Question

Statistically, when doing inference, any event occurring with less than a 5% probability is considered unlikely, or significant. Do you agree with the 5% value? Do you think it should be larger or smaller? Make a case for the cutoff value you prefer, and cite examples to support your argument.

Answer 1

well, 5% means 5/100, or 1/20

so, every 20 times you will get the event you want (statistically speaking)

that is pretty unlikely to happen, but not THAT unlikely

my opinion of unlikely is <1%, cause that is 1/100 or lower, and that means out of 100 times, it will only happen once.

that is unlikely in my opinion ;)

Answer 2

You can think of some examples like... would you want some antibiotic to be (almost) deadly to 5% of the population, or lower? You could argue something like that for a smaller % and so it is favourable to have a smaller probability. Or, you could say perhaps 5% is adequate if you are, perhaps making industrial dried rice and require that the volume of water added to cook the rice is 200 mL for 1 cup of rice, with a margin of 5% so that an extra 10 mL won't make too much of a difference for your preparation (perhaps because the rice grains aren't uniform in size), and that an extra 10 mL of water probably won't ruin your rice anyways since it will probably evaporate with the heat.

Answer 3

The larger margin in the latter case is just that it doesn't require more specialized machinery and so the company doesn't have to worry about getting more efficient (so more costly) machines to do a better job since its not really that important.