Sam Behseta, CHANCE Executive Editor
The lead article in Volume 24, Issue 3 is a remarkable account by statistician Daniel Guzmán, who, in 2010, testified as an expert witness in Guatemala during a case involving two former police agents accused of violently disappearing a Guatemalan union activist in 1984. Guzmán and his coworkers used multi-stage sampling methods to select relevant documents—worthy of court presentation—from millions of scattered pieces in the Guatemalan National Police archives. The sampled materials also were used to verify the authenticity of a separate set of documents directly related to the case. The court’s ruling, which resulted in the conviction of the involved police forces, is not only a triumph for human rights, but also a reflection of the crucial role statisticians can play in serving justice globally.
Also in this issue, we cover three articles about sports. First, Michael Rutter gives a simple Bayesian model for ranking NCAA women’s hockey teams. The forte of Rutter’s proposal is that two seemingly complex parameters, namely the ties and home advantage, can be delicately taken into account in the ranking algorithm. Second, David McCarthy uses another Bayesian model to estimate the uncertainty associated with professional tournament scores in bowling, bowler abilities, and tournament difficulties. Finally, Johan Bring and Marcus Thuresson lay out an argument in defense of the now obsolete two points for a win system, as opposed to the current three points in the European soccer leagues. Using simulation studies and data obtained from the Spanish league La Liga, the authors show the old system would have resulted in fairly relegating teams to the second division and deservedly qualifying better regional teams for the much-celebrated European cup competition.
Jimin Ding and colleagues showcase an application of functional data analysis in characterizing the dominating features of uncertainty of data obtained from actigraphy, an emerging technology for measuring sleeping patterns and circadian activity rhythms.
Rebecca Trempel, Sergey Kyrychenko, and Matthew Moore apply a Poisson regression model to gauge the effect of banning hand-held cellular phones while driving on the insurance claims of car collisions in California, Connecticut, New York, and the District of Columbia. The primary outcome of the study finding—that no significant evidence in the decrease in crash risk after cell-phone laws went into effect—is somewhat counterintuitive. The authors provide a partial justification.
In this installment of Visual Revelations, Howard Wainer articulates support for teacher tenure. This is in sharp contrast to the cliché that granting teachers tenure saves money for the participating states. With the aid of a simple graphical tool, Wainer demonstrates the swelling ratio of superintendent-to-teacher salary in New Jersey after the 1991 decision in that state resulted in tenure for superintendents being abolished.
In O Privacy, Where Art Thou?, John Abowd and Lars Vilhuber continue the discourse started in the previous column by Stephen Fienberg about the many facets of privacy, confidentiality, disclosure, and harm. In addition to stressing the wide-ranging benefits of sharing data with public agencies, the authors accentuate a need for the participation of a larger community of researchers as novel methods of data sharing are streamlined.
I am pleased to announce two new members of the editorial board: Shane Jensen from the University of Pennsylvania will be writing A Statistician Reads the Sports Pages and Christian Robert from Université Paris-Dauphine will write book reviews.
Leave a Reply