Is there an actual sample dataset to illustrate the use of Tetrad?

[jklevy]

It would be helpful if I could see the features of Tetrad illustrated by use of an actual example.

[cg09]

We are making a video to do exactly that. Should be available soon on the CCD website.

On Wed, Dec 26, 2018 at 12:47 PM jklevy notifications@github.com wrote:

It would be helpful if I could see the features of Tetrad illustrated by use of an actual example.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961, or mute the thread https://github.com/notifications/unsubscribe-auth/APmNueH7flmGLSDYLapT_wPKBNEGRnWUks5u87YdgaJpZM4Zh9EI .

[jklevy]

Sounds great. Where is the CCD website? I am hoping that it is a relatively simple example that can highlight all the features. Right now everything is too abstract. Is there a way I can be informed when this is available?

thank you, Jason levy

On Wed, Dec 26, 2018 at 7:49 AM cg09 notifications@github.com wrote:

We are making a video to do exactly that. Should be available soon on the CCD website.

On Wed, Dec 26, 2018 at 12:47 PM jklevy notifications@github.com wrote:

It would be helpful if I could see the features of Tetrad illustrated by use of an actual example.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961, or mute the thread < https://github.com/notifications/unsubscribe-auth/APmNueH7flmGLSDYLapT_wPKBNEGRnWUks5u87YdgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450000336, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp1NeVtzfMhbhPVRWwC5FBVED_0mJks5u87aWgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[cg09]

Mr. Levy,

Just google Center for Causal Discovery and you will find the webpage and links. No one video could illustrate all of the features of the software, unless you wanted to sit watching continuously for a day or two--there is a lot. We will give a video illustrating basic search algorithms. Videos for loading data and for model representation with directed graphs are already available.

I will try to remember to notify you when the next videos are available.

Cb

On Wed, Dec 26, 2018 at 12:52 PM jklevy notifications@github.com wrote:

Sounds great. Where is the CCD website? I am hoping that it is a relatively simple example that can highlight all the features. Right now everything is too abstract. Is there a way I can be informed when this is available?

thank you, Jason levy

On Wed, Dec 26, 2018 at 7:49 AM cg09 notifications@github.com wrote:

We are making a video to do exactly that. Should be available soon on the CCD website.

On Wed, Dec 26, 2018 at 12:47 PM jklevy notifications@github.com wrote:

It would be helpful if I could see the features of Tetrad illustrated by use of an actual example.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961, or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNueH7flmGLSDYLapT_wPKBNEGRnWUks5u87YdgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450000336, or mute the thread < https://github.com/notifications/unsubscribe-auth/AsBzp1NeVtzfMhbhPVRWwC5FBVED_0mJks5u87aWgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450000818, or mute the thread https://github.com/notifications/unsubscribe-auth/APmNuToZ1bTwA2THBGLKBvqs5Vg-NIMSks5u87dxgaJpZM4Zh9EI .

[jklevy]

Great! Yes, I understand that there are videos that explain how to load in data, etc.

If I understand correctly you will contact me when there is a video with a SPECIFIC dataset and example.

thanks so much for your rapid feedback

On Wed, Dec 26, 2018 at 8:15 AM cg09 notifications@github.com wrote:

Mr. Levy,

Just google Center for Causal Discovery and you will find the webpage and links. No one video could illustrate all of the features of the software, unless you wanted to sit watching continuously for a day or two--there is a lot. We will give a video illustrating basic search algorithms. Videos for loading data and for model representation with directed graphs are already available.

I will try to remember to notify you when the next videos are available.

Cb

On Wed, Dec 26, 2018 at 12:52 PM jklevy notifications@github.com wrote:

Sounds great. Where is the CCD website? I am hoping that it is a relatively simple example that can highlight all the features. Right now everything is too abstract. Is there a way I can be informed when this is available?

thank you, Jason levy

On Wed, Dec 26, 2018 at 7:49 AM cg09 notifications@github.com wrote:

We are making a video to do exactly that. Should be available soon on the CCD website.

