Abstract:
The usual chi-squared testing procedure to test independence between variables in two-way contingency tables assumes that the observations are independent or come from simple random sampling. This assumption is invalid for data from complex survey designs, and hence some adjustments to this test have been developed in the recent past.
