Obtain path diagrams for path analysi... PreviousNext
Mplus Discussion > Structural Equation Modeling >
Message/Author
 K Sterling posted on Thursday, July 20, 2006 - 6:26 pm
Hi Drs. Muthen,

Is it possible to obtain path diagrams for path analyses with the new version of MPLUS? If so, could you please tell me what command produces the path diagram? I was unable to find the command in the user's manual. Thanks!
 Linda K. Muthen posted on Thursday, July 20, 2006 - 8:31 pm
No, Mplus does not produce path diagrams.
 Vanessa Wight posted on Thursday, May 26, 2011 - 7:19 am
Can you recommend a way to graph path analyses? Is there a companion software package? I have a reviewer who doesn't like my attempt at a path diagram using excel...help!
 Linda K. Muthen posted on Thursday, May 26, 2011 - 10:54 am
We use Freehand.
 Peter Liberman posted on Wednesday, January 23, 2013 - 9:18 am
I have a question about Diagrammer. I would like to alter a diagram, without changing the results, once it has been generated by running an .imp file. I want to remove control variables and their paths to simplify the diagram for use in a paper, but leaving the rest of the diagram as is. I can move but not cut variables with the Move Tool. The Selection Tool does not seem to do anything, including offering a cut option on the Edit menu.

Thanks for any suggestions. I am using MDiagrammer v 1.0 (1).
 Linda K. Muthen posted on Wednesday, January 23, 2013 - 12:24 pm
There is no cut option for the Diagremmer. But you can use the Delete Object and Delete Arrow options to remove parts of a diagram. These options appear in the context menu by right-clicking on an object or arrow. Deleting an object will delete all arrows going to and from this object. The changes made will not affect the original output diagram. You can open the original output diagram by opening the DGM file at any time.
 Peter Liberman posted on Wednesday, January 23, 2013 - 1:49 pm
Thanks!
 Bogdan Voicu posted on Thursday, January 31, 2013 - 9:11 am
I have some difficulties using Diagrammer. When I try to start it, I get an error message: "No Java? This program needs the latest version of Java to run. Please download it at ...".

First time when I have got the message I was surprised. I knew that Java is installed on my Windows 7/64b. Just to be sure, I reinstalled it. However, Mplus (7) keeps asking me to install it...


I also mention that I encounter the error after a MGCFA, then I tried to open the Diagrammer independently of the analysis, either from Mplus, either using the program itself, but I've got the same results.

I suspect that the error is not generated by Mplus, but somehow by my system. I wonder if somebody else encountered the error. If so, can you share the solution?
 Linda K. Muthen posted on Thursday, January 31, 2013 - 9:45 am
Try downloading the 32-bit Java.
 EFried posted on Wednesday, May 29, 2013 - 1:17 am
Linda, is there a way to increase thickness of lines of the variable boxes (similar to arrow lines)? We'd like to use the MPLUS diagrammer to visualize our SEM in the current paper, but the 1px default thickness is not sufficient to properly see the boxes in the PDF proof of the journal.

Thank you!
 Linda K. Muthen posted on Wednesday, May 29, 2013 - 11:04 am
To change the line thickness on a specific arrow, right-click on the line to bring up the context menu. Then choose “Change line thickness”.

To change the line thickness on all arrows, go to the Diagram menu and choose the option “Change line thickness on all arrows”.

In both cases, you will be prompted for the thickness of the line to use. The default is a thickness of 1.
 EFried posted on Thursday, May 30, 2013 - 2:05 am
Linda, we'd like to change the thickness of the lines of the __boxes__ around variables (round for latent, square for observable). It looks silly if the arrows have 3px width, but all the boxes still have their 1px default width.

We have not found a way to do that. Changing thickness of arrows is very intuitive and works very well.

Thank you
 Linda K. Muthen posted on Thursday, May 30, 2013 - 6:10 pm
The current version does not allow the box lines to be thickened.
 Natalia Dmitrieva posted on Thursday, June 06, 2013 - 11:38 am
I would like to export a figure produced by Mplus diagrammer to a higher resolution pdf document (i.e., 1000 dpi), to meet journal requirements. Do you happen to know if this is possible?