On Wed, Dec 26, 2018 at 12:47 PM jklevy notifications@github.com wrote:

It would be helpful if I could see the features of Tetrad illustrated by use of an actual example.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961, or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNueH7flmGLSDYLapT_wPKBNEGRnWUks5u87YdgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450000336 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1NeVtzfMhbhPVRWwC5FBVED_0mJks5u87aWgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450000818, or mute the thread < https://github.com/notifications/unsubscribe-auth/APmNuToZ1bTwA2THBGLKBvqs5Vg-NIMSks5u87dxgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450003840, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp6lyhLQz7IZVuSqovKSuJUoZ-8ziks5u87zegaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jdramsey]

I couldn't help oversee this conversation. It seems like you want a single dataset to help you understand all of the features of Tetrad. That, it seems, is an impossibility. For instance, there are algorithms that assume all continuous variables, algorithms that assume all discrete variables, and algorithms that assume a mixture of continuous and discrete variables. You need at least three datasets to explore all of these. There are other dichotomies as well. Why not start with analyzing some particular dataset of interest to you? What kind of data do you typically analyze? Maybe the first thing to do is to figure out whether Tetrad is relevant to the type of data you have in mind, or to the questions you have in mind. Do you have an idea?

[jklevy]

dear Joseph

You are very kind to respond so quickly. It must have been quite a chore to program Tetrad. I am simply looking for ANY basic single dataset to understand the system. All continuous variables are fine (although a mixture might of continuous and discrete might be best)

I think most people learn best by exploring a specific dataset

On Wed, Dec 26, 2018 at 6:12 PM Joseph Ramsey notifications@github.com wrote:

I couldn't help oversee this conversation. It seems like you want a single dataset to help you understand all of the features of Tetrad. That, it seems, is an impossibility. For instance, there are algorithms that assume all continuous variables, algorithms that assume all discrete variables, and algorithms that assume a mixture of continuous and discrete variables. You need at least three datasets to explore all of these. There are other dichotomies as well. Why not start with analyzing some particular dataset of interest to you? What kind of data do you typically analyze? Maybe the first thing to do is to figure out whether Tetrad is relevant to the type of data you have in mind, or to the questions you have in mind. Do you have an idea?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450069903, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp9r_BYWQoJwu6ofgUHQ4TsNmy03eks5u9EjGgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jdramsey]

Well, of course, Tetrad has a rather elaborate API for making up such datasets. You could make as many as you like. In fact, if learning the ropes is what you're interested in, I would recommend working with simulated data rather than real, since you know the truth, and you can simply ask whether you're able to recover the truth. I suppose I could make one for you if you like, and tell you the truth.

[jklevy]

dear Joseph

You are very thoughtful. If you have a dataset that you could make up to evaluate the truth that would be fantatsic. I have many uses for the software but prefer to try initial with simulated data, as you mentioned.

On Wed, Dec 26, 2018 at 6:20 PM Joseph Ramsey notifications@github.com wrote:

Well, of course, Tetrad has a rather elaborate API for making up such datasets. You could make as many as you like. In fact, if learning the ropes is what you're interested in, I would recommend working with simulated data rather than real, since you know the truth, and you can simply ask whether you're able to recover the truth. I suppose I could make one for you if you like, and tell you the truth.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450070664, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp4hACvI3bHeaFdJnNhSAuUXtzE23ks5u9EqlgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jdramsey]

Here's a sample you can use:

sample.zip

I included a sample dataset and the corresponding ground truth for what the causal relationships are that were used to generate the data. It's continuous. There are 1000 samples. You should be able to load this data and run various search algorithms on it to see what you get.

[jklevy]

Wonderful, Joseph. I tried to unzip sampledata and hit File-> Open and received this error message

[image: image.png]

On Wed, Dec 26, 2018 at 6:27 PM Joseph Ramsey notifications@github.com wrote:

Here's a sample you can use:

sample.zip https://github.com/cmu-phil/tetrad/files/2711423/sample.zip

I included a sample dataset and the corresponding ground truth for what the causal relationships are that were used to generate the data. It's continuous. There are 1000 samples. You should be able to load this data and run various search algorithms on it to see what you get.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450071315, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp4TFbmCPMElDneuPNtbmqqrkjmRUks5u9Ew8gaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jklevy]

[image: image.png]

On Wed, Dec 26, 2018 at 6:33 PM Jason Levy jlevy@hawaii.edu wrote:

