A researcher wants to estimate the average height of students at a university. They take a random sample of 100 students and calculate a 90% confidence interval. If they were to increase the sample size to 400 students and maintain a 90% confidence level, how would the width of the confidence interval change?