Binomial Distributions

How to recognize binomial random variable: 1. The sample size n is fixed. 2. The n observations (or trials) are independent. 3. There are only 2 possible outcomes for each observation. They are labeled Success and Failure 4. The probability of success is the same for each trial. Let p = success probability and 1 p = failure probability 5. The binomial random variable is . . . X = the number of successes out of n observations. Binomial distributions are identified specifically by two parameters: n and p. X takes on values 0, 1, 2, . . . , n Notation:
X ~ B ( n, p )

-1-

Mean: Variance:

= np

2 = np (1 p )

Standard deviation:

= np(1 p )

The Binomial Probability Formula:

P ( x = r ) = P ( r ) = Cn , r p r (1 p )
n r

( ) , then the probability of observing 16 Suppose successes out of 20 observations is

X ~ B 20,0.7

P (16 ) = C20,16 ( 0.7 ) P (16 ) = 0.1304

16

( 0.3) 4

Find the mean, variance, and standard deviation of X. Mean: Variance:

= np = 20(0.7) = 14

2 = np(1 p ) = 20(0.7)(0.3) = 4.20

-2-

Standard deviation:

= np (1 p ) = 4.20 = 2.05

Table 3 in Appendix II contains probability distributions for a variety of binomial distributions. Look up the following probabilities: For X ~ B(20, 0.7), find P(X=16). Answer: P(X=16) = 0.130 For X ~ B(8, 0.45), find P(X > 5). Answer: P(X > 5) = P(6) + P(7) + P(8) = 0.070 + 0.016 + 0.002 = 0.088

-3-

More examples: 1. A research team found that 10% of all people who wear ties wear them so tight that they actually reduce blood flow to the brain. In a random sample of 9 people wearing ties, what is the probability that . . . a. no ties are too tight?
X ~ B ( 9, 0.10 ) P( x = 0 ) = 0.387

b. at least one tie is too tight?

P( x 1) = 1 P( x = 0 ) =1 0.387 = 0.613

c. more than two ties are too tight? 2. If X ~ B(12, 0.37) find P(x = 5).
12 5 7 P ( 5) = 5 ( 0.37 ) ( 0.63) P ( 5) = 0.2163

P ( x > 2 ) = 1 P( x 2 ) = 1 P( 0 ) P (1) P ( 2 ) P ( x > 2 ) =1 0.387 0.387 0.172 = 0.054

The binomial quota problem: Exercise 20. A large bank vault has several automatic burglar alarms. The probability is 0.55 that a single alarm will detect a burglar. (assume the alarms work independently)
-4-

How many alarms should be used to be 99% certain that a burglar is detected by at least one alarm? Solution: Let X ~ B(n, 0.55) and use Table 3 to see for which value of n is
P ( X 1) =1 P ( X = 0 ) = 0.99 At n = 2, P ( X 1) =1 0.203 = 0.797 At n = 3, P ( X 1) =1 0.091 = 0.909 At n = 4, P ( X 1) =1 0.041 = 0.959

At n = 5, P ( X 1) =1 0.019 = 0.981 At n = 6, P ( X 1) =1 0.008 = 0.992

Finally! We see that n = 6 alarms will do it.

-5-