Wonderful, Joseph. I tried to unzip sampledata and hit File-> Open and received this error message

[image: image.png]

On Wed, Dec 26, 2018 at 6:27 PM Joseph Ramsey notifications@github.com wrote:

Here's a sample you can use:

sample.zip https://github.com/cmu-phil/tetrad/files/2711423/sample.zip

I included a sample dataset and the corresponding ground truth for what the causal relationships are that were used to generate the data. It's continuous. There are 1000 samples. You should be able to load this data and run various search algorithms on it to see what you get.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450071315, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp4TFbmCPMElDneuPNtbmqqrkjmRUks5u9Ew8gaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jdramsey]

When you unzipped it, did you get the two files in it?

[jklevy]

exactly. I sent you the screen shot. Thank you SO MUCH for your guidance. We will have to invite you to Hawaii

On Wed, Dec 26, 2018 at 6:37 PM Joseph Ramsey notifications@github.com wrote:

When you unzipped it, did you get the two files in it?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450072320, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp_ThNoDr2VomRBBi_kt-_qE2cqBUks5u9E6EgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jdramsey]

These are text files. You'll need to open up Tetrad and load them there. The manual should help you out with that.

[jklevy]

When you say load up Tetrad I see this

[image: image.png]

On Wed, Dec 26, 2018 at 6:38 PM Joseph Ramsey notifications@github.com wrote:

These are text files. You'll need to open up Tetrad and load them there. The manual should help you out with that.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450072458, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp9JK0yJwE0ctaofjwtnWP_osz-9Aks5u9E7XgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jklevy]

[image: image.png]

this is how I launch Tetrad from my desktop. This is correct?

Then I see this

[image: image.png]

On Wed, Dec 26, 2018 at 6:40 PM Jason Levy jlevy@hawaii.edu wrote:

When you say load up Tetrad I see this

[image: image.png]

On Wed, Dec 26, 2018 at 6:38 PM Joseph Ramsey notifications@github.com wrote:

These are text files. You'll need to open up Tetrad and load them there. The manual should help you out with that.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450072458, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp9JK0yJwE0ctaofjwtnWP_osz-9Aks5u9E7XgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jdramsey]

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

[jklevy]

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey notifications@github.com wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jklevy]

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey notifications@github.com wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jklevy]

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey notifications@github.com wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jklevy]

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey notifications@github.com wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jdramsey]

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400 , or mute the thread < https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078, or mute the thread https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI .

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

[jklevy]

I understand 100%. The graph is now loaded.

[image: image.png]

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey notifications@github.com wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078, or mute the thread < https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jklevy]

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201. The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey notifications@github.com wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078, or mute the thread < https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[cg09]

Mr. Levy,

There are a lot of search algorithms in TETRAD. None of them reduce to multiple regression. The standard reference for some of them (PC, GES, FCI) is Spirtes, et al., Causation, Prediction and Search, MIT, 2000.

Clark

On Thu, Dec 27, 2018 at 3:56 AM jklevy notifications@github.com wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey <notifications@github.com

wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400 , or mute the thread < https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450107543, or mute the thread https://github.com/notifications/unsubscribe-auth/APmNuRQDwuyvRpe4TZqBu-5NpRAwua0Uks5u9IsqgaJpZM4Zh9EI .

[cg09]

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling < https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201. The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey notifications@github.com wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043, or mute the thread < https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989, or mute the thread https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI .

[jklevy]

Is FCI part of Tetrad. Is this the FCI. It says GFCI https://www.ccd.pitt.edu//wp-content/uploads/2018/10/GFCIc-overview-11_22_2016.pdf

So Tetrad for Path Analysis (no latent variables)

And other tools for CFA and Latent Variable Structural Model

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201. The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey <notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989, or mute the thread < https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jklevy]

Can you help me to confirm where to download the software for latent SEM?

The variables in SEM are measured (observed, manifest) variables (indicators) and factors (latent variables).

So for measured (observed, manifest) variables (indicators) use Tetrad. And for latent variables use FCI or FOFC + Mimbuild.? Where do I download FCI or FOFC + Mimbuild.

thanks so much!!

