2 ms·
I'm the author of this blog. Thank you for calling this out! I'll update the example to fix this :)
by Lior539 3y ago
I'm the author of this blog. Thank you for calling this out! I'll update the example to fix this :)
- hammock 3y agoThe example is fine, you are calling out confounding variables. Just call it confounders, instead of Simpson paradox.
- keithwinstein 3y agoFWIW, the arithmetic in that example also has a glitch. For the "Mobile" "Control" case, 100/3000 is about 3% rather than 10%.