Common Issues in Interpretation II: Non-Normal Distributions

Many statistical models assume that the data is normally distributed like in a bell curve, where a) most observations lie around the mean, b) the frequency of data that falls far from the mean is inversely related with its distance from the mean, and c) the distribution of observations above and below the mean are relatively identical. Often, results from experiments are non-normal, which can lead to spurious findings. Therefore, data transformations often need to be made to account for such problems.

On the TV show, Kitchen Horrors, Bordon Hannity tries to revive failing restaurants by offering advice to their management and chefs. To measure the effectiveness of Hannity's efforts, a fan measures the average reviews of one of Hannity's revived restaurants before and after he offers them advice. With these results:

1) Find the mean difference of the reviews before (Reviews_Before) and after (Reviews_After) Bordon Hannity's advice.
2) Produce kernal density plots of the reviews with the functions plot and density.
3) Determine whether the reviews are normally distributed

Exercise

Common Issues in Interpretation II: Non-Normal Distributions

Instructions