This paper estimates a nightly demand model for National Hockey League attendance for five teams from different divisions. The magnitudes of the coefficients are compared for each team to determine if factorsthat influence attendance affect each team the same. All variables except games played in October were found to be statistically different between at least two teams. Games played between division opponents and previous season success had the largest regional variation while “calendar” variables (weekend and games played in March and April) have the lowest regional differences. The analysis shows that a general attendance model of the NHL is inaccurate and can cause misleading interpretations for purposes of marketing and profit maximization.