Proceedings

  • The learning and teaching support network is a programme funded to promote good practice in teaching and learning in UK higher education. Subject networks have been established in twenty-four different areas, including one for mathematics, statistics and operational research. Among all the different activities of this network, the web, of course, offers a rich source of primary material and a convenient means of dissemination. The web provides a vast collection of material on every subject known to man, including statistics. The aim of the ltsn msor web site is to offer a convenient and filtered gateway to a wide variety of teaching and learning material. This paper describes some of the resources available in statistics in particular. Some of the organisational aspects of setting up the website are also mentioned.

  • Pharmaceutical companies are constantly asked for information by government agencies, market research companies and often carry out their own investigations. However there has been no definitive independent source of information about field-based personnel in the pharmaceutical and healthcare industry. The authors report on the first ever survey of the UK medical sales field force and demonstrate the unique data interrogation tool developed to enable analysis of the data collected including the remuneration, values and perception of the sales force.

  • The availability of comprehensive population registries in Scandinavian countries has facilitated extensive work in epidemiology on associations between risk factors and disease. The area has attracted many statisticians with no previous training in epidemiology. Experience has shown that some statisticians find it difficult to adapt to the practical challenges of this work. Not only is a basic understanding required of the statistical methods involved, but a particular cautious attitude is needed in the interpretation of epidemiological data with inherent uncertainties. An ability to communicate efficiently with coworkers is also essential. Yet the statistician must frequently deal with issues of a biological nature, in addition to technical aspects of data processing. It is difficult to take all these requirements into account in the education of professional statisticians. It is argued that the components not directly connected with statistics should still be integrated into the statistical training of future professionals. If statistics courses include a sufficient amount of relevant data analytic work, the students will be exposed to many of the challenges experienced in epidemiology.

  • In many complex diseases researchers have observed that neither genetic factors nor environmental factors alone determine the disease. This observation generates the hypothesis that human disease is caused by both genetic and environmental factors that act together. This leads to the concept multifactorial causes of disease. On the other hand, the recent compilation of the draft human genome sequence opened the possibility to detect candidate genes for complex diseases and even to study these in relation with environmental factors. The gene-environmental interaction may not be easy to analyze due to the complex structure that the involved factors may have. These factors have different nature that should be treated at different stages of the study. Particular attention should be paid to the study size and design. Epidemiological studies with particular interest in identifying candidate genes that contribute to complex diseases as well as detection of intergenic or gene-environment interactions require large sample sizes because many variables are studied simultaneously. The larger patient populations ensure that individual subgroups retain adequate power to detect significant results with narrow confidence intervals. In the paper we focus on the advantages/disadvantages of classic multifactorial statistical methods applied to the health sciences and the genome scan.

  • As educators, we should not only aim to provide our students with technical skills, but should also help them develop life skills. In recent times there has been an increasing emphasis on communication skills, application skills and reporting skills, but we possibly have not yet sufficiently articulated the social issues associated with good data collection, analysis and reporting. We also need to demonstrate to the students, and through them to the community, the wide field of applicability of statistical techniques, and the need for viewing events from a numerate point of view (among others), in order to interpret what the events mean. There are many social issues that can and should be raised with our students, which can also be used to illustrate statistical techniques. Examples of this, particularly pertinent in South Africa, are issues such as HIV/AIDS, rights of women, etc. For example, HIV/AIDS can be used to discuss regression on indicator variables (HIV negative, HIV+, then later expand to symptomatic and non-symptomatic). This could then be combined with a few questions about whether the class thinks that mortality tables apply to them. This paper focuses on the questions: do statisticians have a social responsibility to students to include such issues among the technical issues, and what is the best way of doing this?

  • One often hears that "data are not information, information is not knowledge, knowledge is not wisdom". But what will turn data into information, information into knowledge, and knowledge into wisdom? The first two facets of this question are at the core of every university course in statistics. They provide a motivation for understanding statistical description and statistical inference, respectively. It is the third facet, the getting of wisdom, which adds depth, resilience and realism to that understanding, yet its importance is often underrated in professional statistics programs. Crucial to the getting of wisdom in this context is a competence to argue back to a statistic and to criticise a statistical argument. Imparting this competence should be a vital concern in designing the program syllabus. In this paper I argue that, by adding a little to the syllabus, such a program can also aid the statistician in opening up for his/her client the client's own path to statistical knowledge and wisdom. Such a move constructively addresses an abiding social issue: the need to enhance the level of numeracy in our alarmingly innumerate society.

  • This paper describes two versions of a teaching experiment that traced the development of Basotho elementary students' thinking with regard to sample space and probability of an event. The instructional design phase of the teaching experiment was informed by a cognitive framework that describes and predicts Basotho elementary students' growth in probabilistic thinking (Polaki, Lefoka, & Jones, 2000). Twelve students (9-10 year olds) drawn from grades 4 and 5 of an elementary school took part in a six-week instructional program. Analysis of qualitative data revealed, amongst other things, a weak and often unstable part-part schema that was minimally effective in enabling the students to order probabilities in 1-dimensional situations, and a stronger and more stable part-part schema that made it possible for some students to experience greater success at listing complete sets of outcomes, and to order probabilities in 1- and 2-dimensional situations.

  • We interviewed 7th and 9th grade students to explore how they summarized and reasoned about data. The students were near the end of an eight-week collaborative research project in which they analyzed data they had collected on the types and frequencies of animals killed on town roads. During our interviews, students worked with data similar to those they had collected to answer questions we posed about conditions that might affect the number of animals struck by cars. To summarize their data, students tended to use a "modal clump," a range of data in the heart of a distribution of values. These clumps appear to allow students to express simultaneously what is average and how variable the data are. Modal clumps may provide useful beginning points for explorations of more formal statistical ideas of center.

  • As teachers of statistics we know the fundamental components of statistical enquiry, be it classical or exploratory. When we turn the focus on ourselves as statistics educators, we run the risk of forgetting some of the fundamental principles of good research - principles that are broader than carrying out statistical significance tests. In this talk I want to present some examples of research in statistics education to illustrate the stages and outcomes that contribute to results that have a scholarly impact on the statistics education community. As a single teacher with a good idea on how to teach "confidence intervals," I do not expect anyone to pay much attention to me. If I can, however, place my ideas in the context of others' ideas or research on teaching confidence intervals; conduct a study - maybe a case study or a controlled experimental<br>design - that is valid for considering the issue I want to promote in teaching about confidence intervals; and have my results refereed by peers in the field; then I can expect people to pay attention to me.

  • Improving the public's understanding of statistical information requires that producers or reporters of statistical messages are aware of: The nature of people's statistics literacy, The factors that affect the difficulty of statistics-related messages, The existence of individual or group differences in statistics literacy; and The information needs of different target audiences. Implications are discussed regarding the need to prepare different types of communicative products and formulate strategies for dissemination and public education

Pages

register