2 ms·
It was indeed self-selected (parents volunteered to administer the test) and notes itself that it did not attempt to control for variables, and so should not be
by ethbr1 3y ago
It was indeed self-selected (parents volunteered to administer the test) and notes itself that it did not attempt to control for variables, and so should not be used for comparison purposes.
Also, the demographics are pretty wonky, as they note:
- 94% white
- 97.2% two-parent marriages
- 93.8% Christian mother (no info on other parent?)
- 2-3x general pop's attainment of 4 year degree or higher
- ~2x likelihood to be in the highest income brackets (50k+)
- Essentially no home schooling families in the lowest income bracket (0.8% vs 12.6% of all families with children)
- suoduandao3 3y agoRegarding the first point you raise, here's a different study: https://www.tandfonline.com/doi/abs/10.1080/15582159.2015.998966 https://www.tandfonline.com/doi/abs/10.1080/15582159.2015.99... I honestly think social sciences aren't all that amenable to the scientific method - too hard to control all variables - and I'm sure someone could cherry pick studies to say the opposite of what I'm posting here. But, what else can we do?
- ethbr1 3y agoAgreed. Social sciences are notoriously difficult to rigorously normalize, mostly due to the natural variation and impracticality of recruiting a valid cohort (in $ or time). It's a miracle even drug development on the harder science side works as well as it does. Thanks! I'll give this one a read later.