Desktop productivity for business analysts and programmers

Testing for normal distribution, identifying & removing outliers

Reply
New Contributor
Posts: 3

Testing for normal distribution, identifying & removing outliers

Hi, I'm a relatively new SAS EG user. I took the first training course recently and learned a lot--but there are a bunch of things I'm having trouble with, hoping others have thoughts on how to do these. Here are a couple:

1) I would like to test to see if a variable in my dataset is normally distributed. How can I do this in SAS EG?

2) I would like to identify and remove outliers from my data set. How can I do this in SAS EG?

Many thanks for any help you can provide, or resources you can point me to.

Grand Advisor
Posts: 17,325

Re: Testing for normal distribution, identifying & removing outliers

For 1:

Tasks>Describe>Distribution Analysis

Click on Tables> Tests for Normality

You'll get a table like follows (for SASHELP.CARS, MPG_CITY).
There are various other options that you can look into. Testing is a good method to find things.

Re 2:

There isn't a set method to identify/remove outliers so you'll have to be more specific, how are you defining an outlier?

Ask a Question
Discussion stats
  • 1 reply
  • 607 views
  • 0 likes
  • 2 in conversation