We now live in an information-oriented society. Massive amounts of data are routinely collected across many fields. Statistics is the science that transforms these data into information that is critical for decision making. Statistics plays a major role in marketing, public policy, social and health sciences, genetics, brain imaging, speech recognition, computational vision, and more. Statistics is a vibrant field that is embarking on an even more exciting journey as we move ahead to the data-rich, information era of the 21^st century.

Figure Skating Judging at the 2002 Winter Olympics

In the Spring 2002 issue of Chance Magazine, Scott M. Berry, Statistical Scientist, revisits the use of ranks in scoring figure skaters. He explores how this system affected the ordering of American skaters Michelle Kwan and Sarah Hughes in the women's singles skating competition at the 2002 Winter Olympics in Salt Lake City, Utah.  "Perhaps the popularity of figure skating derives from its blend of athleticism and artistry," said Berry. "However, there were two figure skating controversies at the 2002 Winter Olympics that did shock the public." Berry also examines empirical evidence for judging errors in the pairs skating competition that left two couples, Berezhnaya-Sikharulidze of Russia and Sale-Pelletier of Canada, with gold medals.

Sampling Mailrooms for Anthrax

Finding a needle in a haystack is difficult. Proving that the needle is NOT IN the haystack is even harder. Demonstrating with confidence that something is not present became a critical concern in the Fall 2001, when mail-associated anthrax attacks struck fear in the US and panic in the capitol. The mailroom serving Senator Daschle was part of a network of approximately 3200 other mailrooms. Contamination of other mailrooms by spores escaping from this or other letters became a grave concern. However, testing all 3200 or so mailrooms for anthrax would be expensive, labor-intensive, and too slow to prevent further cases of this potentially lethal disease. The article investigates whether a sample could have determined the likelihood of further contamination, the the conclusion that approximately 3040 of the 3200, or 95% of them, would have to be sampled to give reasonable assurances (i.e. only a 5% chance of incorrectly declaring all sites safe) that there was no further contamination. Thus sampling, in this instance, would not have saved much time or resources. When investigators aim to find and eliminate all reservoirs of anthrax (or other lethal hazards), sample sizes are large and exactly proportional to the level of certainty the investigators wish to be. If investigators want to be 100% certain there is no needle in a particular haystack, they have to check the entire haystack.

Imperfections in USGA Golf Handicap System

Lawrence L. Kupper has determined through careful statistical analysis that the handicapping system used by the United States Golf Association (USGA) is imperfect. (Chance Magazine, Winter 2001). The USGA handicapping system is designed to level the playing field for amateurs, allowing golfers of varying ability and consistency to play together with nearly equal scores. "Our work has demonstrated empirically [using real data] that golfers who shoot higher scores tend to be more inconsistent than golfers who shoot lower scores. Since the USGA only uses the lowest 10 of the last 20 adjusted scores (called handicap differentials) to determine a golfer's handicap, handicaps for poorer golfers are less representative of (i.e., unreasonably optimistic about) their true average golfing ability. Hence, poorer golfers do not get enough handicap strokes under the current USGA system when competing against better golfers."

SAT and ACT Test Preparation

Derek Briggs analyzes the effect of SAT and ACT preparation courses on students' test results (Chance Magazine, Winter 2001). Briggs performed an in-depth statistical analysis to determine that test preparation has a limited impact on final test scores. The Scholastic Aptitude Test (SAT) is scored on two sections, math and verbal, with a total possible score of 1600. Briggs found that coaching seems to help students improve their scores on the second attempt at taking the test, but only by 14 to 15 points on the math section and 6 to 8 points on the verbal section. The ACT Assessment covers four areas: English, mathematics, reading, and science reasoning, and has a total possible score of 36. ACT coaching was found increase scores on the math section 0 to 0.4 points and the English section 0.3 to 0.6 points. Coaching was shown to have a negative effect on the reading section by 0.6 to 0.7 points. Briggs' findings are clearly different from statements made by commercial coaching companies, who claim that coaching increases test scores by 100 points or more.  "The problem with such claims," according to Briggs, "is that they are made without reference to a control group. The effect of coaching can only be estimated by a careful comparison of the test score gains for coached students relative to uncoached students. The burden of proof should be on coaching companies to validate their claims more scientifically."

Predicting the Unpredictable Election

David Alan Grier, Associate Professor of Computer Science and International Affairs at The George Washington University, describes how errors in polls, especially election polls, have been controversial, but also how they have helped strengthen the tools of social science and increased public understanding of the nature of statistical studies. In "Predicting the Unpredictable Election," (Chance Magazine, Spring 2002) David Alan Grier reviews polling disasters. He also reviews the work of George Gallup and others that have brought us to a time when it appears we trust election projections more than we trust the ballot totals.