A company is looking into the average number of sick leaves each employee has applied for per year. The company has five departments with different number of employees in which the company estimates that the number of sick leaves applied by each department is largely associated with the number of employees in that department.

The company randomly samples three departments, then samples 50 employees with probabilities proportional to the number of employees in each of the selected department to run their analysis.

a) What kind of design is used?

b) What could go wrong with the design proposed?