Outlier tests

for the normal distribution

It seems that JavaScript in your browser is not working.
You must enable JavaScript to be able to use this page.

Paste or write your data set below:

Number separators: Use a space or any other non-numeric character, except a minus sign, period or comma.

Decimal marks: Periods (.) only. Don't use commas as decimal marks; commas with digits adjacent on both sides are interpreted as thousands separators.

Numbers in scientific notation can't be used.

Dixon's test for a single outlier	Any outlier (two sided test) High outlier only (one sided test) Low outlier only (one sided test)
Grubbs' test for a single outlier
Rosner's Extreme Studentized Deviate test for multiple outliers (two sided test):
Significance level (P probability):	(0.00001 - 0.5)
Maximal number of outliers to detect:	(1 - 10)
Iglewicz and Hoaglin's robust test for multiple outliers (two sided test):
Outlier criterion: Modified Z score ≥	(≥1, recommended 3.5)

Test using the log-normal distribution

Your outlier test results:

Please report any errors by emailing us.