How to run and interpret the results of a MANOVA in SPSS is covered in this video (part 1).

Video Transcript: So let's go ahead and get started with our problem in SPSS. Now in SPSS you can see here we have three variables, gender, where we have males and females; I can turn those value labels off for a minute. We have males and females given by 1s and 2s, and we have our two dependent variables, empathic concern and cognitive perspective taking. So notice here we have to have at least two variables once again for MANOVA, so you can see those here, and then we have our grouping variable which is gender. So let's go ahead and run the MANOVA. So to do that we want to go to Analyze and then General Linear Model and then Multivariate. And here the Multivariate dialog box opens. We're going to take our two dependent variables, empathic concern and perspective-taking, move those over to the dependent variables box, and I just pressed and held the control key to grab both of those at once, and then move gender to Fixed Factor(s). And then let's go ahead and go to Options, select gender and move that over to the Display Means for box, check on Descriptive statistics, Estimates of effect size, and Homogeneity tests, and then click Continue and then go ahead and click OK. Next we get our results out, and you can see here our first table shows us our Between-Subjects Factors, and here we have gender, 1 and 2, 1 is male 2 is female, and notice we have 15 adolescents in each group. Next is our Descriptive Statistics table. We have empathic concern for males and females, so here's these two means, and notice in the sample, let's just go ahead and go through this and ask yourself, which of the two genders scored higher on empathic concern? Now we don't know if this is significant yet, but just descriptively, just visually inspecting. Notice that females scored higher than males by 8.27 points, approximately, on empathic concern. And then here on perspective-taking, we can see that females scored about almost three points higher on perspective-taking than did males. OK next we have Box's Test for Equality of Covariance Matrices. Now this test tests an assumption of the MANOVA, which is that the variance-covariance matrices, also referred to as the covariance matrices, as you see here, are equal for the two groups. Now we're going to have a separate video on this assumption to look at it in more detail, but basically in a nutshell it's very similar to the equal variance assumption for the ANOVA; this assumption is the multivariate generalization or extension of the assumption of equal variances for the ANOVA, it's testing the corresponding variances and covariances are equal for the two groups. But for now all you really need to know is that we want to look at our p-value here and we hope that this is greater than .05. But in fact this test, Box's test the equality of covariance matrices, is so sensitive to departures from non- normality, that a lot of people will even use a standard of p is greater than .001. So, basically, we want this result to be not significant. So if we use the more lenient standard and use p is greater than .001, then you can clearly see, well at either level (.05 or .001), this is not significant. So that provides us some evidence that the variance-covariance matrices are equal for the two groups. And that's an assumption of MANOVA, we want that to hold. So, unlike our significance tests on means, we want this to be greater than .05 or greater than .001. And in that other video I'll go into more detail, some of the subtleties of that, and the research findings as well, and what we can do if we do have a p-value that's less than say, .001, for Box's test. OK next we'll go to our Multivariate Tests box

MANOVA
Multivariate analysis of variance
MANOVA in SPSS
Wilk's Lambda
Pillai Trace
DFA
Hotelling's T-squared


For more on inferential Statistics: [ Ссылка ]

Lifetime access to SPSS videos: [ Ссылка ]
Channel Description: [ Ссылка ]
For step by step help with statistics, with a focus on SPSS. Subscribe today!
YouTube Channel: [ Ссылка ]

свернуть