Structural Equation Modeling: Applications in Ecological and Evolutionary Biology

Cover......Page 1
Haf-title......Page 3
Title......Page 5
Copyright......Page 6
Dedication......Page 7
Contents......Page 9
Contributors......Page 11
Preface......Page 13
Section 1 Theory......Page 17
Introduction......Page 19
Definition and specification of a structural equation model......Page 21
Model identification......Page 24
Model estimation......Page 28
Model assessment and fit......Page 33
Model modification......Page 35
Multisample models......Page 37
The LISREL model and program......Page 39
Example of LISREL analysis......Page 43
References......Page 56
Introduction......Page 58
Measurement error......Page 59
Latent variables......Page 64
Hypothesis testing......Page 68
Modeling complex systems......Page 70
Conclusions......Page 73
Appendix 2.1. The EQS program (Bentler, 1992) used to generate a random sample from a population defined by a model…......Page 74
References......Page 75
Introduction......Page 76
Data......Page 78
Construction of the data simulation......Page 82
Data analysis......Page 85
Measurement model......Page 86
Data1......Page 88
Data2......Page 90
Discussion......Page 92
Appendix 3.1.......Page 96
References......Page 99
Introduction......Page 101
Wright’s invention of path analysis......Page 102
Path models in sociology......Page 106
Theoretical constructs and indicators in sociology......Page 108
Models and structures in econometric thought – the identification problem......Page 109
Estimation of parameters in econometric models......Page 111
Estimation by the full information maximum likelihood method (FIML)......Page 112
Other methods of estimation of econometrics models......Page 113
Model building and tests of significance......Page 114
Charles Spearman and the beginning of exploratory and confirmatory factor analysis......Page 115
Maximum likelihood estimation of factor loadings in exploratory factor analysis......Page 120
Maximum likelihood estimation of factor loadings in confirmatory factor analysis and SEM in general......Page 121
Likelihood ratio tests......Page 122
Identification problems in confirmatory factor analysis......Page 123
General comment on exploratory and confirmatory factor analysis......Page 124
The LISREL model: Keesling, Jöreskog, and Wiley......Page 125
Jöreskog and Wiley’s formulations......Page 126
The extension of the LISREL model to include means......Page 128
Generalized least squares......Page 129
The extension of the LISREL model to include dichotomous and ordered variables......Page 130
Distribution free estimators......Page 131
Indices of fit......Page 132
References......Page 133
Cause and correlation......Page 141
Latent variables......Page 143
The justification of the proposed model......Page 144
Sample size and number of indicators per latent variable......Page 145
Evaluation of fit......Page 146
Misspecified models......Page 147
Considerations of power......Page 148
Model modification......Page 150
Existing guidelines......Page 151
References......Page 152
Section 2 Applications......Page 157
Introduction......Page 159
The study at hand......Page 160
Factor analytical approaches......Page 161
Describing intraindividual variability......Page 162
Individual differences and nomothetic generalizations......Page 163
The dynamic nature of multi-occasion data......Page 164
Relationship between standard P-technique and dynamic P-technique......Page 166
Conditions of applicability......Page 169
Advantages of P-technique factor analysis......Page 171
Summary......Page 172
Data collection......Page 173
Empirical tests of dynamic relations in the analyzed covariance matrices......Page 175
Standard P-technique factor patterns......Page 176
Interfactor correlations and factor similarity......Page 179
Coherent patterns of change......Page 180
Construct validity......Page 181
References......Page 183
Introduction......Page 187
Historical perspective......Page 188
Basic terms and concepts......Page 189
Methods......Page 190
Results......Page 191
Regression and principal components......Page 193
Structural equation modeling......Page 195
Consideration of the SEM results......Page 198
Regression does not attempt to explain correlations among predictors......Page 201
PCA provides an empirical characterization of the correlations among predictors......Page 202
SEM has a number of additional strengths......Page 203
SEM is not without limitations......Page 204
The estimation of latent variables using SEM deserves further consideration by ecologists......Page 205
References......Page 206
Introduction......Page 210
How do equivalent models arise?......Page 213
Using equivalent models in the design stage......Page 219
An alternative notation for equivalent models......Page 220
Seed dispersal in St Lucie’s Cherry......Page 222
An interspecific model of gas exchange in leaves......Page 223
Conclusions......Page 225
References......Page 226
Introduction......Page 228
System dynamics......Page 229
Definitions of equivalence......Page 230
Implied covariance......Page 231
Model selection......Page 232
Method......Page 237
Results......Page 238
Discussion......Page 247
References......Page 249
Introduction......Page 251
The LISREL model......Page 252
Estimating an ANOVA model as SEM through the use of dummy-coding......Page 254
ANOVA as SEM......Page 255
An example......Page 257
Model 1......Page 258
Model 2......Page 259
A different parameterization for the ANOVA model......Page 262
A one-way repeated measures ANOVA......Page 265
A two-way repeated measures ANOVA......Page 269
An example......Page 271
Estimating a random coefficients model as SEM......Page 280
Using SEM to estimate the linear curve model......Page 282
An example......Page 285
Appendix 10.1......Page 293
References......Page 295
Introduction......Page 297
Example data......Page 299
An overview of multigroup analysis......Page 301
Illustration of a multigroup analysis......Page 303
Applications of multigroup analysis......Page 306
Appendix 11.1. Data and program commands used in analyses......Page 308
References......Page 311
Introduction......Page 313
A LISREL formulation of the means model......Page 315
A simulated study of phenotypic selection......Page 317
The measurement model and tests of assumptions......Page 318
The means model......Page 319
Discussion......Page 323
EQS simulation for pre-selection group data......Page 325
EQS simulation for post-selection group data......Page 326
References......Page 327
Introduction......Page 328
Development of growth curve methodology......Page 331
Polynomial extraction......Page 335
Approximation of time series using orthogonal polynomials......Page 336
Data example......Page 338
Using polynomial parameters in structural equations modeling......Page 341
Data example using manifest growth curve modeling......Page 343
“Tuckerized” growth curves......Page 353
Discussion......Page 360
References......Page 363
Section 3 Computing......Page 369
Introduction......Page 371
Structural equation modeling......Page 372
A comparison of Amos, EQS, and LISREL......Page 374
System requirements (Table 14.1)......Page 375
Documentation for Amos......Page 377
Documentation for EQS......Page 378
Documentation for LISREL......Page 379
Data management in Amos......Page 381
Data input in Amos......Page 382
Modeling using Amos......Page 383
Modeling using EQS......Page 384
Modeling using LISREL......Page 385
Ease of use......Page 387
A data example: Iris or the struggle for admissibility......Page 388
A comparison of features of Amos, EQS, and LISREL......Page 390
Programming features......Page 391
Amos 3.6......Page 395
EQS 5.7......Page 396
LISREL 8.20......Page 397
Admissibility......Page 398
Figures......Page 400
EQS 5.7......Page 402
LISREL 8.20......Page 403
Appendix 14.1. EQS program file for Fisher’s Iris data......Page 404
Appendix 14.2. LISREL command code for Fisher’s Iris data......Page 405
References......Page 406
Index......Page 408