A method is presented for the coded modulation of digital data, a modulation being carried out that is coded in multiple stages. In this context, the useful bits for the coded modulation are divided into parallel signal streams, and each signal stream is fed to a specific coder to carry out the channel coding. In the respective coder, the useful bits have tail bits added to them, which, in contrast to the useful bits, are channel-coded using a variable code rate, to achieve a preestablished bit number for a specific coder, so that all signal streams that are coded in parallel contain the same number of bits.