Experimental design: Introduction to methods and applications of the dialogue system CADEMO.
(Statistische Versuchsplanung: Einführung in die Methoden und Anwendung des Dialogsystems CADEMO.)

*(German)*Zbl 0802.62064
Stuttgart: Gustav Fischer. x, 386 S. + Diskette (1992).

The basic purpose of this book is to present an introduction to the applications of the dialogue system CADEMO (“Computer Aided Design of Experiments and Modelling”) which has been developed by the authors among others during the last decade. CADEMO is written for IBM PC compatibles. It is one of the first expert systems for experimental design and model choice which attempts to fill the gap left by the huge variety of statistical software packages which are available for the statistical analysis of data. The material of the book is arranged as follows (in rough translation):

Part I: Foundations of experimental design: 1. Introduction (3 pages); 2. Principles of experimental design (27 pages); 3. Allocation of experiments (50 pages); 4. Principles of optimal design (25 pages); 5. Sample size determination (68 pages); 6. Model choice and robustness (5 pages). By the authors’ intention the mathematical level is close to the minimum necessary for using the dialogue system. Also the treatment of problems is restricted to those topics which are already implemented in the system.

Part II: Experimental design with the aid of CADEMO: 1. The dialogue system CADEMO – description and basic structure (30 pages); 2. Description of generally applicable module complexes (45 pages); 3. Description of some module complexes with particular applications (73 pages). The handling of the system is illustrated by a large number of screen printouts. It should be noted that also the dialogues of the system are in German. Several case studies give a flavour how the system could be used.

The text is augmented by an appendix on the basics of probability and an index with more than 500 entries. About one third of the approximately 160 references is linked with those scientists developing CADEMO. The demonstration disk included gives a limited impression of the capabilities of the dialogue system. Finally, a dialogue system like CADEMO can only be used as a device to treat standardized problems. But it can not be a surrogate for the knowledge of the statistical background.

Part I: Foundations of experimental design: 1. Introduction (3 pages); 2. Principles of experimental design (27 pages); 3. Allocation of experiments (50 pages); 4. Principles of optimal design (25 pages); 5. Sample size determination (68 pages); 6. Model choice and robustness (5 pages). By the authors’ intention the mathematical level is close to the minimum necessary for using the dialogue system. Also the treatment of problems is restricted to those topics which are already implemented in the system.

Part II: Experimental design with the aid of CADEMO: 1. The dialogue system CADEMO – description and basic structure (30 pages); 2. Description of generally applicable module complexes (45 pages); 3. Description of some module complexes with particular applications (73 pages). The handling of the system is illustrated by a large number of screen printouts. It should be noted that also the dialogues of the system are in German. Several case studies give a flavour how the system could be used.

The text is augmented by an appendix on the basics of probability and an index with more than 500 entries. About one third of the approximately 160 references is linked with those scientists developing CADEMO. The demonstration disk included gives a limited impression of the capabilities of the dialogue system. Finally, a dialogue system like CADEMO can only be used as a device to treat standardized problems. But it can not be a surrogate for the knowledge of the statistical background.

Reviewer: R.Schwabe (Berlin)

##### MSC:

62Kxx | Design of statistical experiments |

62-04 | Software, source code, etc. for problems pertaining to statistics |

62-01 | Introductory exposition (textbooks, tutorial papers, etc.) pertaining to statistics |