While computing the mean of grouped data, we assume that the frequencies are :

1. evenly distributed over all the classes

2. centred at the class marks of the classes

3. centred at the upper limits of the classes

4. centred at the lower limits of the classes

To calculate the mean, we need a single representative value for each class. We assume that the data points in that interval are balanced around the middle, known as the Class Mark.

Class mark = $\dfrac{\text{Upper limit + lowerlimit}}{2}$

Hence, option 2 is the correct option.