Kline R.B. Principles and Practice of Structural Equation Modeling
Подождите немного. Документ загружается.
424 Subject Index
Structural models
defined, 103
start value suggestions for, 185
types of, 106–110
Structural regression (SR) models
analyzing, 265–268
cause indicators and formative measure-
ment, 280–286
constraint interaction, 295
equivalent versions, 276
estimation
detailed example, 270–275
interpretation of parameter estimates,
269
methods, 269
failing to evaluate the measurement and
structural portions of, 363
fully and partially latent, 119
identification rules for, 144–146
invariance testing of, 288–289
overview, 118–119
research example, 120–121
single indicators in, 276–280
with structured means, 322
Structure coefficient, 232
Structured means
in measurement models
empirical example, 318–322
overview, 316–318
SR models with, 322
Suppression, 26–27
Symbolic processing, 131
SYSTAT 13, 84–85
System matrix, 151–153
Tabulation, 292
Tailored tests, 244–245
Tau-equivalent indicators, 243
TCALIS, 80–81
Test for orthogonality, 243
Test for redundancy, 243
Test–retest reliability, 70
Tetrachoric correlation, 31
Three-indicator rule, 138
Thresholds
for approximate fit indexes, 197–198
defined, 180
model test statistics and, 194
Time-invariant predictors, 314
Time structured data, 304–305
Time-varying predictors, 314
Tolerance, 53
Standardized Root Mean Square Residual
(SRMR), 204, 208–209
Standardized total effects, 167
Standard reflective measurement, 281
Start value problem, 172–175
Start values, 157
common mistakes with, 361
suggestions for measurement models, 263
suggestions for structural models, 185
Stationarity, 108
STATISTICA 9 Advanced program, 85, 224
Statistical beauty, 95, 293
Statistical significance (α)
misinterpretations of, 36–39
threshold value, 194
Statistical tests
misinterpretations of p values, 36–39
role of, 12–13
in SEM, 36, 39–40
standard errors, 33–34
Steiger, J., 224
Steiger–Lind root mean square error of approx-
imation, 204, 205–207
Stem-and-leaf plots, 60, 61, 62
Stepwise regression, 27–28
Stratified sampling, 343
Strictly confirmatory applications, 8
Structural equation modeling (SEM)
a disconfirmatory technique, 16
basic steps, 91–94
characteristics of, 7–13
computer tools, 7 (see also Computer tools)
convergence with multilevel modeling, 348–349
goal of, 189–190
history and development of, 15–16
Internet resources, 3–4
multilevel, 350–353
optional steps, 94–95
popularity of, 13–14
preparing to learn about, 5–7
problems with, 14–15
reporting results of
bottom lines and statistical beauty, 293
confirmation bias, 292–293
data and measure, 291
estimation and respecification, 291–292
identification, 290
overview, 289–290
specification, 290
tabulation, 292
statistical tests in, 36, 39–40
terms, 7–8
Subject Index 425
Validity fallacy, 39
Variables. See also Endogenous variables;
Latent variables; Observed variables
censored, 32
count variables, 16
excluded, 133–135
exogenous, common mistakes with, 357
index variables, 280–281
instrumental, 156
intervening, 105–106
lagged, 316
latent response, 180
manifest, 8–9
mediator, 105–106
outcome variables, 10, 179–180
prior, 316
Variance(s)
configural variance, 252–253, 288
corrected proportions for nonrecursive
models, 187–188
population variance, 155
relative variance, 67–68
sample variances, 155
unexplained variance, 231
unique variance, 113, 201
See also Disturbance variances
Variance inflation factor (VIF), 53–54
Vectors, 54
Visual-PLS, 288
Wald W statistic, 217
Websites, 3–4
Wherry’s equation, 20–21
Wiley, D., 15
“Wizards,” 77
WLS estimation. See Fully weighted least
squares estimation
WLSM. See Mean-adjusted weighted least
squares
WLSMV. See Mean- and variance-adjusted
weighted least squares
Wright, Sewell, 15, 16
Zero-order correlations, 19
z test, 34
Total effect moderation, 335
Total effects
example analysis, 166–169
standardized, 167
unstandardized, 167
Total indirect effects, 167
Tracing rule, 169–171
Transformations, 63–64
Triangle inequality, 50
TSLS estimators, 159
Two-factor model, 235–237
Two-indicator rule, 138
Two-level regression, 345–346
2+ emitted paths rule, 283
Two-stage least squares (TSLS) estimation,
155, 156–157, 341
Two-step modeling, 266, 267, 268
Two-step rule, 144–146
Unconditional intraclass correlation, 344
Unconstrained approach to estimation,
341
Underdetermined models, 125, 126
Underidentification
empirical, 146–147
failure to recognize, 361
Underidentified models, 125, 126
Unexplained variance, 231
Unidimensional measurement, 115
Uniform distributions. See Homoscedasticity
Unique variance, 113, 201
Unit loading identification (ULI) constraints,
127, 128, 130, 269
Unit variance identification, 128, 130
Unit weighting, 245
Univariate normality, 60–63
Univariate Wald W statistic, 217
Unknown weights composite, 282
Unrestricted factor models, 115
Unstandardized path coefficient, 301–302
Unstandardized regression coefficients, 21
Unstandardized residual path coefficient, 105
Unstandardized total effects, 167
Unweighted least squares (ULS) estimation,
176
427
About the Author
Rex B. Kline, PhD, is Professor of Psychology at Concordia University in Montréal,
Quebec, Canada. Since earning a doctorate in clinical psychology, he has conducted
research on the psychometric evaluation of cognitive abilities, child clinical assessment,
structural equation modeling, training of behavioral science researchers, and usability
engineering in computer science. Dr. Kline has published five books, six chapters, and
more than 40 articles in research journals.