Ordinal regression Test of parallel lines

[NJ1979]

Hi,
When I ran a ordinal regression by PLUM functionality in SPSS , for the test
of parallel lines, I received this  messages
"The Chi-Square statistic is computed based on the log-likelihood value of
the last iteration of the general model. Validity of the test is uncertain."
What does it mean and how to fix this ?

Thank you,
NJ




--
Sent from: http://spssx-discussion.1045642.n5.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

[Rich Ulrich]

It almost surely means that the output has some other warning

which says something more.  Maybe, "The model fails to converge"

or something like that.  Is it that, or something else? 

Increasing the number of iterations is an obvious thing to try, though

that won't work if the solution steps are either drifting or cycling.

 - It's conceivable to me that by might get a convergence if you

/dropped/ some specific case, if that broke some numerical tie.

I suspect that your N is relatively small compared to the number

of variables in the model, because that's when problems are more

likely.  In addition to mentioning your N and number of variables,

you could say whether there /looks like/ there is some (big) effect.

--

Rich Ulrich

---

From: SPSSX(r) Discussion <[hidden email]> on behalf of NJ1979 <[hidden email]>
Sent: Monday, May 4, 2020 11:05 AM
To: [hidden email] <[hidden email]>
Subject: Ordinal regression Test of parallel lines

 

Hi,
When I ran a ordinal regression by PLUM functionality in SPSS , for the test
of parallel lines, I received this  messages
"The Chi-Square statistic is computed based on the log-likelihood value of
the last iteration of the general model. Validity of the test is uncertain."
What does it mean and how to fix this ?

Thank you,
NJ                             




--
Sent from: http://spssx-discussion.1045642.n5.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

[David Greenberg]

For some kinds of estimation, the exact sampling distribution is
unknown, or is only known asymptotically (which is the case for
maximum likelihood estimation). In this case, significance tests are
not precise. There is no remedy for this. David Greenberg, Sociology
Dept., NYU

On Tue, May 5, 2020 at 1:12 AM Rich Ulrich <[hidden email]> wrote:

>
> It almost surely means that the output has some other warning
> which says something more.  Maybe, "The model fails to converge"
> or something like that.  Is it that, or something else?
>
> Increasing the number of iterations is an obvious thing to try, though
> that won't work if the solution steps are either drifting or cycling.
>  - It's conceivable to me that by might get a convergence if you
> /dropped/ some specific case, if that broke some numerical tie.
>
> I suspect that your N is relatively small compared to the number
> of variables in the model, because that's when problems are more
> likely.  In addition to mentioning your N and number of variables,
> you could say whether there /looks like/ there is some (big) effect.
>
> --
> Rich Ulrich
>
>
> ________________________________
> From: SPSSX(r) Discussion <[hidden email]> on behalf of NJ1979 <[hidden email]>
> Sent: Monday, May 4, 2020 11:05 AM
> To: [hidden email] <[hidden email]>
> Subject: Ordinal regression Test of parallel lines
>
> Hi,
> When I ran a ordinal regression by PLUM functionality in SPSS , for the test
> of parallel lines, I received this  messages
> "The Chi-Square statistic is computed based on the log-likelihood value of
> the last iteration of the general model. Validity of the test is uncertain."
> What does it mean and how to fix this ?
>
> Thank you,
> NJ
>
>
>
>
> --
> Sent from: http://spssx-discussion.1045642.n5.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

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

[Bruce Weaver]

See the documentation for the CRITERIA sub-command for info on how to try
some of the things Rich suggested.

https://www.ibm.com/support/knowledgecenter/en/SSLVMB_26.0.0/statistics_reference_project_ddita/spss/advanced/syn_plum_criteria.html

Does the same thing happen if you estimate the model via GENLIN rather than
PLUM?  (When using GENLIN, specify "DISTRIBUTION=MULTINOMIAL LINK=CUMLOGIT"
on the /MODEL sub-command.)

HTH.


Rich Ulrich wrote

> It almost surely means that the output has some other warning
> which says something more.  Maybe, "The model fails to converge"
> or something like that.  Is it that, or something else?
>
> Increasing the number of iterations is an obvious thing to try, though
> that won't work if the solution steps are either drifting or cycling.
>  - It's conceivable to me that by might get a convergence if you
> /dropped/ some specific case, if that broke some numerical tie.
>
> I suspect that your N is relatively small compared to the number
> of variables in the model, because that's when problems are more
> likely.  In addition to mentioning your N and number of variables,
> you could say whether there /looks like/ there is some (big) effect.
>
> --
> Rich Ulrich
>
>
> ________________________________
> From: SPSSX(r) Discussion &lt;

> SPSSX-L@.UGA

> &gt; on behalf of NJ1979 &lt;

> drnagesh@

> &gt;
> Sent: Monday, May 4, 2020 11:05 AM
> To:

> SPSSX-L@.UGA

>  &lt;

> SPSSX-L@.UGA

> &gt;
> Subject: Ordinal regression Test of parallel lines
>
> Hi,
> When I ran a ordinal regression by PLUM functionality in SPSS , for the
> test
> of parallel lines, I received this  messages
> "The Chi-Square statistic is computed based on the log-likelihood value of
> the last iteration of the general model. Validity of the test is
> uncertain."
> What does it mean and how to fix this ?
>
> Thank you,
> NJ
>
>
>
>
> --
> Sent from: http://spssx-discussion.1045642.n5.nabble.com/
>
> =====================
> To manage your subscription to SPSSX-L, send a message to

> LISTSERV@.UGA

>  (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

> LISTSERV@.UGA

>  (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





-----
--
Bruce Weaver
[hidden email]
http://sites.google.com/a/lakeheadu.ca/bweaver/

"When all else fails, RTFM."

NOTE: My Hotmail account is not monitored regularly.
To send me an e-mail, please use the address shown above.

--
Sent from: http://spssx-discussion.1045642.n5.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