Class mark = (upper class limit + lower class limit) ÷ 2, the mid-point of a class. For grouped data it stands in for every observation in the class, since the class frequency is assumed to be centred there.