An improved method for evaluating the moderator temperature coefficient (MTC) of reactivity in pressurized water reactors based on a newly established technique is described. Previous work has correlated stochastic fluctuations in the reactor core-exit temperature with similar fluctuations in the in-core neutron flux. The frequency response function between these two stochastic signals has been shown to be proportional to the moderator temperature coefficient. Though this method has been shown to agree well with reactor core design and measured values of the MTC, the method requires over three hours of data processed via the fast Fourier transform, and has certain limitations which suggest that the data analysis be done another way. In the present work, the multivariate autoregressive analysis method is utilized to process under ten minutes of the plant data used previously. Nearly identical results were obtained using only 5 % of the data used in the prior analysis.