The scores of 80 students are grouped as 0–20: 10, 20–40: 18, 40–60: 24, 60–80: 20, 80–100: 8 (class interval: number of students). Assuming the scores are spread evenly within each class, estimate how many students scored MORE than 65.

The mark 65 falls inside the class 60–80, so that class must be split. It spans 20 marks and holds 20 students, and the part above 65 is 80 − 65 = 15 marks wide, that is 15/20 = three quarters of the class, giving $0.75\times 20=15$ students. Every student in 80–100 also scored above 65, adding 8 more, so the estimate is 15 + 8 = 23 students. The value 28 takes the whole of the class 60–80 together with 80–100, which counts the students scoring between 60 and 65 as though they were above 65. The value 20 is the frequency of the class 60–80 on its own, and 15 is the interpolated part with the top class forgotten. This even-spread assumption is the same one that underlies the median formula for grouped data.