Amos & naming variables

[Jill Adelson-2]

I am teaching a structural equation modeling course using Amos.  During lab this week, we ran across an error I have never seen before.  All students were using the same data and building the same model.  (It was just a path analysis model with observed variables only).  Several students got the message that Amos could not calculate estimates because there were unnamed variables in the model.  However, all variables were named. We checked by looking at object properties as well as by pulling up the list of variable names.  If we clicked "new" and rebuilt the model, it would run - so it happened on random computers random times.  Has anyone else experienced this or know what we should do so we don't have to rebuild the models?  Any help or insight would be appreciated.

Thank you,
Jill

--
Jill L. Adelson, Ph.D.
Assistant Professor
Educational Psychology, Measurement, and Evaluation Program
Educational and Counseling Psychology Department
College of Education and Human Development
University of Louisville

[Raimundas Vaitkevicius]

Yes, several of my students experienced this problem in a computer class this May. We didn't manage to find out the cause.

Raimundas Vaitkevčius
statistician and lecturer
Vytautas Magnus university
Kaunas, Lithuania

--- Den fre 2010-09-03 skrev Jill Adelson <[hidden email]>:

Från: Jill Adelson <[hidden email]>
Ämne: Amos & naming variables
Till: [hidden email]
Datum: fredag 3 september 2010 19:04

I am teaching a structural equation modeling course using Amos.  During lab this week, we ran across an error I have never seen before.  All students were using the same data and building the same model.  (It was just a path analysis model with observed variables only).  Several students got the message that Amos could not calculate estimates because there were unnamed variables in the model.  However, all variables were named. We checked by looking at object properties as well as by pulling up the list of variable names.  If we clicked "new" and rebuilt the model, it would run - so it happened on random computers random times.  Has anyone else experienced this or know what we should do so we don't have to rebuild the models?  Any help or insight would be appreciated.

Thank you,
Jill

--
Jill L. Adelson, Ph.D.
Assistant Professor
Educational Psychology, Measurement, and Evaluation Program
Educational and Counseling Psychology Department
College of Education and Human Development
University of Louisville

[SPSS Support-2]

Hi Jill,

In our experience the likely reason for this is that the .amw files in question have somehow become corrupted, making regeneration or using a copy of someone else's would be required.

David Nichols
Statistical Support
SPSS, an IBM Company

From:	Jill Adelson <[hidden email]>
To:	[hidden email]
Date:	09/03/2010 12:37 PM
Subject:	Amos & naming variables
Sent by:	"SPSSX(r) Discussion" <[hidden email]>

---

I am teaching a structural equation modeling course using Amos.  During lab this week, we ran across an error I have never seen before.  All students were using the same data and building the same model.  (It was just a path analysis model with observed variables only).  Several students got the message that Amos could not calculate estimates because there were unnamed variables in the model.  However, all variables were named. We checked by looking at object properties as well as by pulling up the list of variable names.  If we clicked "new" and rebuilt the model, it would run - so it happened on random computers random times.  Has anyone else experienced this or know what we should do so we don't have to rebuild the models?  Any help or insight would be appreciated.

Thank you,
Jill

--
Jill L. Adelson, Ph.D.
Assistant Professor
Educational Psychology, Measurement, and Evaluation Program
Educational and Counseling Psychology Department
College of Education and Human Development
University of Louisville

[anglmill]

This is a huge problem, it just happended to me as well.  When you are working with a complex path model, it takes a VERY long time to build a new one, time that was just wasted for no reason.  Is there a patch for this or something?

Angie L. Miller, Ph.D.
Assistant Research Scientist
Center for Postsecondary Research
Indiana University Bloomington

[baumer66]

Dear

Look at this

https://www-304.ibm.com/support/docview.wss?uid=swg21480913

The problem is with "invisible" variables

Good luck

[autumnwu]

Dear Jill,

I came into the same problem and i tried this way. Transfer Excel data to SPSS file and make new one. Then open this new file in AMOS, drag variables from "list variables in data set" into manifest variables, then rerun AMOS. It works in my case. I also have a try on complex model and if i rebuild the whole model, that really waste too much time. Hope this could help and work in other's case.

[Peter Gries]

I know this was discussed last year, but I wonder if any solutions have been found since then. All my varibales are named in a asimple path model, but AMOS will not run giving me this  "1 variable is unnamed" error message. Ideas?

Pete Gries, The Unviersity of Oklahoma

I have found that sometimes there are "ghost" variables that hide behind the variables of interest.  Try to copy the model to paste it into an excel sheet to see if you can find any unnamed/ unnecessary boxes or circles.

Sent from my iMind

On Dec 9, 2011, at 9:44 PM, Peter Gries <[hidden email]> wrote:

> I know this was discussed last year, but I wonder if any solutions have been
> found since then. All my varibales are named in a asimple path model, but
> AMOS will not run giving me this  "1 variable is unnamed" error message.
> Ideas?
>
> Pete Gries, The Unviersity of Oklahoma
>
> --
> View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Amos-naming-variables-tp2802599p5063520.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

[Colleen Kirk]

I had the same problem.  There were two "ghost" variables in the model.  I used the plugin "name unobserved variables" to name them, and that made their assigned names appear in the model.  Then I could delete them and the model ran.

Very strange.  Thanks for the posts.