Disclosed is a precoding method comprising the steps of: generating a first coded block and a second coded block with use of a predetermined error correction block coding scheme; generating a first precoded signal z1 and a second precoded signal z2 by performing a precoding process, which corresponds to a matrix selected from among the N matrices F[i], on a first baseband signal s1 generated from the first coded block and a second baseband signal s2 generated from the second coded block, respectively; the first precoded signal z1 and the second precoded signal z2 satisfying (z1, z2)T=F[i] (s1, s2)T; and changing both of or one of a power of the first precoded signal z1 and a power of the second precoded signal z2, such that an average power of the first precoded signal z1 is less than an average power of the second precoded signal z2.