The manager of a pizza chain in Albuquerque, New Mexico, wants to determine the average size of their advertised 16-inch pizzas. She takes a random sample of 25 pizzas and records their mean and standard deviation as 16.10 inches and 1.8 inches, respectively. She subsequently computes the 95% confidence interval of the mean size of all pizzas as [15.36, 16.84]. However, she kinds this interval to be too broad to implement quality control and decides to re-estimate the mean based on a bigger sample. Using the standard deviation estimate of 1.8 from her earlier analysis, how large a sample must she take if she wants the margin of error to be under 0.5 inch?