Computer-Aided Multivariate Analysis

by Afifi, A. A.; Clark, Virginia A.

• ISBN13: 9780412730603
• ISBN10: 041273060X
• Edition: 3rd
• Format: Hardcover
• Copyright: 1996-06-01
• Publisher: Chapman & Hall

Summary

Increasingly, researchers need to perform multivariate statistical analyses on their data. Unfortunately, a lack of mathematical training prevents many from taking advantage of these advanced techniques, in part, because books focus on the theory and neglect to explain how to perform and interpret multivariate analyses on real-life data. For years, Afifi and Clark's Computer-Aided Multivariate Analysis has been a welcome exception-helping researchers choose the appropriate analyses for their data, carry them out, and interpret the results. Only a limited knowledge of statistics is assumed, and geometrical and graphical explanations are used to explain what the analyses do. However, the basic model is always given, and assumptions are discussed. Reflecting the increased emphasis on computers, the Third Edition includes three additional statistical packages written for the personal computer. The authors also discuss data entry, database management, data screening, data transformations, as well as multivariate data analysis. Another new chapter focuses on log-linear analysis of multi-way frequency tables. Students in a wide range of fields-ranging from psychology, sociology, and physical sciences to public health and biomedical science-will find Computer-Aided Multivariate Analysis especially informative and enlightening.

