A sample survey is to be conducted in which it is desired to estimate the proportion of persons over 70 years of age who have evidence of a cognitive impairment. This is to be done by taking a sample of households and giving a simple test of cognitive functioning to all members of the household over 70 years of age. A pilot study of 25 households yielded an average of 1.2 persons over 70 years of age per household with standard deviation equal to .8 and an average of .24 persons over 70 years of age per household showing evidence of cognitive impairment with standard deviation equal to .76. It is known that there are 3058 households in the community and 2949 persons over 70 years of age. How many households should be sampled if it is desired to estimate with 95% confidence the proportion of persons over 70 who show evidence of cognitive impairment within 20% of its true value?