NAME: Weights of 1996 US Olympic Rowing Team TYPE: Census SIZE: 26 observations, 3 variables (name, event, weight) DESCRIPTIVE ABSTRACT: The data set provides the weights (in lbs) of the 26 men on the 1996 US Olympic Rowing Team in Atlanta. The data includes the names of the participants and which event they rowed in. The US team participated in 7 of the 8 possible events. This data set is useful for discussing outliers, explanations for outliers, and comparing the robustness of the mean and the median. SOURCES: Data were taken from team member biographies given on the NBC Olympic Web Site (http://www.olympics.nbc.com/sports/rowing/events.html). VARIABLE DESCRIPTIONS: 1. Name: The rowers last name 2. Event: The actual event the team member participated in. LW = "lightweight". Rowers in this event actually have a limit on how much they can weigh the day of the race. Coxswain = In an 8 person crew, one team member gives directions but does not row. 3. Weight: Weight of individual team member in lbs. STORY BEHIND THE DATA: A similar set of data is available from DASL (see reference below) for Oxford and Cambridge crews. We felt observations from the recent Olympics would be more relevant to the students. We also liked that there were 2 interesting features to the dataset: the coxswain and the lightweight rowers. PEDAGOGICAL NOTES: Students can be given the data values (or even collect the data themselves from the web site), but need to have the events available as well (e.g. on a transparency). The data set is small enough to be used as an in class example for constructing histograms or stemplots. The coxswain clearly stands out as a low outlier. Usually at least one student will be able to suggest an explanation for this outlier and the role of the coxswain. Another group of observations also stand out as lower than the rest. When the students return to the source of the data, they see these all correspond to rowers in the lightweight event, so again they can generate an explanation and see why it is always important to consider the context of the data. Discussion can include which values should be included when looking for the "average" weight of the Olympic Team, and that they can't always just throw an outlier away. Students can also compare values for the mean and median Mean Median Full Data Set: 191.85 202.5 Without Coxswain: 194.68 205.00 Without Coxswain or 206.32 207.00 Lightweight rowers allowing the students to see the effect of the low values on the mean and median, leading to discussion of the robustness of the median. REFERENCES: "Crews" Dataset on DASL (http://lib.stat.cmu.edu/DASL/Datafiles/Crews.html) SUBMITTED BY: Beth Chance University of the Pacific Math Department, UOP, Stockton, CA 91711 bchance@legendre.cop.uop.edu