Rules of Classification of Data

The following general rules should be followed while making classification of data in order that the objectives of a classification may be properly meted out.

1. Exhaustability

The classification should be made in an exhaustive manner so that each and every item of the data must belong to any one of the classes without leaving any item to be shown under any class viz. miscellaneous class, or ‘catch all’ class. Provision for a residual class always signifies a weak point in the classification.

2. Exclusiveness

The classification of data should be made in such a manner that the different classes stand mutually exclusive from each other without leaving any room for over-lapping of any class. This means that every item should be confined to a particular class and no item should be included in a female class or vice versa. But the classification of mankind into male, female and blind is not mutually exclusive in as much as a blind can be included in both male and female classes. Such type of overlapping classification should always be avoided.

3. Homogeneity

In the classification of data, it must be seen that all the data included in a particular class are of homogeneous, or similar nature, and accordingly each of the different classes must include the data of homogenous nature. If a class includes any data of heterogeneous or different nature, it will call for another subclassification of the data included in the same class.

4. Consistency

When a basis of classification has been decided for a variable, it should be maintained all through not only for the same variable but also for all its related variable, else, it would be difficult to make a comparative study of the data. For instance, if data relating to sales are classified on the basis of region, its related of data like selling and distributing expends should be classified on the same basis of regions.

5. Flexibility

The manner of classification of data should be such that it should allow for changes in time and situation. The old and outdated classes should allow for changes in time and situation. For this, the whole data should be classified into some major classes, and the datailed  subdivisions of the classes should be left to be done from time to time taking note of the changes in the situation. In this way, both consistency and flexibility of classification can be maintained without any  difficulty.

6. Appropriability

The basis of classification decided, should be appropriate to the nature of the data, else, the very purpose of the classification may be defeated. For instance, if a study is designed to determine the economic condition of a section of the people it would be of no use to classify the number of people on the basis of their region. In such a case. They should be classified on the basis of their income or expenditure range.