Background: Ethnic educational inequality is present in almost all Western societies. One possible explanation for the persisting disadvantages of children from immigrant family backgrounds is discrimination. Individual ethnic discrimination has been reported in the form of ethnic bias in school teachers' evaluations and recommendations. Children of immigrant families tend to be disadvantaged with regard to both school teachers' evaluations of their competencies as well as to recommendations of further steps in their educational career. Purpose: The aim of this research paper is to analyse whether ethnic bias is also present at earlier stages of the educational career, in kindergarten teachers' recommendations. The study, conducted in Germany, investigated (a) whether kindergartens' recommendations of school entry (early, regular or delayed school entry) differed with respect to whether the child had a Turkish migration background or was a child of German origin and (b) whether these potential differences can be explained by children's individual skill levels and family background characteristics. Design, sample and methods: The research was carried out from a quantitative secondary analysis, using data from the longitudinal project "Preschool education and educational careers among migrant children", conducted in Germany since 2006. The sample consists of 742 children, with around half of the sample being children of Turkish origin and half being children of German origin, and their parents. Parents were interviewed and children's individual skills were tested using standardised instruments from the Kaufman Assessment Battery for Children. The effect of Turkish origin on kindergarten teachers' school entry recommendations is analysed, controlling for children's social background, individual skills (non-verbal cognition, German language, mathematics) and children's social behaviour, using a stepwise multinomial logistic regression analysis. Results: The analysis indicated that, although kindergarten teachers' school entry recommendations differed according to a child's ethnic background, there was no ethnic bias evident. Children of Turkish origin were significantly less frequently recommended for early school entry than children of German origin and significantly more often for delayed school entry, but these differences vanished when the parents' educational level was controlled and disappeared when children's individual skills and social behaviour were additionally considered in the model. Conclusions: We conclude that, in the sample studied, there was no ethnic bias in kindergarten teachers' school entry recommendations. The disadvantage of children of Turkish origin as compared to the children of German origin can be fully explained by differences in parental educational level and children's individual skills and social behaviour. The remaining effect of social background, however, should be studied further to clarify whether this effect captures other characteristics such as unobserved skills of the children and differences in parental engagement.