A new approach to measure the association between two categorical variables in a finite population when the sample sizes are small is introduced in this thesis and is referred as Finite Population Association Test (FPAT). FPAT is related to Fisher's Exact Test (FET) as both these tests use the same base, the hypergeometric distribution, for their derivation. FPAT majorly relies on the property of multivariate hypergeometric distribution's convergence to multinomial distribution. The p-value achieved using the FPAT approaches FET's p-value as we increase the size of the population. In addition to the computation and applications of FPAT in the field of Epidemiology, a quick overview of the tests currently available in the field has been discussed