Background for Lesson 5

1 Cumulative Distribution Function

The cumulative distribution function (CDF) exists for every distribution. We define it as
F (x) = P (X ≤ x) for random variable X. If X is discrete-valued, then the CDF is computed
with summation F (x) = xt=−∞ f (t) where f (t) = P (X = t) is the probability mass function
P

(PMF) that we have already seen. If X is continuous, the CDF is computed with an integral
Rx
F (x) = −∞ f (t)dt where f (t) is the probability density function (PDF).

P1
Example: Suppose X ∼ Binomial(5, 0.6). Then F (1) = P (X ≤ 1) = −∞ f (t) =
P−1 P1 5
t
= 50 0.60 (1 − 0.6)5−0 + 51 0.61 (1 − 0.6)5−1 = (0.4)5 +
5−t



t=−∞ 0 + t=0 t 0.6 (1 − 0.6)
5(0.6)(0.4)4 ≈ 0.087.
R2 R2
Example: Suppose Y ∼ Exp(1). Then F (2) = P (Y ≤ 2) = −∞
e−t I{t≥0} dt = 0
e−t dt =
−e−t |20 = −(e−2 − e0 ) = 1 − e−2 ≈ 0.865.

The CDF is convenient for calculating probabilities of intervals. Let a and b be any real
numbers with a < b. Then the probability that X falls between a and b is equal to P (a <
X ≤ b) = P (X ≤ b) − P (X ≤ a) = F (b) − F (a). This concept is illustrated in Figure 1.

F(1) − F(−1) = P(−1 < X < 1)

−3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3 −3 −2 −1 0 1 2 3
x x x

Figure 1: Illustration of using the CDF to calculate the probability of an interval for con-
tinuous random variable X. Probability values are represented with shaded regions in the
graphs.

1
2 Quantile Function
The CDF takes a value for a random variable and returns a probability. Suppose instead
that we start with a number between 0 and 1, call it p, and we wish to find the value x so
that P (X ≤ x) = p. The value x which satisfies this equation is called the p quantile (or
100p percentile) of the distribution of X.

Example: In a standardized test, the 97th percentile of scores among all test-takers is 23.
Then 23 is the score you must achieve on the test in order to score higher than 97% of all
test-takers. We could equivalently call q = 23 the .97 quantile of the distribution of test
scores.

Example: The middle 50% of probability mass for a continuous random variable is found
between the .25 and .75 quantiles of its distribution. If Z ∼ N(0, 1), then the .25 quantile is
−0.674 and the .75 quantile is 0.674. Therefore, P (−0.674 < Z < 0.674) = 0.5.

3 Probability Distributions in R
Each of the distributions introduced in Lesson 3 have convenient functions in R which allow
you to evaluate the PDF/PMF, CDF, and quantile functions, as well as generate random
samples from the distribution. To illustrate, Table 1 lists these functions for the normal
distribution.

Table 1: R functions for evaluating the normal distribution N(µ, σ 2 ).

Function What it does
√
dnorm(x, mean, sd) Evaluate the PDF at x (mean = µ and sd = σ 2 ).
pnorm(q, mean, sd) Evaluate the CDF at q.
qnorm(p, mean, sd) Evaluate the quantile function at p.
rnorm(n, mean, sd) Generate n pseudo-random samples from the normal distribution.

These four functions exist for each distribution, where the d... function evaluates the
density/mass, p... evaluates the CDF, q... evaluates the quantile, and r... generates a
sample. Table 2 lists the d... functions for some of the most popular distributions. The d

2
can be replaced with p, q, or r for any of the distributions, depending on what you want to
calculate. For more details, enter ?dnorm to view R’s documentation page for the normal
distribution. As usual, replace the norm with any distribution to read the documentation
for that distribution.

Table 2: R functions for evaluating the density/mass function for several distributions.
Distribution Function Parameters

Binomial(n, p) dbinom(x, size, prob) size = n, prob = p

Poisson(λ) dpois(x, lambda) lambda = λ
Exp(λ) dexp(x, rate) rate = λ
Gamma(α, β) dgamma(x, shape, rate) shape = α, rate = β
Uniform(a, b) dunif(x, min, max) min = a, max = b
Beta(α, β) dbeta(x, shape1, shape2) shape1 = α, shape2 = β
√
N(µ, σ 2 ) dnorm(x, mean, sd) mean = µ, sd = σ 2
tν dt(x, df) df = ν

Example: Suppose X ∼ Binomial(5, 0.6). Then we can evaluate F (1) = P (X ≤ 1) ≈ 0.087

in R with pbinom(q=1, size=5, prob=0.6). Note also that qbinom(p=0.087, size=5,
prob=0.6) will return 1 as expected.

