## Statistical Analysis of Epidemiologic DataAnalytic procedures suitable for the study of human disease are scattered throughout the statistical and epidemiologic literature. Explanations of their properties are frequently presented in mathematical and theoretical language. This well-established text gives readers a clear understanding of the statistical methods that are widely used in epidemiologic research without depending on advanced mathematical or statistical theory. By applying these methods to actual data, Selvin reveals the strengths and weaknesses of each analytic approach. He combines techniques from the fields of statistics, biostatistics, demography and epidemiology to present a comprehensive overview that does not require computational details of the statistical techniques described. For the Third Edition, Selvin took out some old material (e.g. the section on rarely used cross-over designs) and added new material (e.g. sections on frequently used contingency table analysis). Throughout the text he enriched existing discussions with new elements, including the analysis of multi-level categorical data and simple, intuitive arguments that exponential survival times cause the hazard function to be constant. He added a dozen new applied examples to illustrate such topics as the pitfalls of proportional mortality data, the analysis of matched pair categorical data, and the age-adjustment of mortality rates based on statistical models. The most important new feature is a chapter on Poisson regression analysis. This essential statistical tool permits the multivariable analysis of rates, probabilities and counts. |

### What people are saying - Write a review

We haven't found any reviews in the usual places.

### Contents

1 | |

2 | |

Probabilities | 6 |

Incidence and prevalence | 8 |

Survival probabilities and hazard rates | 12 |

Statistical properties of probabilities calculated from mortality or disease data | 14 |

smoothing transforming and adjusting | 22 |

Variation and Bias | 40 |

general considerations | 248 |

Centering | 249 |

The WCGS additive logistic model | 252 |

Casecontrol sampling | 259 |

The Analysis of Count Data Poisson Model | 263 |

Simplest Poisson model | 264 |

technical description | 265 |

Illustration of the Poisson model | 266 |

A simple model | 41 |

The 𝐭test | 43 |

Selection Bias | 48 |

Confounder bias | 50 |

Ecologic bias | 52 |

Comparison of k groups | 54 |

Interaction contrasts | 58 |

Twoway analysis | 62 |

Misclassification bias | 69 |

Statistical Power and Sample Size Calculations | 75 |

sample mean | 79 |

relative risk | 80 |

onesample test of a proportion | 82 |

twosample test of proportions | 84 |

Loss of statistical power and bias from grouping continuous data | 88 |

Cohort Data Description and Illustration | 93 |

model | 94 |

Birth cohort effect and proportional mortality data | 97 |

Median polish analysis | 100 |

Mean polish analysis | 102 |

Four examples | 104 |

prostatic cancer data | 110 |

Spatial Data Analysis and Estimation | 120 |

Poisson model | 121 |

Nearestneighbor analysis | 125 |

Transformed maps | 130 |

Spatial distribution about a point | 135 |

Timedistance spatial analysis | 138 |

Randomization test | 143 |

randomization test | 146 |

Bootstrap estimation and analysis | 149 |

The 2 X 𝘬 table and the 2 X 2 X 2 Table | 159 |

Independence and homogeneity | 160 |

Regression | 165 |

comparison of two means | 169 |

Ridit probability analysis | 173 |

The 2 X 2 X 2 contingency table | 179 |

The Analysis of Contingency Table Data Logistic Model I | 190 |

discrete case | 191 |

The 2 X 2 X 2 table | 199 |

The 2 X 𝘬 table | 208 |

The 2 X 2 X 𝘬 table | 213 |

The multiway table | 221 |

discrete case | 224 |

Summarizing a series of 2 X 2 tables | 227 |

The Analysis of Binary Data Logistic Model II | 236 |

Bivariate logistic regression | 240 |

Hodgkins disease | 268 |

CHD risk by smoking behavior type | 273 |

application of the Poisson model | 276 |

a twoway classification | 278 |

a threeway classification | 283 |

The Analysis of Matched Data Three Approaches | 291 |

Frequency matching | 292 |

Poststratification | 294 |

continuous variable | 296 |

binary risk factor | 300 |

Confidence interval for the odds ratio | 305 |

Evaluating the estimated odds ratio | 307 |

Disregarding matching | 309 |

Interactions with the matching variable | 311 |

Matched sets using more than one control | 313 |

multilevel categorical risk factor | 317 |

Conditional logistic analysis | 320 |

Life Table Analysis An Introduction | 335 |

construction | 336 |

Life table survival function | 350 |

three applications of life table techniques | 356 |

Competing risks | 372 |

Survival Data Estimation of Risk | 378 |

Parametric model | 379 |

Age adjustment of rates | 382 |

Censored and truncated data | 384 |

parametric estimate | 386 |

nonparametric estimate | 390 |

Mean survival time from censored data | 393 |

Goodnessoffit | 396 |

Twosample data | 399 |

The Wilcoxon test and the Gehan generalization to survival data | 407 |

Survival Data Proportional Hazards Model | 412 |

Simplest case | 413 |

The proportional hazards model | 417 |

Plotting survival curves | 420 |

Four applications of a proportional hazards model | 421 |

Dependency of followup time | 439 |

Appendix | 443 |

Binomial and Poisson probability distributions | 445 |

The odds ratio and its properties | 448 |

Partitioning the chisquare statistic | 454 |

Maximum likelihood estimation and likelihood functions | 457 |

Problems | 464 |

479 | |

487 | |

### Common terms and phrases

additive model age-adjusted age-specific analysis assess behavior type birth blood pressure calculated case/control categorical variables censored CHD event CHD risk chi-square distribution chi-square statistic coefficient cohort column comparison confidence interval confounder bias coronary degrees of freedom denoted difference disease effect evaluate example follow-up groups hazard functions hazard rate increase independent influence interaction leukemia levels likelihood linear log-likelihood statistics log-odds log-rates logarithm logistic model logistic regression lung cancer matched pairs mean survival mean values measure of association median model parameters mortality rates normal distribution null hypothesis number of deaths number of individuals number of model odds ratio p-value pattern person-years Poisson distribution Poisson model population prepregnancy weight probability of death produces proportional hazards model random relationship risk factor risk variables saturated model smoking exposure specific strata summary survival curves survival data survival probability t-test Table test-statistic time/distance total number treatment type-A WCGS data zero

### Popular passages

Page 480 - Statistical Methods for Rates and Proportions, John Wiley and Sons, New York, 1973.