A survey research company uses random digit dialing. They have a contract to estimate the percentage of people watching Spanish-language television in several Southern cities. They take a sample of size 1,000 in Dallas, Texas - which has about 1,350,000 residents. They are satisfied with the accuracy of the estimates for Dallas.

Houston has about twice the population of Dallas, but similar demographics. To get about the same accuracy in Houston as in Dallas, the company should use a sample size of .