Confidence intervals with percentile ranks

classic Classic list List threaded Threaded
9 messages Options
Reply | Threaded
Open this post in threaded view
|

Confidence intervals with percentile ranks

Tanja Gabriele Baudson
Hi all,

I have the feeling there is a pretty simple solution to this, but I
have no idea how to implement it: I have two points of measurement
with one year in between and want to compare children's ranks on a
certain scale (e.g., emotional adaptation in school) in terms of
percentiles. How can I determine whether a child's percentile rank has
changed in a significant way?

Thanks very much
Tanya

--
Tanja Gabriele Baudson
Universität Trier
FB I Psychologie
Hochbegabtenforschung und -förderung
54286 Trier
Fon 0651/201-4558
Fax 0651/201-4578
Email [hidden email]
Web http://www.uni-trier.de/index.php?id=9492

=====================
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
Reply | Threaded
Open this post in threaded view
|

Error message in multiple imputation procedure

Joost van Ginkel
Dear list,

I'm trying to run multiple imputation with 10 imputations for a very
large data set with a lot of missing data. Initially the algorithm
starts running but after the last iteration SPSS generates the following
error message:

"The model cannot be built because a computational error has occurred
during the estimation. No output will be displayed."

The resulting imputed data set contains only 2 of the 10 imputed data
sets. Does anyone have any experience with this error message? And if
so, is there a remedy against this problem? Thank you very much in
advance!

Best regards,

Joost van Ginkel

Joost R. van Ginkel, PhD
Leiden University
Faculty of Social and Behavioural Sciences
PO Box 9555
2300 RB Leiden
The Netherlands
Tel: +31-(0)71-527 3620
Fax: +31-(0)71-527 1721

**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.
**********************************************************************

=====================
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
Reply | Threaded
Open this post in threaded view
|

Re: Error message in multiple imputation procedure (follow-up question)

Mbaye Fall Diallo
Hi Joost and List,
I am in the same situation. I am trying to use SPSS Multiple imputation for missing data but it's not working. I don't why. It's my first experience with multiple imputation. Can someone post something on this (online resources, tutorials, references, etc.) ?

Thanks in advance for your help.

Best, 

Mbaye 

> Date: Wed, 20 Oct 2010 13:46:15 +0200
> From: [hidden email]
> Subject: Error message in multiple imputation procedure
> To: [hidden email]
>
> Dear list,
>
> I'm trying to run multiple imputation with 10 imputations for a very
> large data set with a lot of missing data. Initially the algorithm
> starts running but after the last iteration SPSS generates the following
> error message:
>
> "The model cannot be built because a computational error has occurred
> during the estimation. No output will be displayed."
>
> The resulting imputed data set contains only 2 of the 10 imputed data
> sets. Does anyone have any experience with this error message? And if
> so, is there a remedy against this problem? Thank you very much in
> advance!
>
> Best regards,
>
> Joost van Ginkel
>
> Joost R. van Ginkel, PhD
> Leiden University
> Faculty of Social and Behavioural Sciences
> PO Box 9555
> 2300 RB Leiden
> The Netherlands
> Tel: +31-(0)71-527 3620
> Fax: +31-(0)71-527 1721
>
> **********************************************************************
> This email and any files transmitted with it are confidential and
> intended solely for the use of the individual or entity to whom they
> are addressed. If you have received this email in error please notify
> the system manager.
> **********************************************************************
>
> =====================
> 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
Reply | Threaded
Open this post in threaded view
|

Re: Error message in multiple imputation procedure

Art Kendall
In reply to this post by Joost van Ginkel
When introducing a new topic please make a new post, not a reply to a
previous post.

Art

On 10/20/2010 7:46 AM, Ginkel, Joost van wrote:

> Dear list,
>
> I'm trying to run multiple imputation with 10 imputations for a very
> large data set with a lot of missing data. Initially the algorithm
> starts running but after the last iteration SPSS generates the following
> error message:
>
> "The model cannot be built because a computational error has occurred
> during the estimation. No output will be displayed."
>
> The resulting imputed data set contains only 2 of the 10 imputed data
> sets. Does anyone have any experience with this error message? And if
> so, is there a remedy against this problem? Thank you very much in
> advance!
>
> Best regards,
>
> Joost van Ginkel
>
> Joost R. van Ginkel, PhD
> Leiden University
> Faculty of Social and Behavioural Sciences
> PO Box 9555
> 2300 RB Leiden
> The Netherlands
> Tel: +31-(0)71-527 3620
> Fax: +31-(0)71-527 1721
>
> **********************************************************************
> This email and any files transmitted with it are confidential and
> intended solely for the use of the individual or entity to whom they
> are addressed. If you have received this email in error please notify
> the system manager.
> **********************************************************************
>
> =====================
> 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
Social Research Consultants
Reply | Threaded
Open this post in threaded view
|

