1

I received the following question via email:

I am trying to run a five factor model with uncorrelated factors. Amos keeps saying when I go to run the analysis that one variable is unnamed. I cannot find it for the life of me and am 100% certain that I have labelled all the variables.

After running Amos tutorials in universities, it seems like every second student encounters this error. However, experienced users rarely if ever experience the error.

Questions

  • What causes the "one variable unnamed error" in Amos?
  • How can it be presented?
  • How can a model be recovered?
Jeromy Anglim
  • 33,939
  • 30
  • 115
  • 173

3 Answers3

2

Causes of the problem

Amos requires that all variables in a model have names. All latent variables have to have names that are not in the linked dataset, and all observed variables have to have names in the linked dataset. All names must be unique.

While it is possible just to forget to name a variable, I find that the real problem occurs when a user cannot see why which variable is not named.

From my experience, the most common cause of the unnamed variable error is that there is a variable that the user cannot see. This can occur because (a) a variable was drawn that is particularly small, (b) a variable is drawn on top another variable.

Another possibility is that a label has been added but not a variable name. Labels are optional. Variable names are required.

Fixing the problem

Sometimes it is simpler just to draw the model again rather than try to diagnose and fix the problem. That said, a few options for diagnosing the problem:

  • Use the name unobserved variable plugin. If this leads a name popping up, then that is the problem.
  • Use the Tables View (next to the path diagram table) and look at the list of variables to see if you can any unnamed variables. Note that in version 21, returning to path diagram can destroy your path diagram formatting, but this is meant to be fixed in future versions, and you can always control+z to undo the formatting changes.
  • Try to resize the diagram in various ways in order to uncover an unnamed object; combine selection tools with the reshape tool; try this with and without "preserve symmetries" on.

Avoiding the problem

  • Draw neat path diagrams of appropriate size so that all variables are clearly visible
  • Be careful to correctly drag and drop when drawing variables
  • Drag and drop observed variables from the dataset
  • Use the name unobserved variables plugin for residuals.
Jeromy Anglim
  • 33,939
  • 30
  • 115
  • 173
2

I faced the same problem. I realized that I had created an extra error object by mistake and unfortunately, since it was out of view, I couldn't see that element. This is how I fixed it: I changed the view to legal landscape (View tab > Interface properties > Paper size drop down > Landscape-Legal). And I simply deleted that object.

Z-B
  • 31
  • 2
1

I just had the same problem. I solved it using the table view (you may find it on the lower part of the drawing window) and erasing all the empty lines, which are corresponding to these "ghost variables". Then, everything was OK. Hope this works!

user5135168
  • 11
  • 1