Two big takeaways:

(1) Mindsets matter more than socioeconomic background (though obviously are not the only important factor)

(2) A mix of teacher-directed and inquiry-based education leads to better outcomes, though I think the outcome variable is a bit skewed (teachers are better at teaching to the test than children, so you might decide that a lower test score but more intellectually-curious child is preferable to the converse; the main question is obviously how to measure intellectual curiosity absent test scores)