Methodology

We were able to access lyrics to songs by the Kidz Bop Kidz and the original version of the songs using Genius’s API. Some manual matching was involved to ensure that we chose the correct lyrics if there were multiple songs of the same name. The release year of both versions of each song was collected using Spotify’s API. Again, some manual matching was necessary.

To assess censorship, we first came up with a list of “bad words” and added to it by looking for words that commonly appeared in original songs but not in their Kidz Bop versions. Using regular expressions we tallied how many of each bad word occurred in the original and Kidz Bop versions. To determine censorship trends over time we looked at all of the original songs that had instances of our bad words and calculated what proportion of bad word/original song pairs were removed in the corresponding Kidz Bop versions. Each bad word counts once per song even if it occurs multiple times in the song.

We manually screened a subset of songs for missing and/or unreliable lyrics on Genius, but there could be some lingering artefacts. We also removed some Kidz Bop songs that were on more specialty albums (e.g. holiday albums) from our analysis. Raw data can be found here. Lyrics with particularly offensive words like the n-word and the r-word have been removed from this piece.

This project was originally inspired by a tweet by Andy Baio.