All elements have the same probability of being chosen directly as part of the sample. It can be with replacement or without replacement, depending on whether the selected units will be able to be re-elected, becoming part of the population again. The procedure for selecting the random sample is to choose the desired amount (for example, 20 television programs), number the universe or population to be studied from 0 to 500, for example, all television programs broadcast in one week in generalist television in Spain) and choose the corresponding numbers from the table of random numbers: first by pointing to a random number and following in any direction until completing the next 19, in this case (Wimmer and Dominick, 1996: 71). The final selection (for example: 44, 16, 82, 3,…) must be applied to the mentioned numbering of the programs.

The two requirements for random number formation are that each unit or component of the sample must have the same opportunity as the rest to be chosen and the procedure must be free of any subjective intervention by the researcher (Wimmer and Domminick, 1996: 72).

Among its advantages, it is said that this type of sample does not require a detailed knowledge of the composition of the population. External validity can be inferred statistically. A representative group can be easily obtained (though not in all cases). As disadvantages, it is necessary to prepare a complete list of the population and it is a more expensive procedure than others.