The MANOVA exaggerates these differences and then sees whether there is any contrast between the two. The MANOVA is used to see whether (going to the previous example) the 3 factors help reduce math anxiety or whether they reduce public speaking anxiety. They may be helpful with one, but not helpful with the other. The ANOVA lumps them, but the MANOVA separates them and this distinction is very important. The MANOVA, therefore, also shows the researcher more differences that an ANOVA overlooks, as well as avoiding the possibility of a Type 1 error (namely saying that there is a significant result when there isn't').

On the other hand, it is more complicated than the ANOVA so the researcher can make mistakes. It also leads the researcher to make more assumptions that may be false. The researcher has to also add more degrees of freedom and this may bring more error into the test. Finally, the MANOVA cannot be used in every instance. The two dependent vairables have to be very different for the MANOVA to te4st them since if they are matched or similar, confusion may occur. In the case of the maths and public speaking anxiety, the research can differentiate these two dependent variables so that they do become two separate factors. However, assessing whether there is a difference between two very similar textbooks may result in confusion. The researcher would then be best off using an ANOVA rather than a MANOVA.

3. A researcher has found a significant F. with their MANOVA. What is the general interpretation of the result? What might the next steps be in the analysis, given the significant F. For the MANOVA?

The general interpretation of the result is that all three factors / independent variables have a significant...

He should do the following:
a. examine the Tests of between Subject-effects for each of the dependent variables to see the rate of significance in each. It may be, for instance, that one or more of the three conditions has a stronger impact on maths anxiety than on reading . Or that one condition has an effect on maths and no effect on reading. The researcher will be able to see the contrast of the conditions in each phobia.

b. He should run a Bonferonni test which would further control for the possibility of a Type 1 error (since as we said, given the complexity of MANOVA they are mire susceptible to it).

c. He can use stepdown analysis. This would place the dependent variables in order of priority and test each in turn. The researcher, for instance, may think that the three conditions have stronger impact on maths than on public speaking, or he may be more interested in testing the math phobia than the public speaking one. He should, therefore, use the stepwise procedure to place math anxiety first followed by public speaking, considering maths first and removing the variables related to the 'math' component, before proceeding to test 'public speaking'.

Step-down analysis helps in various ways, not least that it further seperates the two phobias, controlling for effects in each and analyzing possibel differences. It assesses the independent variable effects on each dependent variable.

It is also best for this case when, as mentioned, there seems to be a .4 correlation between the two phobias (between the two independent variables).