Thank you for your help!
Natalia
 Linda K. Muthen posted on Thursday, June 06, 2013 - 3:58 pm
No, the Export to PDF feature does not currently have DPI settings that can be modified.
 Natalia Dmitrieva posted on Thursday, June 06, 2013 - 4:08 pm
Thank you, Linda. Is there any other way to produce a higher resolution figure in Mplus diagrammer?
 Linda K. Muthen posted on Friday, June 07, 2013 - 11:44 am
Not at the present time.
 Natalia Dmitrieva posted on Friday, June 07, 2013 - 1:48 pm
Thank you!
 Jason Edgerton posted on Tuesday, June 11, 2013 - 12:08 pm
I notice that another person posted previously that when they tried to view diagrams of their models Mplus told them to install the latest version of Java. I recently upgraded to Mplus 7.1. When I ran Mplus 7 I had no problem viewing model diagrams, now with Mplus 7.1 it tells me I need the latest version of Java. I have installed the latest version of Java (I've tried both 64 and 32 bit versions) and Mplus still tells me that I need to install Java. Has anyone else come across this problem? Is there a glitch or a solution? BTW, I am running 7.1 in VMWare Fusion with Windows 7 (this configuration gave me no issues with Mplus7).
 Linda K. Muthen posted on Tuesday, June 11, 2013 - 1:01 pm
The 64-bit version is needed. This has worked for everyone so far. Try it again.
 Mya Martin-Glenn posted on Sunday, June 23, 2013 - 5:22 pm
When I use the export to PDF I only get part of my diagram to export (and the paper appears to be landscape instead of portrait. Are there some settings I can change?
 Linda K. Muthen posted on Sunday, June 23, 2013 - 5:29 pm
Please send the files and your license number to support@statmodel.com.
 Jonathan Cottrell posted on Thursday, October 03, 2013 - 3:10 pm
Is there a limit to the number of variables that can be in a model for diagrammer to produce a diagram? And if so, is there any way to increase that limit?
 Linda K. Muthen posted on Thursday, October 03, 2013 - 3:25 pm
I don't think there is an absolute limit but with too many variables the diagram will not look very nice. If you are having a problem, send the files and your license number to support@statmodel.com.
 Fred Danner posted on Sunday, November 10, 2013 - 8:14 am
The diagrammer tool in Mplus 7.11 is excellent. However, one feature option does NOT work for me -- when I select "show only significant" results, none of the non-significant results disappear from the figure. Is there a syntax command that must be in place for this feature to work?
 Linda K. Muthen posted on Sunday, November 10, 2013 - 10:54 am
Please send your output and DGM file to support@statmodel.com. Also, specify which parameter you are concerned with. We have seen the option work in other models that we have tested.
 Frodi Debes posted on Thursday, August 14, 2014 - 12:50 pm
I have model with one latent variable with nine indicators (continuous manifest variables), and thirteen independent variables (some continuous and others dichotomous). When activating the diagrammer from the output file (Alt + D), it draws the figure, but it leaves out the arrow from the independent continuous variable with the smallest absolute effect on the latent variable (STDYX -0.023). All the other arrows are drawn correctly - also the one with the second smallest absolute effect (-0.061). What could be the reason, and can it be fixed?
 Frodi Debes posted on Thursday, August 14, 2014 - 1:04 pm
Sorry! the variable with with the second smallest absolute effect (-0.061) is not drawn either, but all the other are.
 Linda K. Muthen posted on Thursday, August 14, 2014 - 1:04 pm
Please send the relevant files and your license number to support@statmodel.com.
 Simon Schus posted on Thursday, May 28, 2015 - 11:00 am
Hi there :-)

Is there a way to resize the boxes and circles in diagrammer?

In addition, is there a way for the boxes to contain longer descriptive names, other than the variable name? i.e. instead of 'ScaleItem1', could this be replaced with the the wording to the item? I imagine I can do this by right clicking the box I want to rename and clicking Rename, but is there a way to do this in the code in the MPlus input file?

Simon.
 Linda K. Muthen posted on Thursday, May 28, 2015 - 1:24 pm
There is currently no option to change the size of boxes and circles in the Diagrammer. We will add this request to our future list.

There is no option in the Mplus input file to rename text for the circles and boxes.
 Denise Aske posted on Thursday, October 08, 2015 - 8:48 am
Hello,

I was wondering if there were more detailed instructions for using the Mplus Diagrammer v1.3. I am having some trouble getting started.

Thanks!
 Linda K. Muthen posted on Thursday, October 08, 2015 - 11:50 am
Have you looked at the documentation on the website with the user's guide?
 Masoud F. Pour posted on Friday, November 24, 2017 - 9:39 am
Hi
I want to test multiple mediations in an observed path with the following details:
1-5 exogenous variables, all of them sub-scales for a psychological measure.
2-10 mediators, 5 of them are sub-scales of the first construct and the second 5 are sub-scales of a second measure.
3- 2 endogenous variables.
All the variables are observed and continues.

Please could you kindly advice of the relevant syntax for this analysis.

Thanks in advance,
Masoud Fazilat-Pour
 Bengt O. Muthen posted on Friday, November 24, 2017 - 11:32 am
See UG ex 3.11 on our website.
 Masoud F. Pour posted on Sunday, November 26, 2017 - 11:50 am
Dear Bengt
Thanks for the reply. I have checked the section 3.11 in the user guide. The syntax helps to calculate direct effects.
MODEL:
M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 ON X1 X2 X3 X4 X5;
Y1 Y2 ON M1 M2 M3 M4 M5 M6 M7 M8 M9 M10 M11 X1 X2 X3 X4 X5;

For the indirect effects, I used the following syntax:

model indirect:
Y1 IND X1;
Y1 IND X2;
Y1 IND X3;
Y1 IND X4;
Y1 IND X5;
Y2 IND X1;
Y2 IND X2;
Y2 IND X3;
Y2 IND X4;
Y2 IND X5;
output:
stdyx cinterval(bootstrap);

The problem is that the indirect effect command to result in a pooled indirect effect estimate. For example, “Y1 IND X1;” to show the indirect effects of all the mediators but not each of them.
Please could you kindly advise how to calculate the indirect effects for each of the mediators?

Peace,
Masoud
 Bengt O. Muthen posted on Sunday, November 26, 2017 - 4:39 pm
Have a look at the VIA setting (see the UG index).
Back to top
Add Your Message Here
Post:
Username: Posting Information:
This is a private posting area. Only registered users and moderators may post messages here.
Password:
Options: Enable HTML code in message
Automatically activate URLs in message
Action: