In: Statistics and Probability
Here are some real statistics for various countries in 2003: per capita income vs. per capita recorded music sales:
Country Per Cap. Income Per Cap. Music
($ thousands) Sales ($)
Norway 42.4 55.9
United Kingdom 30.9 53.35
United States 42.0 40.43
Australia 32.9 33.84
Switzerland 35.3 34.40
Finland 30.6 26.98
Canada 32.9 20.79
United Arab Emirates 29.1 11.33
Greece 22.8 8.10
Israel 22.3 6.68
Czech Republic 18.1 3.96
South Africa 12.1 3.75
South Korea 20.4 3.34
Mexico 10.1 3.30
Egypt 4.4 0.15
Indonesia 3.7 0.33
4a) Which variable should be considered the dependent (y) variable, and which the independent (x) variable?
4b) According to this model, if a country’s income improves by $3,000, how much to you expect music sales to increase by?
4c) What is the correlation coefficient? Is it statistically significant? How strong is this model?
4d) Why do you think the correlation is as high as it is?
4e) If a region of country has a per-capita income of $25,000, predict its per-capita music sales.
4f) Looking at the U.S, Canada. & Europe only, delete the United Arab Emirates, South Africa, South Korea Mexico, Egypt, and Indonesia from the model. Answer 3b, 3c, and 3e again.
4g) Why might it have been OK to throw away the countries we did in part f?