What is'loss of information'in classified data?



The frequency distribution summarizes the raw data by making it concise and comprehensible. However, it does not show the details that are found in raw data and leads to loss of information.
When the raw data is grouped into classes, an individual observation has no significance in further statistical calculations.
For example, suppose class 20-30 contains 6 observations: 25, 25, 20, 22, 25 and 28. Such data is grouped as a class 20-30, then individual values have no significance and only frequency, i.e.,6 is recorded and not their actual values.
All values in this class are assumed to be equal to the middle value of the class interval or class mark. Statistical calculations are based only on the values of class mark instead of the actual values.