Edward C. answered 03/29/15
Tutor
5.0
(438)
Caltech Grad for math tutoring: Algebra through Calculus
Note that P(2 or more have characteristic) = 1 - P(0 or 1 have characteristic)
P(0 have characteristic) = 0.9810 = 0.8171
since all 10 of the individuals must not have the characteristic
1 individual out of the 10 can have the characteristic in 10 different ways (it can be any 1 out of the 10 individuals) times the probability that 9 in the sample do not have the characteristic and 1 does, which is
P(1 has characteristic) = 10*0.989*0.021 = 0.1667
So P(2 or more) = 1 - 0.8171 - 0.1667 = 0.0162
There are some situations where a binomial distribution like this can be approximated by a Normal distribution and you would end up using a Z-score, but in general the number of samples n must be large enough so that both n*p and
n*(1-p) are both greater than 10. In this case n = 10 and p = 0.02 so n*p = 0.2 so you must resort to using normal probabilistic methods.