According to my research, the coefficient of cubical (thermal) expansion for nitromethane is slightly greater than the coefficient for methanol, and both are about 50% greater than the coefficients for castor oil and synthetic oil. If all of the ingredients are at the same temperature when mixed, the temperature related differences in % by volume values are in the neighborhood of 0.1%, not 2%. If the ingredients are at different temperatures when mixed, the errors can be greater, but still not significant unless the temperature differences are extreme.