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201. The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey <notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989, or mute the thread < https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[cg09]

Just open the search box in TETRAD, when it is connected to a data box. The algorithms are all listed.

On Thu, Dec 27, 2018 at 11:07 AM jklevy notifications@github.com wrote:

Can you help me to confirm where to download the software for latent SEM?

The variables in SEM are measured (observed, manifest) variables (indicators) and factors (latent variables).

So for measured (observed, manifest) variables (indicators) use Tetrad. And for latent variables use FCI or FOFC + Mimbuild.? Where do I download FCI or FOFC + Mimbuild.

thanks so much!!

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences <https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201 . The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey < notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038, or mute the thread < https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450180176, or mute the thread https://github.com/notifications/unsubscribe-auth/APmNuS3TguK_OTNNb-dJTvkEx8lHpCe7ks5u9PBOgaJpZM4Zh9EI .

[cg09]

GFCI is better but FCI is in TETRAD. Just look through the search box options.

On Thu, Dec 27, 2018 at 10:10 AM jklevy notifications@github.com wrote:

Is FCI part of Tetrad. Is this the FCI. It says GFCI < https://www.ccd.pitt.edu//wp-content/uploads/2018/10/GFCIc-overview-11_22_2016.pdf

So Tetrad for Path Analysis (no latent variables)

And other tools for CFA and Latent Variable Structural Model

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences <https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201 . The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey < notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038, or mute the thread < https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450168215, or mute the thread https://github.com/notifications/unsubscribe-auth/APmNuQCzbER2RrfDj84GukI2NYf0PD94ks5u9OLogaJpZM4Zh9EI .

[jklevy]

how do I connect the search box to the data box (see my current screen

thank you!

[image: image.png]

On Thu, Dec 27, 2018 at 6:15 AM cg09 notifications@github.com wrote:

GFCI is better but FCI is in TETRAD. Just look through the search box options.

On Thu, Dec 27, 2018 at 10:10 AM jklevy notifications@github.com wrote:

Is FCI part of Tetrad. Is this the FCI. It says GFCI <

https://www.ccd.pitt.edu//wp-content/uploads/2018/10/GFCIc-overview-11_22_2016.pdf

So Tetrad for Path Analysis (no latent variables)

And other tools for CFA and Latent Variable Structural Model

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences < https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201 . The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey < notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450168215, or mute the thread < https://github.com/notifications/unsubscribe-auth/APmNuQCzbER2RrfDj84GukI2NYf0PD94ks5u9OLogaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450181894, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp7A03Ei9JeYOxI6R7G5Wz0gYAVEsks5u9PIagaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[jklevy]

thank you! how do I connect my databox to the search box?

[image: image.png]

On Thu, Dec 27, 2018 at 6:16 AM Jason Levy jlevy@hawaii.edu wrote:

how do I connect the search box to the data box (see my current screen

thank you!

[image: image.png]

On Thu, Dec 27, 2018 at 6:15 AM cg09 notifications@github.com wrote:

GFCI is better but FCI is in TETRAD. Just look through the search box options.

On Thu, Dec 27, 2018 at 10:10 AM jklevy notifications@github.com wrote:

Is FCI part of Tetrad. Is this the FCI. It says GFCI <

https://www.ccd.pitt.edu//wp-content/uploads/2018/10/GFCIc-overview-11_22_2016.pdf

So Tetrad for Path Analysis (no latent variables)

And other tools for CFA and Latent Variable Structural Model

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences < https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201 . The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey < notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy <jlevy@hawaii.edu

wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450168215, or mute the thread < https://github.com/notifications/unsubscribe-auth/APmNuQCzbER2RrfDj84GukI2NYf0PD94ks5u9OLogaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450181894, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp7A03Ei9JeYOxI6R7G5Wz0gYAVEsks5u9PIagaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[cg09]

There should be a video.

On Thu, Dec 27, 2018 at 11:16 AM jklevy notifications@github.com wrote:

how do I connect the search box to the data box (see my current screen

thank you!

[image: image.png]

On Thu, Dec 27, 2018 at 6:15 AM cg09 notifications@github.com wrote:

GFCI is better but FCI is in TETRAD. Just look through the search box options.

