Home » Additional Features, CHANCE Highlights

Fall Issue Focuses on Elections

1 October 2010 1,629 views No Comment
Mike Larsen, CHANCE Executive Editor


CHANCE 23.3 Cover

Volume 23, Issue 3 of CHANCE focuses on elections. There are articles about elections graphics, statistical election auditing, and epidemiology associated with election days. Lawrence Mosley, Dianne Cook, Heike Hofmann, Chris Kielion, and Barret Schloerke present graphical displays they used to study and predict the 2008 presidential election. An abstract appears in the magazine, while the full article with color graphics appears online. David Rothschild graphically contrasts the stability and accuracy of debiased daily polls, debiased aggregated polls, and prediction market price estimates over time approaching an election. Andy Gelman and Yu-Sung Su use graphical methods to examine vote by education, age, and ethnicity.

In April 2010, the ASA Board of Directors endorsed risk-limiting audits as a preferred methodology for ensuring the accuracy of elections (see ASA Releases Statement on Risk-Limiting Audits or download the PDF). In his article, Phil Stark discusses risk-limiting vote-tabulation audits and the importance of cluster size. His illustrative heuristics are fun to consider and could be useful for teaching purposes. Stark introduced risk-limiting audits in 2007 and conducted six field-pilots of risk-limiting audits in 2008 and 2009.

Jon Hobbs, Luke Fostvedt, Adam Pintar, David Rockoff, Eunice Kim, and Randy Griffiths compare election auditing options in Iowa. They show that investigation of the statistical properties of audit strategies could help improve election laws in Iowa and other states.

Linda Young and Dan McCrea comment on the election audit law in Florida. Although Florida having such a law is positive, the law’s statistical flaws mean it likely would not be able to prevent election controversies like those that occurred in the past.

Don Redelmeier and Rob Tibshirani analyze an increase in road crash data that occurs on Election Day. They use the data to compare absolute risk, relative risk, odds-ratios, and other risk metrics. Is this a reason to vote by mail?

In an article about a topic that will have an impact on elections (and much more) in the future, Mary Mulry and Pat Cantwell give an overview of the 2010 Census Coverage Measurement Program and its evaluations. The U.S. Census is an enormous (and enormously expensive) enterprise; measuring its accuracy is a tremendous and important statistical challenge.

Three articles concern games and sports. David McCarthy, David Groggel, and John Bailer construct a case-control data set to study the probability of throwing a no-hitter in baseball. Matching is common in analysis of observational studies in health, but rather novel in sports.

Vincent Berthet asks whether the poker player with the best hand tends to win, in which case luck dominates, or skilled players do better than their cards would predict. He derives his unique data set from televised poker games.

Martin Jones and Ryan Parker compare three star basketball players from the NBA on several dimensions. Along the way, they illustrate the challenge of carefully interpreting regression coefficients.

In the Here’s to Your Health column, Mark Glickman brings us an article by Arlene Swern about hormone replacement therapy and coronary heart disease in post-menopausal women. The article explains what is meant by “meta-analysis” and how to think about the statistical quality of the conclusions from such a study.

In the Visual Revelations column, Jim Ramsay and Howard Wainer give us inside-out plots. This method is useful for multivariate tabular data. The authors apply their method to data on baseball players.

Jonathan Berkowitz completes the issue with his cryptic crossword puzzle, Goodness of Wit Test #9. As with other articles in this issue, this puzzle asks you to look at data in a fresh way.

In other news, CHANCE magazine cosponsored 10 sessions at the 2010 Joint Statistical Meetings. View the session summaries by visiting the online program, selecting CHANCE as sponsor, and clicking View. The chosen sessions reflect diverse and important interests and applications in the real world. By endorsing these sessions, CHANCE supports those involved in the sessions, raises awareness of the magazine, and encourages submissions.

1 Star2 Stars3 Stars4 Stars5 Stars (1 votes, average: 5.00 out of 5)
Loading...

Comments are closed.