At 06:44 PM 3/2/2014, albert_sun wrote:
>I run the percentile procedure in SPSS (through frequency) over the
>data I have and the 10th percentile and 90th percentile from the
>whole data and from a group of the data is exactly the same, and is
>this something possible?
It's very possible; it's perfectly possible for the distribution of
values in a subgroup to be the same as the distribution in the whole.
The question is, is it plausible? To start with, does your variable
take on a lot of values, or only a few? (It's much easier for
distributions to match by chance if the variable takes on only a few values.)
Do other statistics, like mean and median, match? If *all* statistics
match, it could be that either (1) the subgroup as defined turns out
to include the whole group, or (2) the part of the group outside the
subgroup is a copy of what's in the subgroup.
And that's enough wild speculation. What is the quantity you've
observing, for which you see this phenomenon? What defines the
subgroup, and do you expect it to be very different from the
remainder of the group?
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Locked
