Information matrix (I^1 ), 271
Interactions
additive, 51
assessment of, 170–171, 190, 207–210
coefficients of, 398
confidence interval estimation with,
142–146
confounding and, 215–223
dummy variables and, 405
likelihood ratio test, 150
matching, 404–406
modeling strategy for, 204–230
multiplicative, 49–55
no interaction, 51–52, 63, 211–215
odds ratio for several exposure variables
with, 87–91
precision and, 211–215
product terms, 405
screening, 265
several exposure variables, 246
variables for, 56
Wald tests, 448
Intercept term, 82, 116
Interval estimation.SeeConfidence
interval estimation
Interval variables, 79
Invariance, 467–468
Iterative methods, 113
J
Joint probability, 112, 114, 451
L
L.SeeLikelihood function
Large-sample formula, 121, 447, 473
Largeversussmall number of
parameters debate, 108–110
Least squares (LS) estimation, 106
Likelihood function (L), 111–117
for conditional method, 114, 115
for ordinal model, 478–479
for polytomous model, 450–452
for saturated model, 311
for unconditional method, 114
Likelihood ratio (LR) statistic, 134–138,
449, 519
carrying out, 148
defined, 120
interaction terms, 150
Likelihood statistic, log, 134
Linear regression, 169
Link function, 505, 506
Logistic function, shape of, 6–7
Logistic model, 5–8.SeealsoLogistic
regression
application of, 9–11
defined, 8
follow-up study and, 14–15
interaction, 49–55, 84
matching and, 397–400
multiplicative interaction, 49–55
simple analysis, 43–46
special cases of, 42–66
Logistic regression.SeealsoLogistic
model
ALR, 570–575
basic features of, 4–7
computing odds ratio in, 74–91
conditional, 398, 403, 575–579, 612–614,
642–643, 656–657
defined, 5
introduction to, 2–32
matching and, 392–406
multiple standard, 479–481
ordinal, 466, 481, 620–621, 644–646,
658
polytomous, 434–457, 617–619, 643, 657
statistical inferences for, 117–121,
130–153, 441–444
stratified analysis, 398
unconditional, 602–612, 635–640,
649–654
Logit form, of EVW model, 76
Logit transformation, 16–22
logistic model and, 17
log odds and, 19–21
Log likelihood statistic, 134
Log odds, logit and, 19–21
LR statistic.SeeLikelihood ratio
LS.SeeLeast squares estimationM
Main effect variables, 27, 53
Mantel–Haenszel odds ratio (MOR),
23, 396
Marginal model, 19, 20
Matching, 116
application of, 400–403698 Index