stats

I need to solve for the sample size needed in the table below for a non-normal/skewed distribution of data. Assuming a 90% confidence interval, I need to complete the table below, filling in the blanks with the sample size needed to meet the corresponding population size in the 1st column and the margins of error (+/-3, 5 or 10%) in the 3 right columns. I have performed goodness of fit tests using MathWave's EasyFit software on the data (attached) and the Generalized Extreme Value (GEV) distribution seems to be the best fit. Context is provided below the table. Population Size +/-3% +/-5% +/-10% 25 50 75 100 150 200 300 The numbers are ratings on a scale from 0-10 that students have provided class instructors via post-course evaluation survey. For new instructors, there are few ratings and so there is debate about their accuracy. Our goal is to be able to state a minimum number of ratings needed in order to meet our stated parameters (90% confidence, 3/5/10% error) at each of the population sizes listed. The population sizes represent the total number of students who attended a course and the sample size represent the number of survey responses/ratings received. So, if there were 100 students, we want to know how many ratings/surveys we would need from those 100 students in order to be 90% confident that the ratings given represent all 100 students. Also, to simplify the scenario, we are treating all students as part of the population size, regardless of whether they attended the same course. So, if 20 students were in one course and 30 in a second course, we will just declare a population size of 50 total students. Then, we want to solve for how many of those 50 students would need to provide a rating in order to be within our stated parameters. Please: 1) Complete the table 2) Explain how you calculated the data in the table, citing the distribution used, software or formulas used.