Keywords: simple linear regression, multiple regression, transformation, residuals
The data give the number of medals won by each medal-winning country in the 1992 Summary Olympic Games in Barcelona, Spain, and the 1994 Winter Olympic Games in Lillehammer, Norway. Also given is the population and latitude of each country. Griffiths et al write:
... the media spent a lot of time discussing the number of medals won by each country's athletes. The implication was that the comparison was of some importance. However, larger countries would be expected to win more medals than smaller countries, simply because of their larger populations.
... some viewers, especially those from the smaller countries, felt that the number of medals should be standardised to account for the very wide range of populations, and that a per capita number of medals for a country was a fairer comparison. Others felt that this was unfair to the countries with larger populations - that having twice as many people did not lead to twice as many medals. If standardisation is performed adequately, there should be no systematic relationship between the adjusted medal count and population.
Also countries further from the equator might be expected to do better in the winter olympics.
The data is incomplete in that countries with no medals are not included. These would be mostly smaller population countries.
Data file (tab-delimited text)
|Griffiths, D., Stirling, W. D., and Weldon, K. L. (1998). Understanding Data. Principles and Practice of Statistics. Wiley, Brisbane.|
|Gordon Smyth collected the country latitudes.|