In: Statistics and Probability
As online banking technology improves, and people become more comfortable with its use, banks are increasingly electing to replace full-service branches with ATM-only locations (in order to reduce costs). Imagine a local bank is considering making this transition. In order to ensure that such a switch would not result in a significant loss of customers, the bank branch hires a company to conduct a study to sample a portion of its customer base and solicit their thoughts on the possibility of the branch transitioning to ATM-only. You are an intern, working with the statistical expert tasked with completing the study for the bank. She is interested in your thoughts on the project, and has asked you to write up your answers to the following questions, devoting a short paragraph to each question.
1. Identify the sampling method (Cluster, Simple Random Sample, Stratified Random Sample, Convenience, etc.) that would best ensure that your sample would reflect the overall population of branch customers. Briefly explain your reasoning.
2. If the bank’s population is comprised of 2500 local customers, how many do you think would be good to survey?
1. Identify the sampling method (Cluster, Simple Random Sample, Stratified Random Sample, Convenience, etc.) that would best ensure that your sample would reflect the overall population of branch customers. Briefly explain your reasoning.
Simple Random sampling would best ensure that your sample would reflect the overall population of branch customers. Since, it is a question of transforming only 1 branch, so the population is relatively low. It would save time and cost as well. It will be random so not reason of bias. Other methods require dividing the population into groups and then again randomly selecting the groups. They are time consuming and expensive and they would be more efficient if they are for a larger population eg: All the branches of the bank.
2. If the bank’s population is comprised of 2500 local customers, how many do you think would be good to survey?
Usually 10% of the population is considered a good sample size.
Hence, 250 customers would be good to survey.