Weighing male/female responses (30/70 distribution)

#1
Hi all!

I've got a question. I received 617 responses to a questionnaire however, 445 of them are female (72%). There are two ways to fix this:

1. I could remove 230 female respondents from my data set, bringing my sample back to 385 people. This would leave me with 44% male and 56% female, which would still not be ideal and I would lose massive amounts of beautiful data (I only have 172 male respondents :( )

2. I could assign a certain weight to each respondent. However, because of the large difference between male/female respondents, I'm afraid my data won't be representative anymore.

I could also just separate males and females for the questions I want to answer, thus avoiding a group average :) But this won't look nice in the report, so I'd rather not.

So, the question for you guys: What would you do to make this sample more representative?

Thanks!
Natascha
 
#2
Down weight the the women so that their sum of weight is 0.5. That will also mean that each individual man will have a higher weight.

Don't throw away any data. Of course it is an advantage to have a larger sample, than a smaller!

(The result is an effect of that women are more friendly and more responsible (no kidding))

...and I would lose massive amounts of beautiful data
(You could keep only the females. Then you would ony have "beautiful data" [joke! :) ] )