On Thu, Dec 27, 2018 at 10:10 AM jklevy notifications@github.com wrote:

Is FCI part of Tetrad. Is this the FCI. It says GFCI <

https://www.ccd.pitt.edu//wp-content/uploads/2018/10/GFCIc-overview-11_22_2016.pdf

So Tetrad for Path Analysis (no latent variables)

And other tools for CFA and Latent Variable Structural Model

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences < https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201 . The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey < notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy <notifications@github.com

wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy <jlevy@hawaii.edu

wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy < jlevy@hawaii.edu> wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450168215 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuQCzbER2RrfDj84GukI2NYf0PD94ks5u9OLogaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450181894, or mute the thread < https://github.com/notifications/unsubscribe-auth/AsBzp7A03Ei9JeYOxI6R7G5Wz0gYAVEsks5u9PIagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450182189, or mute the thread https://github.com/notifications/unsubscribe-auth/APmNucvh7eVNTJCaC77TjZDleapPRhABks5u9PJvgaJpZM4Zh9EI .

[cg09]

Or see the manual.

On Thu, Dec 27, 2018 at 11:17 AM jklevy notifications@github.com wrote:

thank you! how do I connect my databox to the search box?

[image: image.png]

On Thu, Dec 27, 2018 at 6:16 AM Jason Levy jlevy@hawaii.edu wrote:

how do I connect the search box to the data box (see my current screen

thank you!

[image: image.png]

On Thu, Dec 27, 2018 at 6:15 AM cg09 notifications@github.com wrote:

GFCI is better but FCI is in TETRAD. Just look through the search box options.

On Thu, Dec 27, 2018 at 10:10 AM jklevy notifications@github.com wrote:

Is FCI part of Tetrad. Is this the FCI. It says GFCI <

https://www.ccd.pitt.edu//wp-content/uploads/2018/10/GFCIc-overview-11_22_2016.pdf

So Tetrad for Path Analysis (no latent variables)

And other tools for CFA and Latent Variable Structural Model

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences < https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201 . The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey < notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy < notifications@github.com> wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy <jlevy@hawaii.edu

wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy < jlevy@hawaii.edu> wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy < jlevy@hawaii.edu

wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450168215 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuQCzbER2RrfDj84GukI2NYf0PD94ks5u9OLogaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450181894, or mute the thread < https://github.com/notifications/unsubscribe-auth/AsBzp7A03Ei9JeYOxI6R7G5Wz0gYAVEsks5u9PIagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450182364, or mute the thread https://github.com/notifications/unsubscribe-auth/APmNuY2-9DtukC0kLeEpSDx580JixGEAks5u9PKhgaJpZM4Zh9EI .

[jklevy]

I am trying to learn how to connect the search box in TETRAD to the data box. Should the search box be a "child" of the "parent" data box. I can't find a video or instructions in the manual on how to do this

http://cmu-phil.github.io/tetrad/manual/#data_box

On Thu, Dec 27, 2018 at 6:12 AM cg09 notifications@github.com wrote:

Just open the search box in TETRAD, when it is connected to a data box. The algorithms are all listed.

On Thu, Dec 27, 2018 at 11:07 AM jklevy notifications@github.com wrote:

Can you help me to confirm where to download the software for latent SEM?

The variables in SEM are measured (observed, manifest) variables (indicators) and factors (latent variables).

So for measured (observed, manifest) variables (indicators) use Tetrad. And for latent variables use FCI or FOFC + Mimbuild.? Where do I download FCI or FOFC + Mimbuild.

thanks so much!!

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences < https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201 . The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey < notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy notifications@github.com wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy jlevy@hawaii.edu wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy jlevy@hawaii.edu wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450180176, or mute the thread < https://github.com/notifications/unsubscribe-auth/APmNuS3TguK_OTNNb-dJTvkEx8lHpCe7ks5u9PBOgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450181313, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp2mw3d5WBR9oc4y3WCjLZuy1Mmjlks5u9PGIgaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[cg09]

It's simple. In the the left hand column there is a box named "Search" Click on it move to the work space and click again. A search box appears. Go to the next to the top of the left hand column. Click on it and move to the workspace and draw an arrow between the data box and search box.

