# Descriptive Statistics: Should I exclude an outlier?

In my study that I am writing up for university:

There were a total of 204 participants, 96 female (47.1%) and 108 male (52.9%). The participants were aged between 20 and 125.

The problem I have is 125 is older than the oldest verified person. Do I include this in my descriptive statistics, and in the general data analyses, or do I exclude from descriptive and data analyses?

Tips would be most welcome!

• It sounds as if that you know that the value is wrong. You surely would have noticed a person of that age in your survey. Perhaps it's facetious; perhaps it's a slip for 25. If I were you, I'd replace it with missing. – Nick Cox Jan 27 '17 at 18:39
• The survey was conducted over the web, so there was no face to face interaction. How would you justify it with 'missing'? – Jamie Jan 27 '17 at 18:55
• The value is wrong and you don't know what value is right; so you enter missing. Let's play a child-like game: my height in 4 metres. Do you believe my height? What height would you enter for me if you trusted the rest of the data on me? I'll readily buy an argument that the entire observation should be struck if someone appears to be messing with the survey. You'll find that's what happens by default with many programs unless you apply multiple imputation. – Nick Cox Jan 27 '17 at 19:00