After the data have been collected and edited properly, the next thing to be done with them is coding. Coding refers to assigning some number or letter to some object for its identification or nomenclature in a secret manner. Thus, coding of data means the assignment of different numbers or letters or both to the different data for their identification and presentation in a secret and mechanical manner. This renders the work of classification and tabulation more affairs and secret from all others. However, coding of the data may be done either before, or after editing of the data but it should always be done either before or at the time of tabulation of the data. While coding the data it must be seen that the different data are grouped into some categories of codes and that the number of such categories are neither too large, nor too small and that each data fits in one category of code only.
In the modern age of computers, coding is done only after translating the data into the language of machines and transferring the quantitative information into punch cards in decimal digits, or binary digits (0 and 1). There are machines to translate the information into the language of machines to make them fit for coding through the computers.
There are, also, machines to de-code the results into the language in which the data were collected originally.
It may be noted that a data processing language known as “Common Business Oriented Language” (COBOL) has been developed by the computer manufactures since 1959. Another language known as Formula Translation (FORTRAN), which is most widely used now-a-days, has been developed by the manufacturers of the International Business Machines (IBM). Thus, there is no much of problems for the programmers in the matter of both coding and decoding.