This exploratory study focuses on concepts and their assessment in K-9 computer science (CS) education. We analyzed concepts in local curriculum documents and guidelines, as well as interviewed K-9 teachers in two countries about their teaching and assessment practices. Moreover, we investigated the 'task based assessment' approach of the international Bebras contest by classifying the conceptual content and question structure of Bebras tasks spanning five years. Our results show a variety in breadth and focus in curriculum documents, with the notion of algorithm as a significant common concept. Teachers' practice appears to vary, depending on their respective backgrounds. Informal assessment practices are predominant, especially in the case of younger students. In the Bebras tasks, algorithms and data representation were found to be the main concept categories. The question structure follows specific patterns, but the relative frequencies of the patterns employed in the tasks vary over the years. Our analysis methods appear to be interesting in themselves, and the results of our study give rise to suggestions for follow-up research.