A general proof using matrices is given proving the equivalence of the Chow test (analysis of covariance) and an appropriate adaptation of the dummy variable technique. Implications of hypothesis testing in the linear regression framework are reviewed for each method. The dummy variable approach is found to have the following advantages: (a) it is more convenient in testing hypotheses regarding the equality of subvectors of the parameter vectors from separate regressions, in particular not requiring the running of new regressions as the Chow test approach sometimes does; and (b) a more general form of hypothesis can be tested, namely that corresponding regression parameters differ by a constant other than zero.