Report 2002: Statistics Group (ST)
Solbjerg Plads 3
DK-2000 Frederiksberg
Head of Group: Hans Kurt Kvist
Telefon: +45 3815 3515
Fax: +45 3815 2500
E-mail: ma.mes@cbs.dk
DK-2000 Frederiksberg
Head of Group: Hans Kurt Kvist
Telefon: +45 3815 3515
Fax: +45 3815 2500
E-mail: ma.mes@cbs.dk
Statistics is about drawing conclusions, i.e. making decisions under conditions of uncertainty. Therefore, the discipline, its principles and methods are applicable to all aspects of corporate management.
Highlights of 2002
- Benchmarking, i.e. comparison of units in terms of efficiency. A number of "(production) units" are considered with a view to comparison of their efficiency. The comparison is internal as various principles lead to identification of a set of efficient units to be compared with the remaining, non-efficient units. The degree of inefficiency is quantified in a number of dimensions, and its dependency on characteristics of the units involved is studied. The goal is to achieve an understanding of the conditions particularly influencing inefficiency in order to obtain a tool leading to a higher degree of efficiency. This work has been going on for a couple of years in cooperation with the Universities of Copenhagen, Oxford and Toronto.
- Single-source data / hierarchical data: The main project concerns the use of statistical models in marketing and is conducted in cooperation with the advertising research forum of the Department of Marketing. It centres around the study of the effect of advertising on individual consumers. Modern data retrieval techniques, e.g. electronic registrations, give rise to very large amounts of data and special-structure data, i.e. multi-panel, hierarchical, multilevel or clustered data. Common methods such as logistic regression are not adequate for data with such a structure and it therefore has to be generalised in various ways. This type of data occurs in several other situations: Educational research, analysis of company data with a hierarchical structure: Group -> subsidiary -> divisions, etc.
- Supply Chain Management: This project was completed in the autumn of 2002. It concerned modelling and solution of planning problems in a supply chain characterised by uncertainties with regard to incoming orders, order sizes, delivery time and the quality of the order delivered.
- Internal seminars, which are held every fortnight during a semester, are widely advertised and some are therefore well-attended. A case in point is Guillermina Eslava's "The Analysis of Binary Clustered Sample Survey Data" which attracted more than 70 participants, including WHO representatives who expressed great interest in the subject. The somewhat cryptic title covers a problem which is fairly analogous to the single source issue of marketing, only here it concerns a large amount of data on the health of young children in Mexico. This information is relevant to primary health sector "management". In countries like Mexico, random sample data will obtain a hierarchical structure as a reflection of the retrieval procedure: rural district -> regions -> counties -> parishes -> households -> children. Guillermina Eslava visited the Group from 2001 to 2002 at the expense of the Mexican government in order to analyse this data set and because of the Group's interest in this type of data.
Academic profile
The academic profile of the Group naturally reflects the discipline of Theoretical Statistics in which the Group staff are deeply rooted. However, we see it in a somewhat wider perspective, including the application of formal quantitative (i.e. mathematical) methods to solve problems related to the operation of enterprises. We thus aim at a profile which is not only directed towards traditional operations analysis in accordance with the formal element of Operations Management, but also towards proper mathematical models of (business) economics "phenomena". Such issues are part of the above highlights and occur in (almost) all business economics disciplines, from modern financial theory via mathematical economics to organisational theory.
Research strategy
The strategy aims at the application and development of mathematical/statistical models capable of contributing to the solution of corporate control problems in a general sense. The strategy is rarely implemented by means of major all-inclusive projects but rather through an effort to solve concrete problems. The latter may be the result of or give rise to PhD projects, and they are often inspired by contacts with researchers in specific disciplines of business economics.
The strategy aims at the application and development of mathematical/statistical models capable of contributing to the solution of corporate control problems in a general sense. The strategy is rarely implemented by means of major all-inclusive projects but rather through an effort to solve concrete problems. The latter may be the result of or give rise to PhD projects, and they are often inspired by contacts with researchers in specific disciplines of business economics.
Research results of 2002 are described summarily in Figure 1 below: There are modest variations compared with previous periods. This reflects the unchanged staff and almost unchanged funding during the period in question (see Figure 2).
Figure 1
Figure 2
Research relations to practice
Relations to practice are important and normal in two respects. For one thing, it is to be hoped that ultimately the problems and solutions that we study will benefit practice, and for another, practice may give rise to new problems and/or solutions. Considering the nature of the discipline, practice may be interpreted broadly. Thus, the Group receives many requests for advice and guidance from other research environments, not least from graduate and PhD students working on empirical projects. With regard to non-academic environments, we traditionally receive a number of informal requests from enterprises/institutions. Two illustrative examples: The occurrence of industrial accidents (from a major contracting company) and changes in the sampling procedure in connection with auditing transactions in a very large international organisation (from an international firm of accountants).
Relations to practice are important and normal in two respects. For one thing, it is to be hoped that ultimately the problems and solutions that we study will benefit practice, and for another, practice may give rise to new problems and/or solutions. Considering the nature of the discipline, practice may be interpreted broadly. Thus, the Group receives many requests for advice and guidance from other research environments, not least from graduate and PhD students working on empirical projects. With regard to non-academic environments, we traditionally receive a number of informal requests from enterprises/institutions. Two illustrative examples: The occurrence of industrial accidents (from a major contracting company) and changes in the sampling procedure in connection with auditing transactions in a very large international organisation (from an international firm of accountants).
The Group only enters sporadic major and more formal cooperation agreements. However, we did enter two such agreements at the end of 2002. One, which is very modest, concerns the role of members of the Executive Board elected by the employees in connection with company management and involves the Department of Finance, the Statistics Group and a number of trade unions.
The other agreement is much more extensive and ambitious. It covers studies of consumer preferences for various types of foods. It was entered between the Royal Veterinary and Agricultural University, The Directorate for Food, Fisheries and Agri Business under the Danish Ministry of Food, Agriculture and Fisheries, the Danish Meat Research Institute, Danish Crown and TiCan. The Group is in charge of planning and analysis of consumer preference studies. In terms of methodology, it intends to apply and continue the development of models to measure consumer preferences. The interest of external participants was aroused, i.a. by previous work on measuring consumer preferences conducted by the Group.
Last updated by Anders Krag 09/02/2005