Re: Confidence intervals with percentile ranks

Art Kendall
In reply to this post by Tanja Gabriele Baudson
Assuming very large norming groups and long test, something like this
could be useful  but do not over interpret.
Open a new instance of SPSS. Copy the syntax below into a syntax window.
Run it.  Is this what you are asking for?

*create some data.
new file.
set seed = 20101019.
INPUT PROGRAM.
LOOP id=1 TO 10.
compute year1 = rnd(rv.uniform(1,99)).
compute year2 = rnd(rv.uniform(1,99)).
compute ydiff = year2 = year1.
compute zyear1 = idf.normal(year1/100,0,1).
compute zyear2 = idf.normal(year2/100,0,1).
compute zdiff = zyear2-zyear1.
compute sort_of_p =1- cdf.normal(abs(zdiff),0,1).
END CASE.
END LOOP.
END FILE.
END INPUT PROGRAM.
FORMATS id (F4.0) year1 year2 ydiff (f3) zyear1 zyear2 zdiff(f5.2)
sort_of_p (f7.3).
var labels
     year1 'percentile in year1'
  year2 'percentile in year2'
  ydiff   'raw change in percentile points CAUTION IN INTERPRETING'
  zyear1 'zscore corresponding to year 1 percentile'
  zyear2 'zscore corresponding to year 2 percentile'
  zdiff   'direction and number of SDs covered by change'
  sort_of_p 'portion of area covered by absolute value of zdiff  CAUTION
IN INTERPRETING'.
t-test pairs = zyear2 zyear1.

list.


Art Kendall
Social Research Consultants


On 10/20/2010 5:27 AM, Tanja Gabriele Baudson wrote:

> Hi all,
>
> I have the feeling there is a pretty simple solution to this, but I
> have no idea how to implement it: I have two points of measurement
> with one year in between and want to compare children's ranks on a
> certain scale (e.g., emotional adaptation in school) in terms of
> percentiles. How can I determine whether a child's percentile rank has
> changed in a significant way?
>
> Thanks very much
> Tanya
>
> --
> Tanja Gabriele Baudson
> Universität Trier
> FB I Psychologie
> Hochbegabtenforschung und -förderung
> 54286 Trier
> Fon 0651/201-4558
> Fax 0651/201-4578
> Email [hidden email]
> Web http://www.uni-trier.de/index.php?id=9492
>
> =====================
> 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
Social Research Consultants
Reply | Threaded
Open this post in threaded view
|

Re: Error message in multiple imputation procedure

SPSS Support-2
In reply to this post by Joost van Ginkel
Dear Joost,

We'd like to get the data you're using so we can see what's happening. I've opened a support case for this and will follow up with you individually.

David Nichols
IBM SPSS Support

From: "Ginkel, Joost van" <[hidden email]>
To: [hidden email]
Date: 10/20/2010 06:48 AM
Subject: Error message in multiple imputation procedure
Sent by: "SPSSX(r) Discussion" <[hidden email]>





Dear list,

I'm trying to run multiple imputation with 10 imputations for a very
large data set with a lot of missing data. Initially the algorithm
starts running but after the last iteration SPSS generates the following
error message:

"The model cannot be built because a computational error has occurred
during the estimation. No output will be displayed."

The resulting imputed data set contains only 2 of the 10 imputed data
sets. Does anyone have any experience with this error message? And if
so, is there a remedy against this problem? Thank you very much in
advance!

