This article draws on the ongoing discussion in some European countries about whether and how non-religious worldviews and pupils should be included in religious education (RE). This discussion lacks empirical knowledge about the specific learning conditions of these pupils in RE -- a desideratum that this article addresses. Building on a terminological classification of non-religiosity proposed by Lois Lee, this study distinguishes between pupils that show anti-religious, indifferent and positive different relations to religion. It investigates whether they differ significantly from most pupils in RE classes regarding active participation and perception of relevance. To this end, this article draws on data from 952 pupils who attended Catholic or Protestant religious education classes in southwestern Germany (state of Baden-Württemberg). The results of the ANOVAs show that pupils with anti-religious, indifferent and positive different relations indeed have lower scores in active participation and perception of relevance in religious education than the majority of pupils, which can be interpreted as an indicator of poorer learning conditions for the former. Furthermore, significant differences were found between pupils with anti-religious, indifferent and positive different relations, which underlines the importance of distinguishing between different types of non-religiosity in the discussion.