data level for Kruskal-Wallis

[J McClure]

Hi,
SPSS v 18. will only allow interval level data for the variable I want
to compare across groups. I thought this test was for ordinal level.
Jan

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

[lucameyer]

Actually Kruskal-Wallis compares median rankings (see e.g:
http://en.wikipedia.org/wiki/Kruskal%E2%80%93Wallis_one-way_analysis_of_variance)
so I you need to rank your ordinsal variable before running the test.

Cheers,
Luca


-----
Mr. Luca Meyer
www.lucameyer.com
--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/data-level-for-Kruskal-Wallis-tp3374966p3377277.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

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

[lucameyer]

Actually Kruskal-Wallis compares median rankings (see e.g: http://en.wikipedia.org/wiki/Kruskal%E2%80%93Wallis_one-way_analysis_of_variance) so I you need to rank your ordinal variable before running the test.

Cheers,
Luca

[Art Kendall]

I believe the ranking is built into the K-W procedure.

People have many meanings for "ordinal data".  If you mean data with pure ranks (about as many values as cases), many non-parametric methods are the same as using the parametric procedures,  but with ranks as input.

If you mean a small number of ordered categories, take a look at categorical regression.  It looks at the data with different levels of measurement, so you can see whether to treat the categories strictly nominal, merely ordered categories, or not-not-too-discrepant from interval level.

Art Kendall
Social Research Consultants

On 2/9/2011 4:29 AM, lucameyer wrote:

Actually Kruskal-Wallis compares median rankings (see e.g:
http://en.wikipedia.org/wiki/Kruskal%E2%80%93Wallis_one-way_analysis_of_variance)
so I you need to rank your ordinal variable before running the test.

Cheers,
Luca


-----
Mr. Luca Meyer
www.lucameyer.com
--
View this message in context: http://spssx-discussion.1045642.n5.nabble.com/data-level-for-Kruskal-Wallis-tp3374966p3377280.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

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