Getting pairwise comparisons for Kruskal Wallis

[Art Kendall]

How can I get pairwise comparisons for the Kruskal Wallis test in NPAR TESTS?

Alternatively,
How can I export the output from NPTESTS to .pdf or .rtf so that all the views are shown?

Also, what have other list members done when publishing results that come from model viewer results when the document is black and white?

Art Kendall
Social Research Consultants
===================== 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]

P.S.
Of course getting the pairwise comparisons other than ranking the variables and running UNIANOVA.

I was hoping there was a trick to get presentable results without a lot of explanations or cutting-and-pasting.

The arrangement in nptests would be great if it could be easily printed for dissertations, journals, etc.

Art


On 2/6/2011 9:45 AM, Art Kendall wrote:

How can I get pairwise comparisons for the Kruskal Wallis test in NPAR TESTS?

Alternatively,
How can I export the output from NPTESTS to .pdf or .rtf so that all the views are shown?

Also, what have other list members done when publishing results that come from model viewer results when the document is black and white?

Art Kendall
Social Research Consultants
===================== 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

[Ryan]

Does this do what you want?

NPTESTS
  /INDEPENDENT TEST (y) GROUP (group) KRUSKAL_WALLIS(COMPARE=PAIRWISE)
  /CRITERIA ALPHA=0.05  CILEVEL=95.

Ryan

On Sun, Feb 6, 2011 at 9:45 AM, Art Kendall <[hidden email]> wrote:

> How can I get pairwise comparisons for the Kruskal Wallis test in NPAR
> TESTS?
>
> Alternatively,
> How can I export the output from NPTESTS to .pdf or .rtf so that all the
> views are shown?
>
> Also, what have other list members done when publishing results that come
> from model viewer results when the document is black and white?
>
> Art Kendall
> Social Research Consultants
> ===================== 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

[Ryan]

Ignore my previous response. I didn't read your entire post carefully.
At the moment, the only way I can think of running all pairwise
comparisons in NPAR TESTS would be the following:

npar tests
 /m-w = y by group (1,2)
 /m-w = y by group (1,3)
 /m-w = y by group (2,3).

That is, you could run multiple Mann-Whitney U tests. These tests are
equivalent to running separate Kruskal Wallis H tests for each
pairwise comparison. As I'm sure you are aware, these tests are NOT
the same as running post-hoc pairwise tests using NPTESTS.

Ryan

On Sun, Feb 6, 2011 at 9:45 AM, Art Kendall <[hidden email]> wrote:

> How can I get pairwise comparisons for the Kruskal Wallis test in NPAR
> TESTS?
>
> Alternatively,
> How can I export the output from NPTESTS to .pdf or .rtf so that all the
> views are shown?
>
> Also, what have other list members done when publishing results that come
> from model viewer results when the document is black and white?
>
> Art Kendall
> Social Research Consultants
> ===================== 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

[Marta Garcia-Granero]

Hi Art:

Did you take a look at http://gjyp.nl/marta/KruskalWallisPostHoc.sps ?

The macros cover both asymptotic methods (see Sheskin "HANDBOOK OF PARAMETRIC AND NONPARAMETRIC STATISTICAL PROCEDURES") and exact methods based on pairwise Mann-Whitney's tests followed by different methods of EWER control (Bonferroni, Holm, Hochberg...).

HTH,
Marta GG

El 06/02/2011 17:49, Art Kendall escribió:

P.S.
Of course getting the pairwise comparisons other than ranking the variables and running UNIANOVA.

I was hoping there was a trick to get presentable results without a lot of explanations or cutting-and-pasting.

The arrangement in nptests would be great if it could be easily printed for dissertations, journals, etc.

Art


On 2/6/2011 9:45 AM, Art Kendall wrote:

How can I get pairwise comparisons for the Kruskal Wallis test in NPAR TESTS?

Alternatively,
How can I export the output from NPTESTS to .pdf or .rtf so that all the views are shown?

Also, what have other list members done when publishing results that come from model viewer results when the document is black and white?

Art Kendall
Social Research Consultants
===================== 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

[Art Kendall]

Thank you I'll take a look at that.

Art

On 2/7/2011 9:32 AM, Marta García-Granero wrote:

Hi Art:

Did you take a look at http://gjyp.nl/marta/KruskalWallisPostHoc.sps ?

The macros cover both asymptotic methods (see Sheskin "HANDBOOK OF PARAMETRIC AND NONPARAMETRIC STATISTICAL PROCEDURES") and exact methods based on pairwise Mann-Whitney's tests followed by different methods of EWER control (Bonferroni, Holm, Hochberg...).

HTH,
Marta GG

El 06/02/2011 17:49, Art Kendall escribió:

P.S.
Of course getting the pairwise comparisons other than ranking the variables and running UNIANOVA.

I was hoping there was a trick to get presentable results without a lot of explanations or cutting-and-pasting.

The arrangement in nptests would be great if it could be easily printed for dissertations, journals, etc.

Art


On 2/6/2011 9:45 AM, Art Kendall wrote:

How can I get pairwise comparisons for the Kruskal Wallis test in NPAR TESTS?

Alternatively,
How can I export the output from NPTESTS to .pdf or .rtf so that all the views are shown?

Also, what have other list members done when publishing results that come from model viewer results when the document is black and white?

Art Kendall
Social Research Consultants
===================== 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

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