Interactive book reading (IBR) is an effective method to foster children's language development. Previous studies pointed at smaller gains for children at risk for language and literacy impairments due to socio-economic, ethnic, or minority background or due to home language. However, possible differences in interaction patterns with children at risk were not investigated. Therefore, this research provides an integrated view on input and interaction during reading with a group of at-risk and a group of not-at-risk children. Videos of small-group reading sessions were transcribed verbatim (16 hours, 53 minutes, and 23 seconds of IBR), divided into 18,995 single units of language and coded with a good interrater agreement (? = 0.84, p < 0.001). A literature-based detailed coding scheme was used for coding. Negative binomial models show similar general input patterns. However, upon closer examination, significant differences on specific features in reading between both groups are present, showing that IBR to children with an at-risk background is related to less stimulating input and interaction during reading. To counter Matthew effects and use the full potential of IBR for every group of children, it is imperative that teachers and other adult readers guarantee qualitative interaction on all input features during IBR.