An investigator took samples on size 10 for a study of life of batteries. Later he took samples of size 20. What is your opinion on change in sampling error?