This is really standard drag and drop.

On Fri, Dec 28, 2018 at 12:30 AM jklevy notifications@github.com wrote:

I am trying to learn how to connect the search box in TETRAD to the data box. Should the search box be a "child" of the "parent" data box. I can't find a video or instructions in the manual on how to do this

http://cmu-phil.github.io/tetrad/manual/#data_box

On Thu, Dec 27, 2018 at 6:12 AM cg09 notifications@github.com wrote:

Just open the search box in TETRAD, when it is connected to a data box. The algorithms are all listed.

On Thu, Dec 27, 2018 at 11:07 AM jklevy notifications@github.com wrote:

Can you help me to confirm where to download the software for latent SEM?

The variables in SEM are measured (observed, manifest) variables (indicators) and factors (latent variables).

So for measured (observed, manifest) variables (indicators) use Tetrad. And for latent variables use FCI or FOFC + Mimbuild.? Where do I download FCI or FOFC + Mimbuild.

thanks so much!!

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences < https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201 . The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey < notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy <notifications@github.com

wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy jlevy@hawaii.edu wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy <jlevy@hawaii.edu

wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy < jlevy@hawaii.edu> wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450180176 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuS3TguK_OTNNb-dJTvkEx8lHpCe7ks5u9PBOgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450181313, or mute the thread < https://github.com/notifications/unsubscribe-auth/AsBzp2mw3d5WBR9oc4y3WCjLZuy1Mmjlks5u9PGIgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450293088, or mute the thread https://github.com/notifications/unsubscribe-auth/APmNufRptazZjn2N3Di_RMBkvBTen2C7ks5u9axugaJpZM4Zh9EI .

[jklevy]

Yes, yes easy enough with your instructions. Thank you so much Clark! I do see the FCI algorithm in Tetrad search box. However, I also see GFCI, FOFC and Mimbuild. Are all four not also part of Tetrad?

[image: image.png]

On Thu, Dec 27, 2018 at 8:50 PM cg09 notifications@github.com wrote:

It's simple. In the the left hand column there is a box named "Search" Click on it move to the work space and click again. A search box appears. Go to the next to the top of the left hand column. Click on it and move to the workspace and draw an arrow between the data box and search box.

This is really standard drag and drop.

On Fri, Dec 28, 2018 at 12:30 AM jklevy notifications@github.com wrote:

I am trying to learn how to connect the search box in TETRAD to the data box. Should the search box be a "child" of the "parent" data box. I can't find a video or instructions in the manual on how to do this

http://cmu-phil.github.io/tetrad/manual/#data_box

On Thu, Dec 27, 2018 at 6:12 AM cg09 notifications@github.com wrote:

Just open the search box in TETRAD, when it is connected to a data box. The algorithms are all listed.

On Thu, Dec 27, 2018 at 11:07 AM jklevy notifications@github.com wrote:

Can you help me to confirm where to download the software for latent SEM?

The variables in SEM are measured (observed, manifest) variables (indicators) and factors (latent variables).

So for measured (observed, manifest) variables (indicators) use Tetrad. And for latent variables use FCI or FOFC + Mimbuild.? Where do I download FCI or FOFC + Mimbuild.

thanks so much!!

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy notifications@github.com wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences < https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201 . The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey < notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy < notifications@github.com

wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy < jlevy@hawaii.edu> wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy < jlevy@hawaii.edu

wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy < jlevy@hawaii.edu> wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450180176 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuS3TguK_OTNNb-dJTvkEx8lHpCe7ks5u9PBOgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450181313 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp2mw3d5WBR9oc4y3WCjLZuy1Mmjlks5u9PGIgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450293088, or mute the thread < https://github.com/notifications/unsubscribe-auth/APmNufRptazZjn2N3Di_RMBkvBTen2C7ks5u9axugaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450301918, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp96ViyTW8Fd3fWj9vH4flICYkhYyks5u9b85gaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

[cg09]

yes they are

On Fri, Dec 28, 2018 at 6:20 AM jklevy notifications@github.com wrote:

Yes, yes easy enough with your instructions. Thank you so much Clark! I do see the FCI algorithm in Tetrad search box. However, I also see GFCI, FOFC and Mimbuild. Are all four not also part of Tetrad?

