nach oben

2010 | Buch

Statistical Analysis of Management Data

verfasst von: Hubert Gatignon

Verlag: Springer New York

Enthalten in: Springer Professional "Wirtschaft+Technik" , Springer Professional "Technik" , Springer Professional "Wirtschaft"

Einloggen, um Zugang zu erhalten

Über dieses Buch

Statistical Analysis of Management Data provides a comprehensive approach to multivariate statistical analyses that are important for researchers in all fields of management, including finance, production, accounting, marketing, strategy, technology, and human resources. This book is especially designed to provide doctoral students with a theoretical knowledge of the concepts underlying the most important multivariate techniques and an overview of actual applications. It offers a clear, succinct exposition of each technique with emphasis on when each technique is appropriate and how to use it. This second edition, fully revised, updated, and expanded, reflects the most current evolution in the methods for data analysis in management and the social sciences. In particular, it places a greater emphasis on measurement models, and includes new chapters and sections on:

confirmatory factor analysis

canonical correlation analysis

cluster analysis

analysis of covariance structure

multi-group confirmatory factor analysis and analysis of covariance structures.

Featuring numerous examples, the book may serve as an advanced text or as a resource for applied researchers in industry who want to understand the foundations of the methods and to learn how they can be applied using widely available statistical software.

Inhaltsverzeichnis

Frontmatter

Chapter 1. Introduction

Abstract

This book covers multivariate statistical analyses that are important for researchers in all fields of management, whether finance, production, accounting, marketing, strategy, technology, or human resources management. Although multivariate statistical techniques such as those described in this book play key roles in fundamental disciplines of the social sciences (e.g., economics and econometrics or psychology and psychometrics), the methodologies particularly relevant and typically used in management research are the focus of this study.

Hubert Gatignon

Chapter 2. Multivariate Normal Distribution

Abstract

In this chapter, we define univariate and multivariate normal distribution density functions and then we discuss tests of differences of means for multiple variables simultaneously across groups.

Hubert Gatignon

Chapter 3. Reliability Alpha, Principle Component Analysis, and Exploratory Factor Analysis

Abstract

In this chapter, we will discuss the issues involved in building measures or scales. We focus on two types of analysis: (1) the measurement of reliability with Cronbach’s alpha and (2) the verification of unidimensionality using factor analysis. In this chapter, we concentrate on exploratory factor analysis and we only introduce the notion of confirmatory factor analysis in this chapter. The next chapter develops in detail the confirmatory factor analytic model and examines the measures of convergent and discriminant validity.

Hubert Gatignon

Chapter 4. Confirmatory Factor Analysis

Abstract

As mentioned in the last chapter, a measurement model of the type illustrated in Fig. 4.1 is assumed in confirmatory factor analysis.

Hubert Gatignon

Chapter 5. Multiple Regression with a Single Dependent Variable

Abstract

This chapter covers the principles which are basic to understanding properly the issues involved in the analysis of management data. This chapter cannot constitute the depth which goes into a specialized econometric book. It is however designed to provide the elements of econometric theory essential for a researcher to develop and evaluate regression models. Multiple regression is not a multivariate technique in a strict sense in that a single variable is the focus of the analysis: a single dependent variable. Nevertheless, the multivariate normal distribution is involved in the distribution of the error term, which, combined with the fact that there are multiple independent or predictor variables, leads to considering simple multiple regression within the domain of multivariate data analysis techniques.

Hubert Gatignon

Chapter 6. System of Equations

Abstract

In this chapter, we consider the case where several dependent variables are explained by linear relationships with other variables. Independent analysis of each relationship by ordinary least squares Ordinary Least Squares OLS could result in incorrect statistical inferences either because the estimation is not efficient (a simultaneous consideration of all the explained variables may lead to more efficient estimators for the parameters) or may be biased in cases where the dependent variables influence each other.

Hubert Gatignon

Chapter 7. Canonical Correlation Analysis

Abstract

In canonical correlation analysis, the objective is to relate a set of dependent or criterion variables to another set of independent or predictor variables. In order to do that, we find a scalar, defined as a linear combination of the dependent variables, as well as a scalar defined as a linear combination of the independent variables. The criterion used to judge the relationship between this set of independent variables with the set of dependent variables is simply the correlation between the two scalars. Canonical correlation analysis then consists in finding the weights to apply to the linear combinations of the independent and dependent variables that will maximize the correlation coefficient between those two linear combinations. The problem can be represented graphically as in Fig. 7.1

Hubert Gatignon

Chapter 8. Categorical Dependent Variables

Abstract

In this chapter, we consider statistical models to analyze variables where the numbering does not have any meaning and, in particular, where there is no relationship between one level of the variable and another level. In these cases, we are typically trying to establish whether it is possible to explain with other variables the level observed of the criterion variable. The chapter is divided in two parts. The first part presents discriminant analysis, which is a traditional method in multivariate statistical analysis. The second part introduces quantal choice statistical models. The models are described, as well as their estimation. Their measures of fit are also discussed.

Hubert Gatignon

Chapter 9. Rank-Ordered Data

Abstract

When the criterion variable is defined on an ordinal scale, the typical analyses based on correlations or covariances are not appropriate. The methods described in Chapter 6 do not use the ordered nature of the data and, consequently, do not use all the information available. In this chapter, we present methodologies that take the ordinal property of the dependent variable into account.

Hubert Gatignon

Chapter 10. Error in Variables – Analysis of Covariance Structure

Abstract

In this chapter, we bring together the notions of measurement error discussed in Chapters 3 and 4 with the structural modeling of simultaneous relationships presented in Chapter 6. We will demonstrate that a bias is introduced when estimating the relationship between two variables measured with error if that measurement error is ignored. We will then present a methodology for estimating the parameters of structural relationships between variables which are not observed directly: analysis of covariance structures. We will discuss especially the role of the measurement model as discussed in the chapter on the confirmatory factor analytic model.

Hubert Gatignon

Chapter 11. Cluster Analysis

Abstract

The objective of cluster analysis is to group observations (e.g., individuals) in such a way that the groups formed are as homogeneous as possible within each group and as different as possible across groups.

Hubert Gatignon

Chapter 12. Analysis of Similarity and Preference Data

Abstract

Similarity data in management research are typically collected in order to understand the underlying dimensions determining perceptions of stimuli such as brands or companies. One advantage of such data is that it is cognitively easier for respondents to provide subjective assessments of the similarity between objects than to rate these objects on a number of attributes which they may not even be aware of. Furthermore, when asking respondents to rate objects on attributes, the selection of the attributes proposed may influence the results while, in fact, it is not clear that these attributes are the relevant ones.

Hubert Gatignon

Chapter 13. Appendices

Abstract

The data sets described below can be downloaded from the web at: http://www.insead.edu/˜gatignon. Three different kinds of information, which correspond to typically available data about markets, are provided for analysis: industry, panel, and survey data. In addition, scanner data are provided for a product category in the form typically available in practice.

Hubert Gatignon

Backmatter

Titel: Statistical Analysis of Management Data
verfasst von: Hubert Gatignon
Verlag: Springer New York
Electronic ISBN: 978-1-4419-1270-1
Print ISBN: 978-1-4419-1269-5
DOI: https://doi.org/10.1007/978-1-4419-1270-1