Extremes in Explorative Data Analysis

[Simon J. Buechner]

What's the criterion SPSS uses in order to classify a value as an
"extreme" in explorative data analysis?

Regards,
Sim

[Daniel Robertson]

WRT boxplots, a case whose value is 3*IQR above Q3 or 3*IQR below Q1 is
classified as extreme. See "SPSS 15.0 Algorithms", p.252 (quoted below):

"The boundaries of the box are Tukey’s hinges. The length of the box is
the interquartile range based on Tukey’s hinges. That is,

IQR = Q3 - Q1

Define

STEP = 1.5 IQR

A case is an outlier if

Q3 + STEP <= y_i < Q3 + 2STEP
or
Q1 - 2STEP < y_i <= Q1 - STEP

A case is an extreme if

y_i <= Q3 + 2STEP
or
y_i <= Q1 - 2STEP"

hope this helps,
Dan R.


Simon J. Buechner wrote:
> What's the criterion SPSS uses in order to classify a value as an
> "extreme" in explorative data analysis?
>
> Regards,
> Sim
>

--
Daniel Robertson
Sr. Research and Planning Associate
Institutional Research and Planning
Cornell University
440 Day Hall, Ithaca NY 14853-2801
ph:607.255.9642 / irp.cornell.edu