Outliers

[Patricia Ali]

Dear All,

I have a quick question,

Is it necessary to remove outliers from survey data? Would that affect the data I am observing as it would not be a reflection of the actual survey. Isn't taking into account the population weight sufficient, especially in the case where my data passes the linktest, swilk e text, vif test, qnorm etc.

Thanks

[Marcos Almeida]

Quite generally, the outliers are not to be removed, but inspected. If they were not mistyped, if we are convinced they are real values, they become part and parcel of the analysis, let alone the data.

[Patricia Ali]

Thanks Marcos!