I'm responding to your statement that you "need to run a higher-order factorial MANOVA instead of performing MANOVA 3 times separately for each of my independent variables". 

First, I think you meant to say for each of your dependent variables. 

Second, when I see people using MANOVA, I always wonder if they are using it as a precursor to univariate ANOVAs, and if they think it is buying them some protection against Type I error.  If that is why you are doing MANOVA, I suggest you read the classic 1989 article by Huberty & Morris (link below). 

HTH.

http://wweb.uta.edu/management/Dr.Casper/Fall10/BSAD6314/Coursematerial/Huberty%20&%20Morris%201989%20-%20MANOVA%20vs.%20ANOVA.pdf

An interaction occurs when the level of one factor influences the response of another. Suppose that you are looking at times for someone to run a mile. If you use gender and age as 2 variables, you will find that in general, males are faster than females and younger people are faster than older people. However, young women are faster than old men. 

If we are talking about 2 types of medications, if one medication works better for males than females, this is another type of interaction.  

Something to watch out for with interactions, if you variables are not coded properly, you can have issues with collinearity. Interactions between your categorical variables should not be an issue. Between continuous variables will be an issue. 

I'm responding to your statement that you "need to run a higher-order factorial MANOVA instead of performing MANOVA 3 times separately for each of my independent variables". 

First, I think you meant to say for each of your dependent variables. 

Second, when I see people using MANOVA, I always wonder if they are using it as a precursor to univariate ANOVAs, and if they think it is buying them some protection against Type I error.  If that is why you are doing MANOVA, I suggest you read the classic 1989 article by Huberty & Morris (link below). 

HTH.

http://wweb.uta.edu/management/Dr.Casper/Fall10/BSAD6314/Coursematerial/Huberty%20&%20Morris%201989%20-%20MANOVA%20vs.%20ANOVA.pdf

Hi,

Your question: Any good reference for how to conduct Higher-Order Factorial Designs MANOVA in SPSS.

Interaction variable is a variable constructed from an original set of variables to try to represent either all of the interaction present or some part of it.

In exploratory statistical analyses, it is common to use products of original variables as the basis of testing whether interaction is present with the possibility of substituting other more realistic interaction variables at a later stage. When there are more than two explanatory variables, several interaction variables are constructed, with pairwise-products representing pairwise-interactions and higher order products representing higher order interactions. Then, we go for MANOVA. 

Kindly see the following link to do MANOVA in SPSS.

http://www.ucdenver.edu/academics/colleges/nursing/Documents/PDF/MANOVAHowTo.pdf

Thanks.

Hi, Parisa.

You asked this, "I have 3 Dependent Variables and 3 Independent Variables in a study. Some of my IVs are categorical and some of them are continuous. I think I need to run a higher-order factorial MANOVA instead of performing MANOVA 3 times separately for each of my independent variables."

    Presuming that Bruce Weaver is right and you meant performing *ANOVA* 3 times for each *DV,* the piece that catches my attention is that not all your IVs are categorical.  You can run a multivariate General Linear Model (which is how SPSS runs MANOVA) with 3 DV and 3 IV but the IV need to be categorical for the analysis to estimate all the possible interactions between them (1x2, 1x3, 2x3, 1x2x3).  The advantage is that you do not have to calculate product variables, the disadvantage is that you lose a lot of information whenever you categorize a continuous measure.

    You can create the product of two (or three) IVs and include them in a multiple regression analysis to get information about statistical significance, but only for 1 DV at a time.

    Frankly, I think you need structural equations modeling to get at what you want to know.  However, you can get close to a final model by parsing it out (sets of IV at a time, each DV individually, building to an informed set of tests).  Good luck!

I agree with Julia, this URL will be helpful

https://www.youtube.com/watch?v=m0zV_wFGA1I

Best regards

Julia: when you say "You can create the product of two (or three) IVs and include them in a multiple regression analysis to get information about statistical significance, but only for 1 DV at a time." I don't know what you mean. You can reproduce exactly any interaction model in GLM by manually creating the product variables and entering them.

It can be a pain, but I actually prefer this as it is easier (n my view) to control the coding of predictors. That said, I rarely use SPSS these days so I'd run this in R (which has much nicer syntax):

lm( y ~ x1 + x2 + x3, data=my.data) # main effects

lm( y ~ (x1 + x2 + x3)^2, data=my.data) # all two-way interactions and main effects

lm( y ~ x1 + x2*x3, data=my.data) # one two-way interaction and all main effects

lm( y ~ x2*x2*x3, data=my.data) # all main effects and interactions

You are right, Thom, but in SPSS multiple regression is only run with one dependent variable.   I don't read R code, but it looks to me that your code only has one y, when Parisa has y1, y2, and y3.  That is why I would suggest SEM as a way to put together the full model with the three outcomes modeled correctly.

OK that makes sense. Though I'm generally against using MANOVA for these kinds of models as I don't think the outcomes are readily interpretable.

Multivariate multilevel models or SEM would be more flexible (which can be both good an bad).