Best regards,

Joost van Ginkel

Joost R. van Ginkel, PhD
Leiden University
Faculty of Social and Behavioural Sciences
PO Box 9555
2300 RB Leiden
The Netherlands
Tel: +31-(0)71-527 3620
Fax: +31-(0)71-527 1721

**********************************************************************
This email and any files transmitted with it are confidential and
intended solely for the use of the individual or entity to whom they
are addressed. If you have received this email in error please notify
the system manager.
**********************************************************************

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

Reply | Threaded
Open this post in threaded view
|

Re: Error message in multiple imputation procedure (follow-up question)

SPSS Support-2
In reply to this post by Mbaye Fall Diallo
Hi Mbaye,

Are you getting the same error message as Joost, or is something different happening? There is an example or case study tutorial included in the Help under the Missing Values Option.

David Nichols
IBM SPSS Support

From: Mbaye Fall Diallo <[hidden email]>
To: [hidden email]
Date: 10/20/2010 07:32 AM
Subject: Re: Error message in multiple imputation procedure (follow-up              question)
Sent by: "SPSSX(r) Discussion" <[hidden email]>





Hi Joost and List,
I am in the same situation. I am trying to use SPSS Multiple imputation for missing data but it's not working. I don't why. It's my first experience with multiple imputation. Can someone post something on this (online resources, tutorials, references, etc.) ?

Thanks in advance for your help.

Best,

Mbaye


> Date: Wed, 20 Oct 2010 13:46:15 +0200
> From: [hidden email]
> Subject: Error message in multiple imputation procedure
> To: [hidden email]
>
> Dear list,
>
> I'm trying to run multiple imputation with 10 imputations for a very
> large data set with a lot of missing data. Initially the algorithm
> starts running but after the last iteration SPSS generates the following
> error message:
>
> "The model cannot be built because a computational error has occurred
> during the estimation. No output will be displayed."
>
> The resulting imputed data set contains only 2 of the 10 imputed data
> sets. Does anyone have any experience with this error message? And if
> so, is there a remedy against this problem? Thank you very much in
> advance!
>
> Best regards,
>
> Joost van Ginkel
>
> Joost R. van Ginkel, PhD
> Leiden University
> Faculty of Social and Behavioural Sciences
> PO Box 9555
> 2300 RB Leiden
> The Netherlands
> Tel: +31-(0)71-527 3620
> Fax: +31-(0)71-527 1721
>
> **********************************************************************
> This email and any files transmitted with it are confidential and
> intended solely for the use of the individual or entity to whom they
> are addressed. If you have received this email in error please notify
> the system manager.
> **********************************************************************
>
> =====================
> 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

Reply | Threaded
Open this post in threaded view
|

Re: Error message in multiple imputation procedure (follow-up question)

George Kyakulaga
In reply to this post by Mbaye Fall Diallo
On Thu, 21 Oct 2010 15:45:37 -0500, SPSS Support <[hidden email]>
wrote:

>Hi Mbaye,
>
>Are you getting the same error message as Joost, or is something different
>happening? There is an example or case study tutorial included in the Help
>under the Missing Values Option.
>
>David Nichols
>IBM SPSS Support
>
>
>From:
>Mbaye Fall Diallo <[hidden email]>
>To:
>[hidden email]
>Date:
>10/20/2010 07:32 AM
>Subject:
>Re: Error message in multiple imputation procedure (follow-up question)
>Sent by:
>"SPSSX(r) Discussion" <[hidden email]>
>
>
>
>Hi Joost and List,
>I am in the same situation. I am trying to use SPSS Multiple imputation
>for missing data but it's not working. I don't why. It's my first
>experience with multiple imputation. Can someone post something on this
>(online resources, tutorials, references, etc.) ?
>
>Thanks in advance for your help.
>
>Best,
>
>Mbaye
>
>> Date: Wed, 20 Oct 2010 13:46:15 +0200
>> From: [hidden email]
>> Subject: Error message in multiple imputation procedure
>> To: [hidden email]
>>
>> Dear list,
>>
>> I'm trying to run multiple imputation with 10 imputations for a very
>> large data set with a lot of missing data. Initially the algorithm
>> starts running but after the last iteration SPSS generates the following
>> error message:
>>
>> "The model cannot be built because a computational error has occurred
>> during the estimation. No output will be displayed."
>>
>> The resulting imputed data set contains only 2 of the 10 imputed data
>> sets. Does anyone have any experience with this error message? And if
>> so, is there a remedy against this problem? Thank you very much in
>> advance!
>>
>> Best regards,
>>
>> Joost van Ginkel
>>
>> Joost R. van Ginkel, PhD
>> Leiden University
>> Faculty of Social and Behavioural Sciences
>> PO Box 9555
>> 2300 RB Leiden
>> The Netherlands
>> Tel: +31-(0)71-527 3620
>> Fax: +31-(0)71-527 1721
>>
>> **********************************************************************
>> This email and any files transmitted with it are confidential and
>> intended solely for the use of the individual or entity to whom they
>> are addressed. If you have received this email in error please notify
>> the system manager.
>> **********************************************************************
>>
>> =====================
>> 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,

