The purpose of this study was to assess the reliability of coding mother-neonate interactions. Two coders independently examined videos of two mother-neonate dyads and coded five interactions: mother looking at neonate's face, mother smiling at neonate, mother joggling neonate, mother talking to neonate, and mother touching neonate. High inter-rater reliability was predicted for the first three interactions, which were presumed easy to identify and measure. Reliability analysis of 1,427 interactions revealed that the two coders agreed less than would be expected by chance alone. This suggests that the process was highly dependent on the individual coders. Reevaluating current coding methods will be essential to develop meaningful standards for reliable coding, which in turn should help us better understand mother-neonate bonding.