Let me start by saying that Hans Rosling is one of a kind. He is a data rock star. Rarely do we find such a passion for data combined with a cause. The visualization technique behind gapminder is simple and elegant and leaves no tool in the shed unused, applying them all masterfully to demonstrate what has happened to the world in the last decades, demonstrating that our aggregate truth needs updating. In this video you can see a sample of his narration of a data movie showing what time did to the scatter plotting of countries along number of children per woman and children mortality rate. It’s only when he says “and as child mortality fell women chose to have fewer and fewer babies” that you realize that even such a beautiful technique can obscure and create confusion about causality and its direction. I am not an expert but the sentence seems to suggest that women had more children to replace lost ones suggesting some direction of causality but it is reasonable to conjecture that the other direction is a priori as likely: if you don’t have any children then you have zero child mortality.