially determined. Consequently, the white minority dominates the research community and intellectual discourse as it does other socio‐economic and political spheres of society. This situation guarantees the constant reproduction and perpetuation of the social relations of racial domination.

As an agent that generates knowledge and new ideas, research as an academic and intellectual tool of enquiry is an instrument of social control, producing new concepts, language and theoretical abstractions which are not accessible to those outside its multi‐farious disciplines. Insofar as the largest proportion of practitioners of these specialised disciplines is drawn from the dominant group, research has itself become a pivotal part of the dominant ideology. Its role is inevitably and inextricably bound up with the processes of systematic reproduction of the relations of domination.

The aim of this viewpoint is therefore to explore various ways in which research bodies and intellectual discourse in general in South Africa can be deracialised and be made more representative of the social make‐up of society.     

Developing a tutoring module for presentation of computerized patient management problems

The simulation problem is one of the most commonly used formats in computer-assisted instruction (CAI). In medical education, a patient simulation problem is known as a patient management problem (PMP). A number of computerized PMP systems are currently available. However, to date, there is no record of PMP systems having been developed using artificial intelligence. This is noteworthy since artificial intelligence techniques could help in the generation of more effective and intelligent instructional systems.

This paper analyzes and evaluates existing instructional simulation software in clinical medicine in terms of effective educational design attributes. Components of effective PMP systems thus identified are incorporated into a computerized instructional system which uses artificial intelligence techniques for teaching problem-solving and diagnostic skills to undergraduate medical students.     

Polyhedral Combinatorics in Combinatorial Optimization

Polyhedral combinatorics is a subarea of combinatorial optimization of increasing practical importance. It deals with the application of the theory of linear systems and linear algebra to combinatorial problems. The paper is not intended as a survey on polyhedral combinatorics but it reviews some of the main concepts and proof techniques.     

Distribution-free tests for comparing several treatments with a control

Abstract  In this paper we propose distribution-free treatments versus control tests for the one- and two-way classifications. The null distribution properties are shown to be equivalent to those of well-known statistics for which critical values have been extensively tabled. Asymptotic relative efficiencies with respect to corresponding normal theory competitors are derived and optimal designs for the allocation of experimental units to the treatments and control are obtained.     

The occurrence of outliers in the explanatory variable considered in an errors-in-variables framework

Summary The problem of estimating the slope of a linear relationship between two jointly normally distributed random variables is considered when outliers may occur in the explanatory variable. It will be studied as a special case of an errors-in-variables problem where the explanatory variable is measured which a nonnormally distributed error. In this more general model and under certain conditions a consistent estimator can be given with a normal limiting distribution. Applications to cases of outliers in the explanatory variable will be presented.