The invention relates to a method for encoding digital data, in particular of data processed in a microprocessor unit. In the method according to the invention for a respective data word (A, B, C) of a series of data words to be encoded subsequently a parity code (P(A), P(B), P(C)) is computed on the basis of the data of the respective data word (A, B, C). Further the respective data word (A, B, C) is altered with the aid of the data word (A, B, C) preceding it in the series, wherein the altered data word (Aa, Ba, Ca) and the parity code (P(A), P(B), P(C)) represent the encoded data word (Ac, Bc, Cc) and the encoded data word (Ac, Bc, Cc) can be decoded with the aid of the data word (A, B, C) preceding it in the series.