formula for median in SPSS

[Biggert, Robert (Sociology)]

In teaching a stats class, I compared some medians, some calculated by
hand and some in SPSS.  Given the consensual definition of a median as
the  (n+1)/ 2 value in an ordered distribution, I was assuming the two
sets of medians would be identical.  However, some weren't.  The data
set I was using was small (n=15) and for some of the variables there
were duplicate values at or near the median.  I'm assuming SPSS used the
formula with some correction factor but couldn't find the formula in
SPSS or on the internet.  So my question is: How does SPSS calculate the
median?

Thanks, RB

 

====================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

[Art Kendall]

click <help>
click <algorithms>
click <frequencies>
click <p-th percentile>

Art Kendall
Social Research Consultants

Biggert, Robert (Sociology) wrote:

> In teaching a stats class, I compared some medians, some calculated by
> hand and some in SPSS.  Given the consensual definition of a median as
> the  (n+1)/ 2 value in an ordered distribution, I was assuming the two
> sets of medians would be identical.  However, some weren't.  The data
> set I was using was small (n=15) and for some of the variables there
> were duplicate values at or near the median.  I'm assuming SPSS used the
> formula with some correction factor but couldn't find the formula in
> SPSS or on the internet.  So my question is: How does SPSS calculate the
> median?
>
> Thanks, RB
>
>
>
> ===================
> 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
>
>
>

=====================
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

[ViAnn Beadle]

I'm curious as to what the "consensual" way to calculate a median is. IIRC,
there are about 4 ways to do this. You'll find the algorithms at

http://support.spss.com/ProductsExt/SPSS/Documentation/Statistics/algorithms
/index.html


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Biggert, Robert (Sociology)
Sent: Thursday, February 07, 2008 1:11 PM
To: [hidden email]
Subject: formula for median in SPSS

In teaching a stats class, I compared some medians, some calculated by
hand and some in SPSS.  Given the consensual definition of a median as
the  (n+1)/ 2 value in an ordered distribution, I was assuming the two
sets of medians would be identical.  However, some weren't.  The data
set I was using was small (n=15) and for some of the variables there
were duplicate values at or near the median.  I'm assuming SPSS used the
formula with some correction factor but couldn't find the formula in
SPSS or on the internet.  So my question is: How does SPSS calculate the
median?

Thanks, RB



=======
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

=====================
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

[Richard Ristow]

At 05:59 PM 2/7/2008, ViAnn Beadle wrote:

>I'm curious as to what the "consensual" way to calculate a median is. IIRC,
>there are about 4 ways to do this. You'll find the algorithms at
>
>http://support.spss.com/ProductsExt/SPSS/Documentation/Statistics/algorithms

Thanks!

And since I had to poke a little to find the discussion: The
documentation for EXAMINE goes into the definitions of 'median' (see
the treatment of 'Percentiles') the most completely of any I've found yet.

=====================
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