Estimated variance-covariance matrix,
117–118, 132–133
Evans County Heart Disease Study, 61,
62, 118, 146, 158, 188, 223–230,
599
Events-trials (ET) format, 307
E, V, W model, 55–64
general, 56–57
logit form of, 76
several exposure variables, 85–92, 245–246
Exchangeable correlation structure, 511
Exchangeable sets, 407
Exposure variables, 4, 46
EVW model, 245–247
odds ratio for, 87–91
single, 45
Extra variation, 523
F
False negatives (FN), 350
False positive rate (FPR), 355
1–specificity (synonym for FPR), 349,
353, 355
False positives, (FP), 350
Family-wise error rate, 280
Fixed correlation structure, 515
Fixed effects, 576, 579
Follow-up studies, 11–15
Forward selection approach
an option for screening variables, 264
criticized, 264
Full model, 135
Fully parameterized model, 308–311
G
GEE.SeeGeneralized estimating
equations
Generalized linear mixed model
(GLMM), 580–587
Generalized linear models (GLM),
503–506, 526
Generalizing estimating equations
(GEE) model, 492–538, 540–565
ALR model and, 569, 574
Aspirin-Heart Bypass Study, 551–555
asymptotic properties of, 516
defined, 506–507
dichotomous data and, 646–648, 658–660
GLM and, 526
Heartburn Relief Study, 555–557
Infant Care Study, 542–550
matrix notation for, 524–525
parameters in, 526
score-like equations and, 524–528
statistical properties of, 516
General odds ratio formula, 84
GLM.SeeGeneralized linear models
GLMM.SeeGeneralized linear mixed
model
Gold standard model, 212–215, 219
Goodness of fit (GOF), 304–325
Group prediction, 307, 310
Group saturated, 307, 310H
Hat symbol, 9
Heartburn Relief Study, 555–557
Hierarchical backward-elimination
approach (HBWE), 184–185
Hierarchically well-formulated (HWF)
model, 181–184, 406
Hierarchy principle, 185
applying, 216
product terms and, 191
rationale for, 186
retained variables, 185–187
Hosmer–Lemeshow statistic, 318–325
examples, 320–325
formula, 320
table of observed and expected cases,
319
HWF.SeeHierarchically well-
formulated model
Hypothesis Generating Model, 260
Hypothesis testing, 9, 117, 120, 132–153I
Identity matrix, 511
Independent correlation structure, 511
Independent of coding tests, 182–183
Independent variables, 4–5
Indicator variables, 398–399
Individual prediction, 310
Infant care study, 493–498, 542–550, 559
Inference, statistical, 117–121, 130–153,
441–444.Seealsospecific
methods
Influential observations, 173, 275–278Index 697