As for the second part of your question: It is good if students are able to do something when we have just taught them how to do it, but it can't really be called learning unless they can still do it sometime later. So a delay posttest should be conducted quite a while after the posttest, to see whether students have remembered what you taught them and can apply it. Then we could say that they learnt something.
A delayed posttest in language teaching research is essential to assess long-term retention, the application of language skills in real-life contexts, and the impact of forgetting. It helps confirm the lasting effectiveness of a teaching method, serving as evidence of sustainable learning. The delayed posttest should be conducted several weeks or months after the initial posttest, allowing researchers to gauge learners' ability to retain and apply their language skills over time. This sequence of assessments, starting with a pretest, followed by an immediate posttest, and concluding with a delayed posttest, provides a comprehensive picture of language learning outcomes and the durability of those outcomes.