imageWhen discussing data quality issues, we frequently address issues where inconsistent data can lead to analysis problems.

The 2012 Olympics present an interesting variant where the input data is the same, but leads to a different answer…

When looking at the different positions of the countries competing in the Olympics, Britain and most of the world use a system whereby the number of Gold medals is the prime decider of rank, followed by Silver and then Bronze. On the 5th August, this leads to the following top 10 countries:

Rank Country Gold Silver Bronze Total
1 China 29 16 14 59
2 United States 27 14 15 56
3 Britain 16 10 10 36
4 South Korea 10 4 6 20
5 France 8 7 9 24
6 Germany 5 10 7 22
7 Italy 5 5 3 13
8 Kazakhstan 5 0 0 5
9 North Korea 4 0 1 5
10 Russia 3 16 15 34

However, for some reason, the US media have used a different system to the rest of the world where rank is based on the total number of medals and then the numbers of Gold, Silver and Bronze. This leads to the following result:

Rank Country Gold Silver Bronze Total
1 China 29 16 14 59
2 United States 27 14 15 56
3 Britain 16 10 10 36
4 Russia 3 16 15 34
5 Japan 2 11 12 25
6 France 8 7 9 24
7 Germany 5 10 7 22
8 South Korea 10 4 6 20
9 Australia 1 12 7 20
10 Italy 5 5 3 13

So what are the effects of the different approach?

The rank of the top 3 countries are unchanged, but the US old Olympic rivals Russia are 4th rather than 10th, however the two Koreas are pushed down the rankings. Australia also makes an appearance in the US version of the top 10.

So, same input data, very different output.

What other similar examples are you aware of?

Tagged on:         

One thought on “The same data… but different answer

  • 5th August 2012 at 8:30 pm
    Permalink

    Both ranking methods leave a bad taste, as at first glance none of them is satisfactory. Okay, let’s assign a weight to each type of medal and rank nations by the weighted sum, but can we agree on the weighting factors? And is it anyway fair to compare large nations with smaller countries? So maybe we should calculate the number of medals per participating athletes of each nation or compare to their number of residents?!
    Way off the mark – the simple, unspoken rule is “The winner takes it all” – and that’s what counts (and what should be counted).

    Reply

Leave a Reply

Your email address will not be published. Required fields are marked *

This site uses Akismet to reduce spam. Learn how your comment data is processed.