[image: image.png]

On Thu, Dec 27, 2018 at 8:50 PM cg09 notifications@github.com wrote:

It's simple. In the the left hand column there is a box named "Search" Click on it move to the work space and click again. A search box appears. Go to the next to the top of the left hand column. Click on it and move to the workspace and draw an arrow between the data box and search box.

This is really standard drag and drop.

On Fri, Dec 28, 2018 at 12:30 AM jklevy notifications@github.com wrote:

I am trying to learn how to connect the search box in TETRAD to the data box. Should the search box be a "child" of the "parent" data box. I can't find a video or instructions in the manual on how to do this

http://cmu-phil.github.io/tetrad/manual/#data_box

On Thu, Dec 27, 2018 at 6:12 AM cg09 notifications@github.com wrote:

Just open the search box in TETRAD, when it is connected to a data box. The algorithms are all listed.

On Thu, Dec 27, 2018 at 11:07 AM jklevy notifications@github.com wrote:

Can you help me to confirm where to download the software for latent SEM?

The variables in SEM are measured (observed, manifest) variables (indicators) and factors (latent variables).

So for measured (observed, manifest) variables (indicators) use Tetrad. And for latent variables use FCI or FOFC + Mimbuild.? Where do I download FCI or FOFC + Mimbuild.

thanks so much!!

On Thu, Dec 27, 2018 at 4:59 AM cg09 notifications@github.com wrote:

You can make any variable you want latent by opening the data file and deleting the corresponding column. But for latent variable modeling, you should run FCI or FOFC + Mimbuild.

I don't know what sort of data you have, but if your aim is to find a multiple indicator model, you should try FOFC + Mimbuild.

Clark

On Thu, Dec 27, 2018 at 9:14 AM jklevy <notifications@github.com

wrote:

Is there a way to start with say, something simple like Exploratory Structural Equation Modeling <

https://www.annualreviews.org/abs/doi/10.1146/annurev-clinpsy-032813-153700?intcmp=trendmd

. A simple diagram with latent variables (Intelligence and Academic Performance) is shown below:

[image: image.png]

We are intererested in applying the traditional approaches from standard SEM papers, particularly in social sciences < https://www.annualreviews.org/doi/pdf/10.1146/annurev.psych.51.1.201 . The above is very classic (from wiki). Latent variables are drawn as circles. In your example there are no latent variables? -- only manifest or measured variables (shown as squares?).... Residuals and variances are often drawn as double headed arrows into an object. Note latent IQ variable fixed at 1 to provide scale to the model.

On Wed, Dec 26, 2018 at 11:10 PM Joseph Ramsey < notifications@github.com

wrote:

The underlying causal structure is in the graph file I sent. Just put a graph box o the workbench, double-click, and load it (as text) from the file menu.

On Thu, Dec 27, 2018 at 3:59 AM jklevy < notifications@github.com

wrote:

I do see the manual below. For the data set you provided is there an underlying causal structure we should start with?

http://cmu-phil.github.io/tetrad/manual/

On Wed, Dec 26, 2018 at 10:55 PM Jason Levy < jlevy@hawaii.edu> wrote:

I am not an expert. But I am interested in understanding the various modeling tools that TETRAD provides for your awesome dataset.

For example, I am interested in SEM. My understanding is that multiple regression a special case of structural equations modeling (SEM). So perhaps with your test set you could help me write down a model in a SEM way, and see if it reduces to the simple case of multiple regression.

If so, then I assume we should use the multiple regression tools.. If not, use the complete SEM framework. I would love to see some videos or specific examples but the user manual seems a little abstract. I don't want to limit my use of TETRAD to SEM but that seems like the most ubiquitous concept (since multiple variable regression is a subset)

On Wed, Dec 26, 2018 at 7:51 PM Jason Levy < jlevy@hawaii.edu

wrote:

OK, I figured out how to load your awesome continuous dataset: it has 10 continuous variables and 1000 records

I just followed this video https://www.ccd.pitt.edu/video-tutorials/

I was able to validate that the data is correct. Is there another nice video that gives examples on how to use Tetrad

[image: image.png]

