Why do scientists trust their experiments?
Clinical trial
Why do scientists trust their experiments?
Imagine you're tasting five new ice cream flavors to find the best one. You want to make sure you're not picking the best flavor just because you liked the first one you tried.
To avoid picking favorites, you don't taste the first flavor. Instead, you mix up the flavors and let someone else decide which you liked best. This way, you don't let your first choice influence your final decision.
Example
You taste Flavor A, and it's good. You taste Flavors B, C, D, and E, and they're all good too. Someone else tastes them and says Flavor B is the best. You didn't know you liked Flavor A the most because you didn't taste it first.
Remember this
Double-blind experiments prevent bias by having someone other than you decide which treatment you're getting, making sure you don't pick favorites based on first impressions.
Text adapted from Wikipedia, licensed under CC BY-SA 4.0.
Related concepts
the back-door criterion identifies: sufficient adjustment sets for causal estimation
Can we trust studies without random experiments?
an instrumental variable does: isolates causal effect when you can't randomize
Why can't we always trust what we see?
Bias vs variance: high bias = underfitting, high variance = overfitting
Can a perfect fit to past data predict future events?
log-loss / cross-entropy loss penalizes: confident wrong predictions more heavily
How wrong predictions hurt more than they're supposed to
the condition number κ(A) measures: sensitivity of Ax=b to perturbations
How small changes affect big outcomes
Controlling for a variable
Confounders influence both treatment and outcome
Swipe through 100 ML concepts daily
Open Pocket Polymath