What do you do about an outlier or an influential observation?

-In general, you do not want to remove values from your study unless they greatly distort your results or you find that there were errors in their collection or reporting.

-Before removing observations, one should investigate their origins – if possible.

-For influential observations, they often exist because they are the only explanatory variable in that range of values. Is it possible to sample additional observations in that range without distorting the results of the study?