Logistic Regression: A Self-learning Text, Third Edition (Statistics in the Health Sciences)

basic features of, 392–394
case-control studies, 392–409
category matching, 393–394
cohort studies, 409–413
confidence intervals, 393
exchangeability and, 407–408
follow-up data, 409–413
interaction and, 404–406
logistic model and, 397–400
major disadvantage, 394
matching factors, 392
ML estimation and, 401
pooling, 407–409
precision and, 393
stratification, 394–397
validity and, 394
Mathematical models, 5
Matrix algebra, 117–118, 132–133,
507, 524
Maximum likelihood (ML) methods,
106–107
numerical example, 147–153
overview, 106
statistical inferences using, 130–153
subject-specific saturated model, 312
unconditionalversusconditional,
107–109
McNemar test, 396–397, 412
Meaningful change in OR,
concept of, 217
Methicillin-resistance infection (MRSA)
example, 244
Method 0, 265–270
MI dataset, 600
Mixed logistic model (MLM),
580, 584
ML.SeeMaximum likelihood methods
MLM.SeeMixed logistic model
Modeling strategy
confounding and, 203–230
example of, 188–192
guidelines for, 165–192
interaction and, 203–230
overview of, 169–173
rationale for, 168–169
several exposure variables, 244–262
Moderate samples, 121
MOR.SeeMantel–Haenszel odds ratio
Multicollinearity, 172, 280

Multilevel outcomes, 432

Multiple linear regression, 169

Multiple standard logistic regressions,

453, 479–481

Multiple testing, 172

Multiplicative interaction, 49–53

Multivariable problem, 4–5

N

No interaction model, 63, 78–79, 83–85,

149–150, 152–153, 211–215

Nominal exposure variable, 82–84

Normal distribution, 141

Nuisance parameters, 116, 526, 575

Null hypotheses, 54, 280

O

Odds, 18–19

Odds ratio (OR), 11–13

adjusted, 26, 27, 77

computation of, 25–26, 64, 74–91

confidence limits, 141, 472–475

confounders and, 87–91, 221

correlation measure and, 571

examples of, 22–23

exchangeable, 573

as fixed, 60, 79

formula for, 22–25, 84

invariance of, 468

logistic regression and, 74–91

MOR and, 396

risk ratio and, 15–16

three categories, 437–441

One-to-one matching, 393

OR.SeeOdds ratio

Ordinal logistic regression, 466–472

SAS and, 620–621

SPSS and, 644–646

Stata and, 658

Ordinal models, 466–472

Ordinal variables, 79

P

Pair matching, 110–111, 393–398,

400–403

Parameterizing, of model, 470

Parameters, number of, 108

Pearson statistic, 322

Perfect fit, 305

Index 699