On the relevance of the coupling model to experiments