Deleting Data

Home > Statistics Every Writer Should Know > The Stats Board > Discusssion

Deleting Data
Message posted by John Dueman on October 8, 2000 at 12:00 AM (ET)

I was asked for the top 10 employers from a survey, and in the results were two junk companies "20" and "-". I asked if I should delete these before printing them out and received a scalding rebuke "You can't do proper data analysis if you delete data". Is there some authoritative basic statement somewhere about omitting obviously errouneous data points? What should I tell this person?

READERS RESPOND:
(In chronological order. Most recent at the bottom.)

Re: Deleting Data
Message posted by SOF on October 8, 2000 at 12:00 AM (ET)

If you have obviously erroneous data, then your data is comtaminated. The problem with editing data is that you then have a judgement sample. To make statistical statements you need a random sample. Also if two of you data points are obviously erroneous, what can you say about the rest of the data ?

Your $5 contribution helps cover part the $500 annual cost of keeping this site online.