I also experienced the similar problems yesterday. I would get an error
message saying the constraints I imposed need adjusting. However, if a
scale is 0-5,how can you modify the minimum which must necessarily zero.
When I used direct logs,Ireceived an error message saying there was a
computational error and the process has to stop.

Strangely, the error stopped happening based on some new analyses I did a
minute ago so I could copy and paste the error messages. However, the
system proceeded without any error messages this time. I wonder if there
was a bug in the system which has now been eliminated. Please, advise.

=====================
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
Reply | Threaded
Open this post in threaded view
|

Re: Error message in multiple imputation procedure (follow-up question)

SPSS Support-2
In reply to this post by Mbaye Fall Diallo
Hi George,

I've also opened a support case for you and will contact you individually. We'll need to look at your data and commands to see what's going on.

David Nichols
IBM SPSS Support

From: George Kyakulaga <[hidden email]>
To: [hidden email]
Date: 10/22/2010 01:10 AM
Subject: Re: Error message in multiple imputation procedure (follow-up              question)
Sent by: "SPSSX(r) Discussion" <[hidden email]>





On Thu, 21 Oct 2010 15:45:37 -0500, SPSS Support <[hidden email]>
wrote:

>Hi Mbaye,
>
>Are you getting the same error message as Joost, or is something different
>happening? There is an example or case study tutorial included in the Help
>under the Missing Values Option.
>
>David Nichols
>IBM SPSS Support
>
>
>From:
>Mbaye Fall Diallo <[hidden email]>
>To:
>[hidden email]
>Date:
>10/20/2010 07:32 AM
>Subject:
>Re: Error message in multiple imputation procedure (follow-up question)
>Sent by:
>"SPSSX(r) Discussion" <[hidden email]>
>
>
>
>Hi Joost and List,
>I am in the same situation. I am trying to use SPSS Multiple imputation
>for missing data but it's not working. I don't why. It's my first
>experience with multiple imputation. Can someone post something on this
>(online resources, tutorials, references, etc.) ?
>
>Thanks in advance for your help.
>
>Best,
>
>Mbaye
>
>> Date: Wed, 20 Oct 2010 13:46:15 +0200
>> From: [hidden email]
>> Subject: Error message in multiple imputation procedure
>> To: [hidden email]
>>
>> Dear list,
>>
>> I'm trying to run multiple imputation with 10 imputations for a very
>> large data set with a lot of missing data. Initially the algorithm
>> starts running but after the last iteration SPSS generates the following
>> error message:
>>
>> "The model cannot be built because a computational error has occurred
>> during the estimation. No output will be displayed."
>>
>> The resulting imputed data set contains only 2 of the 10 imputed data
>> sets. Does anyone have any experience with this error message? And if
>> so, is there a remedy against this problem? Thank you very much in
>> advance!
>>
>> Best regards,
>>
>> Joost van Ginkel
>>
>> Joost R. van Ginkel, PhD
>> Leiden University
>> Faculty of Social and Behavioural Sciences
>> PO Box 9555
>> 2300 RB Leiden
>> The Netherlands
>> Tel: +31-(0)71-527 3620
>> Fax: +31-(0)71-527 1721
>>
>> **********************************************************************
>> This email and any files transmitted with it are confidential and
>> intended solely for the use of the individual or entity to whom they
>> are addressed. If you have received this email in error please notify
>> the system manager.
>> **********************************************************************
>>
>> =====================
>> 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,

I also experienced the similar problems yesterday. I would get an error
message saying the constraints I imposed need adjusting. However, if a
scale is 0-5,how can you modify the minimum which must necessarily zero.
When I used direct logs,Ireceived an error message saying there was a
computational error and the process has to stop.

Strangely, the error stopped happening based on some new analyses I did a
minute ago so I could copy and paste the error messages. However, the
system proceeded without any error messages this time. I wonder if there
was a bug in the system which has now been eliminated. Please, advise.

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