Permutation tests are non-parametric statistical tests that assess the significance of an observed effect by comparing it to the distribution of effects generated by randomly shuffling the data.

food_consumption.csv contains data on several countries' food consumption per food category and their respective CO2 emissions. We are interested in average consumption per food category.

In this case study, we observed that the average consumption of poultry (21.22) across all countries (130 in this dataset) seems to be higher than the average consumption of fish (17.29).

But is this observed result statistically significant? A permutation test will help us assess it.

H0 = avg poultry consumption = avg fish consumption H1 = avg poultry consumption > avg fish consumption

• Through the permutation test, we are able to reject the null hypothesis that the average poultry consumption is equal to the average fish consumption.

• This means that the our initially observed result (average poultry consumption is higher than fish consumption) is statistically significant.