Correlation between variables, without the total cases, but only with sum absolute numbers

#1
Hello.

I have the sum numbers of the results of a questionnaire that been made to 5 different countries.
The people that have been asked are approximately 1000 for each country.
In one question the people had to evaluate a variable depended on their perception, from 1 to 10 (1=lowest, 10=highest).
I clustered the answers into 3 categories (1-3: low, 4-6: medium, 7-10: high).
Now i want to check if some demographics affected the perception of the people that they answered 7-10 (high).
The problem is that i don't have the total cases, what each one (of the people) answered, but only the sum absolute numbers for each possible answer for each country ie. 100 males from UK answered 5.
Is there any way to examine the correllation between the demographic variables and the answer?

Thank you in advance