Contingency Tables/Chi Square Test

#1
Hi all,
Please help statistics gurus! I've searched through some of the postings but couldn't find the specific answer I was looking for. If I have population data from a census survey and am interested in examining relationships between two variables, can I simply cross tabulate the variables in a contingency table and describe the observed relationship as is? Since I do not have sample data, nor am I interested in treating it as sample data, then is it correct to assume that I do not need to perform the chi-square measure of association test? Does cell count still matter?

Please provide clear, specific answers explaining the reasoning behind your answer.
 
#3
From what I know about population data, all you want or need to do is describe what is going on. So, I agree with you, there is no need for inferential stats like the Chi-Square test. I do think that you could still use cartain "measures of associaton" like the Odds Ratio (OR). For example, males are twice as likley to agree with whatever compared to females. But there would be no need for confidence intervals, since you are not trying to generalize to a greater group.