2) I am going to use Cluster Sampling in this case. There are 10 zone exist in data and each of zone have 10 various of cars. In 30 out of 100 cars data, fill up Regular gas cars are 18 and fill up Premier gas cars are 12. Cluster Sampling usually provides less precision than SRS. On the other hand, Cluster Sampling is less expensive and time consuming than a fully random sample.
3)