A company would like to look at the average number of vacation days their employees take each year. They would like their estimate to be within 1.1 days of the true average number of vacation days taken by employees at the 95% confidence level. Another branch of the company did a similar study and found a standard deviation of about 4.8 vacation days. What is the smallest sample size they could take to generate an error of no more than 1.1 days?