I have several questions.

With 3 DVs, you may be considering MANOVA. My question 3 above is a question that Huberty & Morris (1989) raised in their classic article on MANOVA:

Article Multivariate Analysis Versus Multiple Univariate Analyses

If you are considering MANOVA, see also this article:

Article MANOVA: A Procedure Whose Time Has Passed?

Another option to consider, particularly if your DVs are not really quantitative variables (for which means & SDs are defensible) would be to turn things on their head by estimating a logistic regression model with gender as the outcome and with your 3 DVs as the explanatory variables. Frank Harrell described this approach in his well-known textbook on regression modeling: See the excerpt in the attached PDF. But note that it requires having a sufficient number of observations in each gender category. By the 20:1 rule of thumb Harrell describes in the Datamethods Author Checklist (link below), one would need at least 60 observations in each gender category with 3 explanatory variables in the model.

• https://discourse.datamethods.org/t/author-checklist/3407#arrow_rightoverfitting-and-lack-of-model-validation-26

HTH.

Bruce Weaver Hi thank you for your response!

1. The DVs are measured as quantitative variables on a likert scale questionnaire form.

2. I have 3 categories of gender (M/F/other) I have not calculated sample size for each as of yet as Qualtrics is showing discrepancy in response number to participant rate.

3. I believe it is a multivariate research question.

Thank you for the articles provided I will look at these

I have thought about multivariate logistic regression but was struggling in how to group my DVs together.

For context my research question is on public perception of male victims of intimate partner violence .

Multinomial Logistic Regression