13.11.2012 Views

Introduction to Categorical Data Analysis

Introduction to Categorical Data Analysis

Introduction to Categorical Data Analysis

SHOW MORE
SHOW LESS

Create successful ePaper yourself

Turn your PDF publications into a flip-book with our unique Google optimized e-Paper software.

SUBJECT INDEX 355<br />

ordinal versus nominal treatment<br />

of data, 41–45, 118–119<br />

quasi symmetry, 258–260<br />

scores, choice of, 43–44, 119, 195<br />

testing independence, 41–45, 232<br />

trend in proportions, 41–45<br />

Ordinal quasi symmetry, 258–260<br />

Overdispersion, 80–84, 192–193, 280,<br />

283–284, 304–305<br />

Paired comparisons, 264–266<br />

Parameter constraints, 113, 206, 221<br />

Partial association, 49<br />

partial table, 49<br />

same as marginal association, 224<br />

Partitioning chi-squared, 39–40, 62, 329<br />

Pearson chi-squared statistic, 35, 37, 61<br />

chi-squared distribution, 35–36<br />

comparing proportions, 26<br />

degrees of freedom, 35, 37, 62, 327<br />

goodness of fit, 86, 145–147, 184, 212<br />

independence, 35–38, 41–43, 61, 333<br />

loglinear model, 212<br />

and residuals, 86–87, 148<br />

sample size for chi-squared<br />

approximation, 40, 156–157, 329<br />

sample size, influence on statistic, 61<br />

two-by-two tables, 26, 40<br />

Pearson, Karl, 325–327<br />

Pearson residual, 86–87, 148<br />

Binomial GLM, 148<br />

GLM, 86–87<br />

independence, 38–39<br />

Poisson GLM, 87<br />

Penalized quasi likelihood (PQL), 317<br />

Perfect discrimination, 153<br />

Poisson distribution, 74<br />

mean and standard deviation, 74<br />

negative binomial, connection with, 81<br />

overdispersion, 80–84<br />

generalized linear mixed model, 324<br />

Poisson loglinear model, 75, 205, 324<br />

Poisson regression, 75–84<br />

residuals, 87<br />

Poisson GLM, 74–84, 205, 334–335<br />

Population-averaged effect, 249, 279<br />

Power, 160–162<br />

Practical vs. statistical significance, 61, 140,<br />

218–219<br />

Prior distribution, 317<br />

Probability estimates, 6, 68, 100, 108–109,<br />

176, 245<br />

Probit model, 72–73, 135, 328<br />

Proportional odds model, 182 See also<br />

Cumulative logit model<br />

Proportions<br />

Bayesian estimate, 17<br />

confidence intervals, 9–10, 20, 26<br />

dependent, 34, 244–252<br />

difference of, 26–27, 246–247<br />

estimating using models, 68, 100, 108,<br />

176<br />

independent, 26<br />

inference, 6–16, 26–31<br />

ratio of (relative risk), 27–28, 32<br />

as sample mean, 7<br />

significance tests, 8, 13–16, 19<br />

standard error, 8, 19, 26<br />

P -value and Type I error probability,<br />

14, 20, 47–48<br />

Quasi independence, 261–263, 274, 329<br />

Quasi likelihood, 280<br />

Quasi symmetry, 257–259, 263, 265, 274,<br />

340<br />

R (software). See<br />

www.stat.ufl.edu/∼aa/cda/<br />

software.html<br />

Random component (GLM), 66<br />

Random effects, 298–318, 341–342<br />

bivariate, 311–313<br />

predicting, 299, 303, 313, 316<br />

Random intercept, 298<br />

Ranks, 44–45<br />

Rasch model, 307, 328<br />

Rater agreement, 260–264<br />

Rates, 82–84, 97<br />

Receiver operating characteristic (ROC)<br />

curve, 143–144<br />

Regressive-logistic model, 288–289<br />

Relative risk, 27–28, 32, 328<br />

confidence interval, 28, 58<br />

and odds ratio, 32, 33<br />

Repeated response data, 276

Hooray! Your file is uploaded and ready to be published.

Saved successfully!

Ooh no, something went wrong!