Table of Contents

Preface	xiii	(4)
Preface to the second edition	xvii	(2)
Preface to the first edition	xix
Part One Preparation for Analysis	1	(82)
1 What is multivariate analysis?	3	(9)
1.1 How is multivariate analysis defined?	3	(1)
1.2 Examples of studies in which multivariate analysis is useful	3	(3)
1.3 Multivariate analyses discussed in this book	6	(3)
1.4 Organization and content of the book	9	(2)
References	11	(1)
2 Characterizing data for future analyses	12	(9)
2.1 Variables: their definition, classification and use	12	(1)
2.2 Defining statistical variables	12	(1)
2.3 How variables are classified: Stevens's classification system	13	(3)
2.4 How variables are used in data analysis	16	(1)
2.5 Examples of classifying variables	17	(1)
2.6 Other characteristics of data	18	(1)
Summary	18	(1)
References	19	(1)
Further reading	19	(1)
Problems	19	(2)
3 Preparing for data analysis	21	(27)
3.1 Processing the data so they can be analyzed	21	(1)
3.2 Choice of computer for statistical analysis	22	(1)
3.3 Choice of a statistical package	23	(5)
3.4 Techniques for data entry	28	(6)
3.5 Data management for statistics	34	(6)
3.6 Data example: Los Angeles depression study	40	(3)
Summary	43	(2)
References	45	(1)
Further reading	46	(1)
Problems	46	(2)
4 Data screening and data transformation	48	(23)
4.1 Making transformations and assessing normality and independence	48	(1)
4.2 Common transformations	48	(6)
4.3 Assessing the need for and selecting a transformation	54	(10)
4.4 Assessing independence	64	(3)
Summary	67	(1)
References	67	(1)
Further reading	68	(1)
Problems	68	(3)
5 Selecting appropriate analyses	71	(12)
5.1 Which analyses?	71	(1)
5.2 Why selection of analyses is often difficult	71	(1)
5.3 Appropriate statistical measures under Stevens's classification	72	(4)
5.4 Appropriate multivariate analyses under Stevens's classification	76	(3)
Summary	79	(1)
References	79	(1)
Further reading	80	(1)
Problems	80	(3)
Part Two Applied Regression Analysis	83	(142)
6 Simple linear regression and correlation	85	(39)
6.1 Using linear regression and correlation to examine the relationship between two variables	85	(1)
6.2 When are regression and correlation used?	85	(1)
6.3 Data example	86	(2)
6.4 Description of methods of regression: fixed-X case	88	(5)
6.5 Description of methods of regression and correlation: variable-X case	93	(1)
6.6 Interpretation of results: fixed-X case	94	(2)
6.7 Interpretation of results: variable-X case	96	(4)
6.8 Further examination of computer output	100	(8)
6.9 Robustness and transformations for regression analysis	108	(3)
6.10 Other options in computer programs	111	(1)
6.11 Special applications of regression	112	(3)
6.12 Discussion of computer programs	115	(2)
6.13 What to watch out for	117	(1)
Summary	118	(1)
References	118	(2)
Further reading	120	(1)
Problems	121	(2)
7 Multiple regression and correlation	124	(42)
7.1 Using multiple linear regression to examine the relationship between one dependent variable and multiple independent variables	124	(1)
7.2 When are multiple regression and correlation used?	125	(1)
7.3 Data example	125	(3)
7.4 Description of techniques: fixed-X case	128	(2)
7.5 Description of techniques: variable-X case	130	(7)
7.6 How to interpret the results: fixed-X case	137	(3)
7.7 How to interpret the results: variable-X case	140	(3)
7.8 Residual analysis and transformations	143	(5)
7.9 Other options in computer programs	148	(6)
7.10 Discussion of computer programs	154	(3)
7.11 What to watch out for	157	(3)
Summary	160	(1)
References	160	(1)
Further reading	161	(1)
Problems	162	(4)
8 Variable selection in regression analysis	166	(31)
8.1 Using variable selection techniques in multiple regression analysis	166	(1)
8.2 When are variable selection methods used?	166	(1)
8.3 Data example	167	(3)
8.4 Criteria for variable selection	170	(3)
8.5 A general F test	173	(2)
8.6 Stepwise regression	175	(6)
8.7 Subset regression	181	(4)
8.8 Discussion of computer programs	185	(2)
8.9 Discussion and extensions	187	(4)
8.10 What to watch out for	191	(2)
Summary	193	(1)
References	193	(1)
Further reading	194	(1)
Problems	194	(3)
9 Special regression topics	197	(28)
9.1 Special topics in regression analysis	197	(1)
9.2 Missing values in regression analysis	197	(5)
9.3 Dummy variables	202	(7)
9.4 Constraints on parameters	209	(3)
9.5 Methods for obtaining a regression equation when multicollinearity is present	212	(2)
9.6 Ridge regression	214	(5)
Summary	219	(1)
References	220	(1)
Further reading	221	(1)
Problems	221	(4)
Part Three Multivariate Analysis	225	(218)
10 Canonical correlation analysis	227	(16)
10.1 Using canonical correlation analysis to analyze two sets of variables	227	(1)
10.2 When is canonical correlation analysis used?	227	(1)
10.3 Data example	228	(1)
10.4 Basic concepts of canonical correlation	229	(5)
10.5 Other topics related to canonical correlation	234	(3)
