In a survey to estimate prevalence of obesity, order the following samples in the expected order of the width of the confidence interval for the estimate prevalence, from smallest to widest, assuming a constant roh.
1. A cluster sample of 100 clusters with 10 persons in each cluster
2. A two-stage sample of 50 clusters with 20 persons selected at random from each cluster
3. A simple random sample of 1000 persons
4. A cluster sample of 10 clusters with 100 persons in each cluster