ABSTRACT
The most straightforward method of testing the hypotheses is to employ ordinary least squares regression analysis, using some measure of the rate of arms diffusion as a dependent variable. In order to test the hypotheses related to the omitted variables, some means would have to be devised to generate measures which capture the theoretical essence of their models without using core state GDP or world military spending as proxy variables. There are a number of different possible ways to solve an autocorrelation problem. One method is to assume that the autocorrelation results from an inadequate specification of the autoregressive term in the model. From the conceptual perspective, the diffusion term(s) incorporate into the model the idea that there is an “underlying” rate of diffusion, which can be “disturbed” by other exogenous factors.
