Worth remembering that this is causal inference, so we are taking correlations and trying to imply causality: could well be wrong. The evidence for a quadratic relationship is clear though (~13%…