Example: Suppose Y ∼ Exp(1). The middle 80% of probability mass is located between
the 0.1 and 0.9 quantiles. To find these quantiles of the Exp(1) distribution, save them as
a vector in R: a = c(0.1, 0.9) followed by qexp(p=a, rate=1) which returns the vector
(0.105, 2.303). Therefore, we have P (0.105 < Y ≤ 2.303) = 0.8.

Practice: The remaining lessons require many calculations with distributions. Here are a
few problems to practice in R. Answers are given in blue.

1. Let X ∼ Pois(3). Find P (X = 1). (0.149)

2. Let X ∼ Pois(3). Find P (X ≤ 1). (0.199)

3. Let X ∼ Pois(3). Find P (X > 1). (0.801)

4. Let Y ∼ Gamma(2, 1/3). Find P (0.5 < Y < 1.5). (0.078)

3
 5. Let Z ∼ N(0, 1). Find z such that P (Z < z) = 0.975. (1.96)

6. Let Z ∼ N(0, 1). Find P (−1.96 < Z < 1.96). (0.95)

7. Let Z ∼ N(0, 1). Find z such that P (−z < Z < z) = 0.90. (1.64)

4 Probability Distributions in Excel

Excel also provides convenient functions for evaluating probability distributions. There are
two primary functions which we can modify to accomplish each of the four tasks of computing
a PDF/PMF, computing a CDF, computing a quantile, and generating a pseudo-random
sample. These functions are demonstrated for the normal distribution in Table 3.

Table 3: Excel functions for evaluating the normal distribution N(µ, σ 2 ). Replace x, mean = µ,
√
standard dev = σ 2 , and probability with numbers or cell references.
Function What it does

NORM.DIST(x, mean, standard dev, FALSE) Evaluate the PDF at x (cumulative = FALSE).
NORM.DIST(x, mean, standard dev, TRUE) Evaluate the CDF at x (cumulative = TRUE).
NORM.INV(probability, mean, standard dev) Evaluate the quantile function at probability.
NORM.INV(RAND(), mean, standard dev) Generate one sample (probability=RAND()).

The RAND() function generates a pseudo-random draw from the Uniform(0, 1) distribution,
which if passed through the normal quantile function, produces a draw from the normal
distribution.
The .DIST and .INV functions are available for most of the distributions listed in Table 4,
which gives the PDF/PMF function for each. More information can be obtained by searching
these functions in the Excel help menu.

Example: Suppose X ∼ Binomial(5, 0.6). Then we can evaluate F (1) = P (X ≤ 1) ≈ 0.087

in Excel by entering = BINOM.DIST(1, 5, 0.6, TRUE). Note also that = BINOM.INV(5,
0.6, 0.087) (where trials = 5, probability s = 0.6, and alpha = 0.087) will return 1
as expected.

Example: Suppose Y ∼ Exp(3). The middle 80% of probability mass is located between
the 0.1 and 0.9 quantiles. To find these quantiles of the Exp(1) distribution, save 0.1 and

4
Table 4: Excel functions for evaluating the density/mass function for several distributions.
Replace the arguments with numbers or cell references.
Distribution Function Parameters

Binomial(n, p) BINOM.DIST(x, trials, probability s, FALSE) trials = n, probability s = p

Poisson(λ) POISSON.DIST(x, mean, FALSE) mean = λ
Exp(λ) EXPON.DIST(x, lambda, FALSE) lambda = λ
Gamma(a, b) GAMMA.DIST(x, alpha, beta, FALSE) alpha = a, beta = 1/b
Beta(α, β) BETA.DIST(x, alpha, beta, FALSE) alpha = α, beta = β
√
2
N(µ, σ ) NORM.DIST(x, mean, standard dev, FALSE) mean = µ, standard dev = σ2
tν T.DIST(x, deg freedom, FALSE) deg freedom = ν

0.9 in two cells, say A1 and A2. Note that Excel does not have a EXPON.INV() function,
so we rely on the fact that the the exponential distribution is a special case of the gamma
distribution with a = 1. We calculate the quantiles by entering = GAMMA.INV(A1, 1, 1/3)
and = GAMMA.INV(A2, 1, 1/3) which yields 0.035 and 0.768 (note that the GAMMA functions
in Excel use a scale parameter instead of a rate parameter, hence we use 1/3 instead of 3).
Therefore, we have P (0.035 < Y ≤ 0.768) = 0.8.

Practice: Now try the practice exercises from Section 3 using Excel.