Conditional likelihood, 107–111,
115–116
Conditional logistic regression, 398, 403,
575–579
SAS and, 612–614
SPSS and, 642–643
Stata and, 656–657
Condition indices (CNI’s), 271
Confidence intervals, 121
estimation of, 120, 140–141
interactions and, 142–146
large-sample formula, 121, 447, 473
matching, 393
narrower, 214
odds ratios and, 141, 472–475
one coefficient, 140–142
Confounding
assessment, 171–172, 204–230
causal diagrams for, 177
change-of-estimate rule, 251, 255
general procedure, 216
interactions and, 215–223
modeling strategy for, 204–220
odds ratio and, 87–91, 221
potential, 56, 57, 60, 62, 399
precision and, 211
screening, 264
several exposure variables, 250–251, 254
Consistency, 516, 565
Controls, 392
Control variables, 55
Cook’s distance, 276
Cook’s distance-type measures, 274
Correlation.Seealsospecific procedures,
parameters
autoregressive, 498, 512–513, 545
block diagonal matrix and, 509
covariance and, 500–502
defined, 500
effects of ignoring, 497
exchangeable, 511
fixed, 515
independent, 511
matrix forms, 507–509
stationary m-dependent, 514
structures, 507–510
types of, 511–516
unstructured, 510, 514–515
Covariance
correlation and, 500–502
defined, 500
matrix forms, 117–118, 132–133, 671
Covariate pattern, 307–310
number of covariate patterns (G),
307–308, 314–317
Cox proportional hazards regression,
169
Cross-sectional studies, 11–15
Cut-point (for classification), 348–352D
Data layout for correlated analysis, 499
Data sets, computer.SeeComputer data
sets
Deciles of risk, 318–319
Degrees of freedom, 134, 449
Delta-beta, 275–276
Dependence, and coding, 182–183
Descriptive analyses, 181
Deviance, 312–317
events trials deviance (DevET), 314–315
subject specific deviance (DevSS),
315–317
Diagnostic table, 348
Diagonal matrix, 508
Dichotomous data, 5, 440, 502–503
coding schemes for, 77–82
GEE and, 646–648, 659–660
Discriminant function analysis, 107
Discrimination measures
c statistic, 359, 367–368
gamma, 367–368
Somer’s D, 367–368
Tau-a, 367–368
Discriminatory performance (DP),
304–305, 356
Dispersion, 523
Dummy variables, 82, 398, 399
interaction and, 405
matching and, 109E
Effect modifiers, 57, 405
Eligible subsets, 214
Elimination, candidates for, 216
Epidemiologic study framework, 8696 Index