# Chi-squared test of independence for biased data

I'm working with a survey dataset consisting of 28807 observations (8470 males and 20337 females).

I'm trying to determine the association between dichotomous variables, for instance, sex (Male, Female) vs. hipertension diagnostic (Yes, No). What worries me is that the female/male proportions do not reflect the proportion of the real population. For instance, in my dataset, more than 70% of the observations are women.

So, the contingency "table" (contTable) is:

Male 7471 (No) 999 (Yes)

Female 16832 (No) 3505 (Yes)

I'm using the Chi-squared test of independence to determine if this kind of dichotomous variables are associated, but I'm not sure if this unbalance in the date is going to affect the result.

I used chisq.test(contTable) to perform the test and the results were: X-squared = 133.7446, df = 1, p-value < 2.2e-16

To my understanding, since p<<<0.05 this clearly indicates that my variables are dependent (right?). Will I get the same result if my population included 50% men and 50% women?