Community Manager Community Manager

Should we exclude outliers, and if not, how can we label them?

Julian Parris suggests that in cases where the outlier is clearly not a result of a collection mistake, it can be useful to label them and run the analyses two ways - with and without outliers - to see if the outliers change the relationships among the data or help you glean information from the data.

(view in My Videos)

Article Labels
Article Tags