MANOVA

[deleted_user]

I am trying to do a MANOVA using the Student Version. I have 38 questionnaire items as dependent measures and 2 blocking variables. I am beginning to think it is not possible.with the Student Version. Please advise. .

[Paige]

> I am trying to do a MANOVA using the Student Version.
> I have 38 questionnaire items as dependent measures
> and 2 blocking variables. I am beginning to think it
> is not possible.with the Student Version. Please
> advise. .

And the error you are seeing is .... ???

[Doc_Duke]

When you say "student version" are you talking about the "Learning Edition"? LE is just limited by the maximum number of observations in any one procedure; it used to be 1000. GLM can technically handle that big a dependent variable set. However, I suspect other problems. For instance, GLM uses listwise deletion, so you may be getting an effective sample size that is so low that the estimates cannot be computed.

As Paige said, more specificity would be needed for us to comment more specifically.

Doc Muhlbaier
Duke

[deleted_user]

Thanks for your reply doc@Duke and Paige,

Yes, I am a new user of Learning edition and trying to use the "analyze" presets. GLM appears to allow only 1 dv, I have 38. MULTIVARIATE was very useful for a FA and PCA I conducted, but I do not see a Manova selection. I did successfully run a discriminant analysis.

[Paige]

Docfreedaddy

Two people have asked for information about the error you are seeing; and you haven't provided that information. Until you do so, we can't help you.

GLM does allow more than one dependent variable.

[deleted_user]

Thanks Paige,

I did not mention seeing an error, my problem is entering multiple dv's as the GLM Window prompt reads "Dependent variable: limit 1".

[Paige]

Oh.

The "analyze presets" are something that I am not familiar with. However, it is highly likely that these "analyze presets" are set up for ANOVA only; and this one is not set up for MANOVA. Perhaps there is another one set up for MANOVA?

MANOVA can definitely be performed in SAS PROC GLM via creating the proper SAS code.

[deleted_user]

Thanks guys,

The challenge for me is that I have not used the codes and will need to find a guide book and learn it. I am trying to help my 14 yr. old daughter who won her division and placed in the state science fair with a very sophisticated project she is now writing up for publication. I haven't touched an anatysis since grad school many years ago.

If editors are OK with the discrim, I am good. I may just learn the codes anyway as soon as I have a bit more time available. Any suggestions are welcome.

[Doc_Duke]

Enterprise Guide, which is the user interface to Learning Edition, does only allow one dependent variable in ANOVA. I don't have Learning Edition installed, so I can't tell if you can get to the code to modify it.

You may be able to set it up as a univariate ANOVA and then click on the "display code" button and then the "insert code" button. If you can get to the point of inserting code, then you can add as many dependent variables as you wish.

If you can't get to "insert Code", then you can't do MANOVA in LE.