Not quite; Simpson's paradox is when splitting a group into subgroups reverses the direction of correlation in every subgroup of the split, which is not the cas

Not quite; Simpson's paradox is when splitting a group into subgroups reverses the direction of correlation in every subgroup of the split, which is not the case in the example presented here. Aside: it's not always true that the direction of correlation in the sliced groups is more correct than that of the larger group. You need causal analysis to know when it is incorrect or correct to slice the data to eliminate confounders.