Frederick Mosteller and David L. Wallace
With a
New Introduction by John Nerbonne
The 1964 publication of Inference and Disputed Authorship made the cover of Time magazine and drew the attention of academics and the public alike for its use of statistical methodology to solve one of American history's most notorious questions: the disputed authorship of The Federalist papers.
Back in print for a new generation of readers, this classic volume applies mathematics, including the once-controversial Bayesian analysis, into the heart of a literary and historical problem by studying frequently-used words in the texts. The reissue of this landmark book will be welcomed by anyone interested in the juncture of history, political science, and authorship.
Frederick Mosteller (1916-2006) was
professor of statistics at Harvard University. David L. Wallace is professor emeritus of statistics at the University of Chicago.
- Chapter 1 The Federalist Papers As a Case Study
- 1.1 Purpose
- 1.2 The Federalist Papers
- 1.3 Early work
- 1.4 Recent work—pilot study
- 1.5 Plots and honesty
- 1.6 The plan of the book
- Chapter 2 Words and Their Distributions
- 2.1 Why words?
- 2.2 Variation with time
- 2.3 How frequency of use varies
- 2.4 Correlations between rates for different words
- 2.5 Pools of words
- 2.6 Word counts and their accuracies
- 2.7 Concluding remarks
- Chapter 3 The Main Study
- 3.1 Introduction to Bayes' theorem and its applications
- 3.2 Handling unknown parameters of data distributions
- 3.3 Selection of words
- 3.4 Log odds
- 3.5 Log odds by words and word groups
- 3.6 Late Hamilton papers
- 3.7 Adjustments to the log odds
- Chapter 4 Theoretical Basis of the Main Study
- 4.1 The negative binomial distribution
- 4.2 Analysis of the papers of known authorship
- 4.3 Abstract structure of the main study
- 4.4 Odds factors for the negative binomial model
- 4.5 Choosing the prior distributions
- 4.6 Magnitudes of adjustments required by the modal approximation to the odds factor
- 4.7 Correlations
- 4.8 Studies of regression effects
- 4.9 A logarithmic penalty study
- 4.10 Techniques in the final choice of words
- Chapter 5 Weight-Rate Analysis
- 5.1 The study, its strengths and weaknesses
- 5.2 Materials and techniques
- 5.3 Results for the screening and calibrating sets
- 5.4 Regression effects
- 5.5 Results for the disputed papers
- Chapter 6 A Robust Hand-Calculated Bayesian Analysis
- 6.1 Why a robust study?
- 6.2 Papers and words
- 6.3 Log odds for high-frequency words
- 6.4 Low-frequency words
- 6.5 The procedure for low-frequency words
- 6.6 Bayesian discussion for low-frequency words
- 6.7 Log odds for 2000-word set and validating set
- 6.8 Disputed papers
- Chapter 7 Three-Category Analysis
- 7.1 The general plan
- 7.2 Details of method
- 7.3 Groups of words
- 7.4 Results for the screening and calibrating sets
- 7.5 Regression effects
- 7.6 Results for the joint and disputed papers
- Chapter 8 Other Studies
- 8.1 How word rates vary from one text to another
- 8.2 Making simplified studies of authorship
- 8.3 The Caesar letters
- 8.4 Further analysis of Paper No. 20
- 8.5 How words are used
- 8.6 Scattered investigations
- 8.7 Distributions of word-length
- Chapter 9 Summary of Results and Conclusions
- 9.1 Results on the authorship of the disputed Federalist Papers
- 9.2 Authorship problems
- 9.3 Discrimination problems
- 9.4 Remarks on Bayesian studies
- 9.5 Summing up
- Chapter 10 The State of Statistical Authorship Studies in 1984
- 10.1 Scope
- 10.2 Computers, concordances, texts, and monographs
- 10.3 General empirical work
- 10.4 Poetry versus prose
- 10.5 Authorship studies similar to the Junius or Federalist studies
- 10.6 Homogeneity problems
- 10.7 Anonymous translation
- 10.8 Forensic disputes
- 10.9 Concluding remarks
- Appendix
- Reference
- Index
12/15/2007
