# No group differences between group A and group B emerged for sex p=0.6 - How was this tested?

I wanted to report the demographics of my sample, which is a common thing to do in medicine. It means to report mean age, gender proportion etc of your groups. I turned to a paper that conducted a similar experiment as I did and I saw in their demographics table, that they also did tests to test whether the control and the patient group have different demographics. Now I was a bit confused, as they report a p-value for the gender. The only sentence I found regarding this is, that they "found no group difference between group A and B for sex and in the table, there is a p-value 0.6. My question is, what type of test can be used to test for difference in gender between two groups? I am a bit confused and unsure what to do.

• You need to tell us *group differences with respect to what *? Both genders equally represented? Gender is of no importance in some regression/other model? which? ... Feb 18, 2021 at 14:04
• I tried to rephrase the question, but I think what I am trying to ask is, is there a test that allows to test for difference in gender distribution between two groups. Like, a test so I can say "my groups had roughly the same ratio of female/male or my groups differed a lot and were not balanced"..
– CST
Feb 18, 2021 at 17:44
• It would help to post a link to AND a citation for the paper. I like something called the G-test for testing proportion equality, but chi-squared and Fisher’s exact test are popular, too. Context might give away what they did, even if they aren’t explicit, though it also would not surprise me if they consider it tangential to their main point and not important enough to warrant much discussion.
– Dave
Feb 18, 2021 at 17:56
• Altered brain metabolism contributes to executive function deficits in school-aged children born very preterm Schnider 2020 et al., pubmed.ncbi.nlm.nih.gov/32590836
– CST
Feb 18, 2021 at 18:01

You didn't specify how you did the sampling, but if simple random sampling, the number of M in a sample of size $$n$$ from a population with equal number of M, F would be $$\mathcal{Binomial}(n, p=0.5)$$. So you can do a test based on the binomial distribution. That is probably what the paper you referenced did.