Principles of multivariate analysis. A user’s perspective.

*(English)*Zbl 0678.62001
Oxford Statistical Science Series, 3. Oxford: Clarendon Press. xxi, 563 p. £65.00 (1988).

The book is aimed at the (potential) user of multivariate methods. To cater best for an applied readership the author organized the text according to a problem-oriented format. This sets it in contrast to many other texts on multivariate methods which are in a technique-oriented format. The book is for those who are not willing to cope with full mathematical treatment. But there is enough theory explained that it is not a cook-book. It rather keeps the theory in its place.

It is problem oriented in that way that it proceeds as a researcher is likely to do when analysing real data sets. The first part is on ‘looking at multivariate data’. A lot of graphics is included here, and many examples are also given. Examples continue to be presented in the whole text, whereas graphics are missed in later parts.

Part two presents basics on ‘samples, populations, and models’. Essentially the reader finds here the distribution theory necessary for inference. ‘Analysing ungrouped data’ and ‘analysing grouped data’ are the next two parts. Grouping is understood as the presence of a group structure as when individuals have been sampled from different populations. The last part is concerned with ‘analysing association’ among variables. Multivariate regression, factor analysis and latent- structure analysis are special topics.

The book is large, but it is a wide field which should be covered. In that the author has been rather successful. The missing of any hint to existing statistical software is surprising nowadays, especially in an applied text. Overall it is an excellent book, well written and very recommondable to the beginning user of multivariate methods. The more experienced would be expected to like to have a traditionally organized text within reach too, because it is a bit difficult to find a method when looking for a special one.

It is problem oriented in that way that it proceeds as a researcher is likely to do when analysing real data sets. The first part is on ‘looking at multivariate data’. A lot of graphics is included here, and many examples are also given. Examples continue to be presented in the whole text, whereas graphics are missed in later parts.

Part two presents basics on ‘samples, populations, and models’. Essentially the reader finds here the distribution theory necessary for inference. ‘Analysing ungrouped data’ and ‘analysing grouped data’ are the next two parts. Grouping is understood as the presence of a group structure as when individuals have been sampled from different populations. The last part is concerned with ‘analysing association’ among variables. Multivariate regression, factor analysis and latent- structure analysis are special topics.

The book is large, but it is a wide field which should be covered. In that the author has been rather successful. The missing of any hint to existing statistical software is surprising nowadays, especially in an applied text. Overall it is an excellent book, well written and very recommondable to the beginning user of multivariate methods. The more experienced would be expected to like to have a traditionally organized text within reach too, because it is a bit difficult to find a method when looking for a special one.

Reviewer: R.Schlittgen

##### MSC:

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

62Hxx | Multivariate analysis |

62-07 | Data analysis (statistics) (MSC2010) |