10.6 Discussion of computer programs	237	(2)
10.7 What to watch out for	239	(1)
Summary	240	(1)
References	241	(1)
Further reading	241	(1)
Problems	241	(2)
11 Discriminant analysis	243	(38)
11.1 Using discriminant analysis to classify cases	243	(1)
11.2 When is discriminant analysis used?	244	(1)
11.3 Data example	245	(1)
11.4 Basic concepts of classification	246	(7)
11.5 Theoretical background	253	(2)
11.6 Interpretation	255	(4)
11.7 Adjusting the value of the dividing point	259	(3)
11.8 How good is the discriminant function?	262	(3)
11.9 Testing for the contributions of classification variables	265	(1)
11.10 Variable selection	266	(1)
11.11 Classification into more than two groups	267	(2)
11.12 Use of canonical correlation in discriminant function analysis	269	(3)
11.13 Discussion of computer programs	272	(3)
11.14 What to watch out for	275	(1)
Summary	276	(1)
References	277	(1)
Further reading	277	(1)
Problems	278	(3)
12 Logistic regression	281	(25)
12.1 Using logistic regression to analyze a dichotomous outcome variable	281	(1)
12.2 When is logistic regression used?	281	(1)
12.3 Data example	282	(1)
12.4 Basic concepts of logistic regression	283	(2)
12.5 Interpretation: categorical variables	285	(3)
12.6 Interpretation: continuous and mixed variables	288	(1)
12.7 Refining and evaluating logistic regression analysis	289	(7)
12.8 Applications of logistic regression	296	(3)
12.9 Discussion of computer programs	299	(2)
12.10 What to watch out for	301	(1)
Summary	302	(1)
References	302	(1)
Further reading	303	(1)
Problems	304	(2)
13 Regression analysis using survival data	306	(24)
13.1 Using survival analysis to analyze time-to-event data	306	(1)
13.2 When is survival analysis used?	306	(1)
13.3 Data examples	307	(2)
13.4 Survival functions	309	(5)
13.5 Common distributions used in survival analysis	314	(3)
13.6 The log-linear regression model	317	(2)
13.7 The Cox proportional hazards regression model	319	(1)
13.8 Some comparisons of the log-linear, Cox and logistic regression models	320	(4)
13.9 Discussion of computer programs	324	(2)
13.10 What to watch out for	326	(1)
Summary	327	(1)
References	328	(1)
Further reading	328	(1)
Problems	328	(2)
14 Principal components analysis	330	(24)
14.1 Using principal components analysis to understand intercorrelations	330	(1)
14.2 When is principal components analysis used?	330	(1)
14.3 Data example	331	(2)
14.4 Basic concepts of principal components analysis	333	(3)
14.5 Interpretation	336	(9)
14.6 Use of principal components analysis in regression and other applications	345	(3)
14.7 Discussion of computer programs	348	(2)
14.8 What to watch out for	350	(1)
Summary	351	(1)
References	352	(1)
Further reading	352	(1)
Problems	352	(2)
15 Factor analysis	354	(27)
15.1 Using factor analysis to examine the relationship among P variables	354	(1)
15.2 When is factor analysis used?	354	(1)
15.3 Data example	355	(1)
15.4 Basic concepts of factor analysis	356	(2)
15.5 Initial factor extraction: principal components analysis	358	(4)
15.6 Initial factor extraction: iterated principal components	362	(3)
15.7 Factor rotations	365	(6)
15.8 Assigning factor scores to individuals	371	(1)
15.9 An application of factor analysis to the depression data	372	(2)
15.10 Discussion of computer programs	374	(2)
15.11 What to watch out for	376	(1)
Summary	377	(1)
References	378	(1)
Further reading	379	(1)
Problems	379	(2)
16 Cluster analysis	381	(29)
16.1 Using cluster analysis to group cases	381	(1)
16.2 When is cluster analysis used?	381	(2)
16.3 Data example	383	(2)
16.4 Basic concepts: initial analysis and distance measures	385	(6)
16.5 Analytical clustering techniques	391	(7)
16.6 Cluster analysis for financial data set	398	(6)
16.7 Discussion of computer programs	404	(2)
16.8 What to watch out for	406	(1)
Summary	406	(1)
References	407	(1)
Further reading	407	(1)
Problems	408	(2)
17 Log-linear analysis	410	(33)
17.1 Using log-linear models to analyze categorical data	410	(1)
17.2 When is log-linear analysis used?	410	(1)
17.3 Data example	411	(2)
17.4 Notation and sample considerations	413	(2)
17.5 Tests of hypotheses and models for two-way tables	415	(4)
17.6 Example of a two-way table	419	(2)
17.7 Models for multiway tables	421	(4)
17.8 Tests of hypotheses for multiway tables: exploratory model building	425	(6)
17.9 Tests of hypotheses: specific models	431	(1)
17.10 Sample size issues	432	(2)
17.11 The logit model	434	(3)
17.12 Discussion of computer programs	437	(2)
17.13 What to watch out for	439	(1)
Summary	440	(1)
References	440	(1)
Further reading	441	(1)
Problems	441	(2)
Appendix A Lung function data	443	(3)
Table A.1 Code book for lung function data set	444	(1)
Table A.2 Lung function data set	445	(1)
Further reading	446	(1)
Appendix B Lung cancer survival data	446	(3)
Table B.1 Lung cancer data	447	(2)
Index	449