Mixed ANOVA Unequal Gender Ratio

[bobbyk]

Dear all,

For my thesis I'm doing a factorial Mixed ANOVA in which I look at gender differences. As it's a mixed ANOVA, everyone goes through the same conditions. The problem however is that I have 10 men and 26 women in my sample. Therefore the groups that I compare when looking at gender, aren't equal. I was wondering if there was a specific test I could do to calculate how problematic this difference in gender ratio is and if there is a test I could do to make up for this discrepancy?

Thank a lot in advance.

 

[fed2]

I think if gender is independent then it is not a real issue, except possible there is some reduction in the power of some tests. What do you mean by 'makeup for this discrepancy'?

 

[obh]

Hi Bob,

The calculation is different for unbalance groups since there is a correlation between the factors.
You need to choose a function that supports an unbalanced design. there are 3 types that deal with the correlation in different ways. Usually using the conservative types 2 or 3. (If you use R you may use Anova function from "car")

 

[bobbyk]

I think if gender is independent then it is not a real issue, except possible there is some reduction in the power of some tests. What do you mean by 'makeup for this discrepancy'?

Well there are non-parametric tests that you can use to make up for the violation of assumptions/ in which you don't need to comply to certain assumptions right? So I was wondering if there was a similar test you can do when there is an unequal gender ratio. But if I understand your reply, there isn't one, or at least you don't have to do one (if gender is independent)?

 

[bobbyk]

Hi Bob,

The calculation is different for unbalance groups since there is a correlation between the factors.
You need to choose a function that supports an unbalanced design. there are 3 types that deal with the correlation in different ways. Usually using the conservative types 2 or 3. (If you use R you may use Anova function from "car")

What makes you say that there is a correlation?

 

[obh]

What makes you say that there is a correlation?

When having unbalance design it cause the correlation, when you know value of one categorical variable the probability to get a value from the other categorical variable is higher or lower, depend if the unbalance is up or down.

 

[bobbyk]

When having unbalance design it cause the correlation, when you know value of one categorical variable the probability to get a value from the other categorical variable is higher or lower, depend if the unbalance is up or down.

Okay, I see. Thanks sm for answering!

 

[fed2]

yeah i think your worried about something you shouldn't worry about. It depends what hypotheses you are testing, but in pretty much any practical situation, its a non-issue.

 

[bobbyk]

yeah i think your worried about something you shouldn't worry about. It depends what hypotheses you are testing, but in pretty much any practical situation, its a non-issue.

Okay, thank you so much!