On Wed, Dec 26, 2018 at 6:55 PM Jason Levy < jlevy@hawaii.edu> wrote:

You are so kind, Joseph. I appreciate you taking time out of your busy schedule to help us get started and running. Talk to you tomorrow. I apologize for the late emails but it is only 7 pm in Honolulu now

On Wed, Dec 26, 2018 at 6:47 PM Joseph Ramsey < notifications@github.com> wrote:

I'm not seeing any of the images you're sending, sorry. I'm falling asleep, but I can try to help some more tomorrow.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450073400

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp9bIljMXCOj3061gKJhXws57rdNmks5u9FDagaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450108078

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AJZZR6My_DTQu_jo6IrfmzoGKCQPLFhrks5u9IvngaJpZM4Zh9EI

.

-- Joseph D. Ramsey Special Faculty and Director of Research Computing Department of Philosophy 135 Baker Hall Carnegie Mellon University Pittsburgh, PA 15213

jsph.ramsey@gmail.com Office: (412) 268-8063 http://www.andrew.cmu.edu/user/jdramsey

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub <

https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450110043

,

or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp1YdffkibRaaWXTMRE2akEhDMeJ1ks5u9I5vgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450157989 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuRlx623pcvDq8T6RpxRJsyALl2Skks5u9NWtgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450166038 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp2xPtIrMPrr0eATmiN7aBisDpMchks5u9OBVgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450180176 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNuS3TguK_OTNNb-dJTvkEx8lHpCe7ks5u9PBOgaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub < https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450181313 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/AsBzp2mw3d5WBR9oc4y3WCjLZuy1Mmjlks5u9PGIgaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub <https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450293088 , or mute the thread <

https://github.com/notifications/unsubscribe-auth/APmNufRptazZjn2N3Di_RMBkvBTen2C7ks5u9axugaJpZM4Zh9EI

.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450301918, or mute the thread < https://github.com/notifications/unsubscribe-auth/AsBzp96ViyTW8Fd3fWj9vH4flICYkhYyks5u9b85gaJpZM4Zh9EI

.

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450343941, or mute the thread https://github.com/notifications/unsubscribe-auth/APmNuW5d3FVFd2_FdDRZ3HKSnnVy1QFTks5u9f5ggaJpZM4Zh9EI .

[jdramsey]

"Tetrad" is the tool that contains the pieces that you see. There is some history to the project; to get a sense of how the original algorithms work, see Spirtes, Glymour, and Scheines (2000), Causation, Prediction, and Search. But may pieces have been added since then (and these pieces improved). The tool itself used to be in Pascal; that version dates back to 1991, I think. Since 1998 it's been programmed in Java. There are, as you can see from the Center for Causal Discovery website, tools for accessing the algorithms via R, Python, command line, and a supercomputer. You said you were having trouble finding that website, but if you just Google it you'll find it.

[jklevy]

Thanks Joseph sure i see the CCD website. i just never found a specific case study or example worked out (would be good to have a video or full example)

If GFCI is part of TETRAD .... I cannot understand why someone wrote "GFCI is better but FCI is in TETRAD. "

On Fri, Dec 28, 2018 at 9:34 AM Joseph Ramsey notifications@github.com wrote:

"Tetrad" is the tool that contains the pieces that you see. There is some history to the project; to get a sense of how the original algorithms work, see Spirtes, Glymour, and Scheines (2000), Causation, Prediction, and Search. But may pieces have been added since then (and these pieces improved). The tool itself used to be in Pascal; that version dates back to 1991, I think. Since 1998 it's been programmed in Java. There are, as you can see from the Center for Causal Discovery website, tools for accessing the algorithms via R, Python, command line, and a supercomputer. You said you were having trouble finding that website, but if you just Google it you'll find it.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/cmu-phil/tetrad/issues/961#issuecomment-450415029, or mute the thread https://github.com/notifications/unsubscribe-auth/AsBzp7sWRuFHPAM6GK7TcJNXJtjSW9T_ks5u9nI4gaJpZM4Zh9EI .

-- Jason Levy Professor Disaster Preparedness and Emergency Management University of Hawaii Phone: (808)-689-2492 Email: jlevy@hawaii.edu