Index
A
Additive interaction, 53, 77
Adjusted odds ratio, 26, 27, 82
ALR.SeeAlternating logistic
regressions
Alternating logistic regressions (ALR),
570–575
Area under the ROC curve (AUC),
355–365
c statistic formula, 359
geometrical approach, 362–365
Aspirin-Heart Bypass Study, 551–555,
572–575
Asymptotic properties, 516
Autoregressive correlation structure,
512
AR1, 498, 513, 545
B
Backdoor path, 177
Background odds, 20
Backward elimination approach,
184–185
criticized, 264
an option for screening variables, 263
Baseline odds, 20
Berkson’s bias, 178
“Best” model, guidelines for, 168
Biased estimates, 117, 123
in causal diagrams, 175–179
Binomial-based model, 112
Black/White Cancer Survival Study,
434, 470
Block diagonal matrix, 509
Bonferroni correction, 281–282
C
C measures, 274
Case-control study, 11–15, 392, 400–403
intercept in, 116
pair-matched, 109–111, 395, 400–401
Category matching, 393
Causal diagrams, 175–179
for common effect, 177–178
for confounding, 177
example with several exposures, 263
Change-of-estimate rule, 251, 255
Chi-square tests, 134, 138
Mantel–Haenszel, 134, 396
Chunk test, 207–208
several exposure variables, 248
Classification table, 348, 350–352
CLR.SeeConditional logistic regression
Coding
for dichotomous variables, 77–82
tests dependent on, 183
tests independent of, 183–184
Collinearity, 270–275
condition indices, 271–273
EVW model, 173–174
diagnosing, 271
qualitative, 180
variance decomposition proportions
(VDP’s), 271–273
when screening variables, 263, 267
Common effect from two or more
causes, 177–178
conditioning on, 178
Complete separation of points, 358
Computer data sets, 599–602
Conditional estimate, 111, 403