4 ms·
Ignoring the issues with comparing random sampling to self selection, your analogy falls flat due to the population differences. Here you are assuming a racial/
by Jtype 5y ago
Ignoring the issues with comparing random sampling to self selection, your analogy falls flat due to the population differences. Here you are assuming a racial/sex population distribution equal to the color population distribution of M&Ms, which is drastically different in even the most diverse countries in the world. If your starting population is 90% green M&Ms then pulling 6 greens in a row is quite high.
Pulling grants because the racial/sex output isn't what you want is unethical and anti-science.
- rak1507 5y ago<pedantry> if the population is 90% green, the chance of getting 6 greens in a row is only about 53% (tends to 53% as size -> inf) </pedantry>
- bell-cot 5y agoNo, my point was not that simple. No large human population will have the same race / sex distribution as a standard bag of M&M's. But anyone who got an "A" in Probability 101 can look at the actual distribution of M&M colors, and do the calculations. AND - as rak1507 effectively points out - "all 6 green" is both grounds for a touch of suspicion, and utterly wrong as a basis for concluding actual non-random selection. My other point was about motivated reasoning. One timeless hallmark of human bigotry (by sex, race, religion, age, social class, caste, nobility of blood, alma mater, or whatever) is that, so long as the outcomes of selection are what the bigots desire, then there really is no argument too flawed, nor excuse too weak, to proclaim that the outcomes are fair. But if the outcomes go against the bigots' preferences, then the standards of proof go way, way up.