How to compute the robust measure of scale Qn?

[aminkarimid]

Hello everybody,

I use the PROC UNIVARIATE statement to find outleirs.

Here are my codes:

proc univariate data=sampledata05 robustscale plot;
var  price;
run;

There is a warning?

WARNING: The number of nonmissing observations for variable Price is too large to compute the robust
         measure of scale Qn. The statistic Qn is set to missing

How can I fix that?

Thanks in advance

[ballardw · Accepted solution]

How many non-missing values do you have for Price? The message says that it can't compute the requested value because you have too many points. If you look into details of Qn there is a factorial component (part of an N choose R ) to the calculation and hence for large N cannot be computed in practical terms.

[aminkarimid]

My dataset has no missing values.
Thanks.