What are the ABC’s of PMF’s?

A probability mass function, often abbreviated PMF, provides a clear map detailing the probability that a discrete random variable takes on a certain value. This function is formally defined as (P(X = x)), where (X) is the random variable and (x) is one of the potential outcomes. The resulting probability must satisfy several key mathematical properties to be considered a valid PMF, ensuring consistency within the framework of probability theory. Essentially, the PMF allows statisticians to summarize the likelihood of all possible discrete outcomes stemming from a random experiment.

The Foundational Example: Rolling a Single Die

To properly introduce the concept of the PMF, let us consider the classic scenario of rolling a fair, six-sided die one time. If we designate (X) as the discrete random variable representing the number that the die lands on, then (X) can take on the integer values {1, 2, 3, 4, 5, 6}. Since the die is fair, each outcome is equally likely. The probability that (X) takes on any one of these specific values is uniform, resulting in the following set of defined probabilities described by the Probability Mass Function:

• P(X=1): 1/6
• P(X=2): 1/6
• P(X=3): 1/6
• P(X=4): 1/6
• P(X=5): 1/6
• P(X=6): 1/6

This example highlights the core purpose of the PMF: it establishes that there is an equal chance that the die could land on any number between 1 and 6, thereby quantifying the randomness inherent in the physical act of rolling the die. This simple illustration provides a clear visual and numerical representation of how probability is distributed across all possible outcomes for a discrete event.

Here’s how we would formally represent these probabilities using the notation of a probability mass function, often shown graphically or in table form:

This visual representation succinctly links the potential discrete outcomes (the results of the roll) with their corresponding probabilities (the likelihood of those results). The PMF structure ensures that every possible outcome in the sample space is accounted for, and its probability is explicitly stated.

Further detailed representations, often utilized in textbooks and scholarly documentation, clearly separate the defined probability values from the outcomes they represent, solidifying the mathematical relationship:

Essential Properties and the Concept of Support

One of the most defining characteristics of any valid probability mass function is the requirement that the sum of all probabilities across all possible outcomes must precisely equal 1 (or 100%). This property ensures that the function covers the entire sample space exhaustively, guaranteeing that some outcome must occur. For our standard die example, this condition is easily satisfied:

Sum of probabilities = 1/6 + 1/6 + 1/6 + 1/6 + 1/6 + 1/6 = 6/6 = 1.

The support of a probability mass function refers to the complete set of values that the discrete random variable can potentially take on. It represents the domain over which the PMF is defined and yields a positive probability. In the case of the single die roll, the support would be the set ({1, 2, 3, 4, 5, 6}), as these are the only values the outcome of the dice roll can possibly assume. Any value outside this set is assigned a probability of zero.

A crucial implication of the definition of support is that outside of this defined set, the value for the PMF is identically equal to zero. For instance, the probability that the dice lands on “0” or “7” or “8” is exactly zero, since none of these numbers are included within the support of the random variable (X). This mathematical constraint ensures that the PMF only deals with meaningful, observable outcomes within the context of the defined experiment.

Probability Mass Functions in Practice: Standard Distributions

While the dice roll is an excellent pedagogical tool, PMFs are far more frequently encountered in practical applications through well-established theoretical distributions that model common real-world phenomena. The two most common and critical examples of probability mass functions in practice are the Binomial Distribution and the Poisson Distribution. These distributions provide specific formulas that generate PMFs for events characterized by repeated trials or occurrences over a fixed interval.

The Binomial Distribution PMF

The Binomial Distribution models the number of successes in a fixed number of independent Bernoulli trials, where each trial has only two possible outcomes (success or failure) and the probability of success remains constant. If a random variable (X) follows a Binomial distribution, then the probability that (X) equals (k) successes out of (n) trials can be found using the following fundamental PMF formula:

P(X=k) = _nC_k * p^k * (1-p)^n-k

This formula incorporates the principles of combinations and probability laws to calculate the exact likelihood of observing (k) successes. The components of this formula are defined as follows:

• n: Represents the total number of trials in the experiment.
• k: Represents the desired number of successes we are calculating the probability for.
• p: Represents the probability of success on any single, given trial.
• _nC_k: This is the binomial coefficient, calculating the number of ways to obtain exactly (k) successes in (n) trials without regard to order.

For example, suppose we flip a fair coin 3 times ((n=3)) where the probability of success (getting a head) is (p=0.5). We can use the formula above to determine the probability of obtaining 0, 1, 2, or 3 heads during these 3 independent flips. This demonstrates the power of the PMF in quantifying outcomes for repetitive events:

• P(X=0) = ₃C₀ * .5⁰ * (1-.5)^3-0 = 1 * 1 * (.5)³ = 0.125
• P(X=1) = ₃C₁ * .5¹ * (1-.5)^3-1 = 3 * 0.5 * (.5)² = 0.375
• P(X=2) = ₃C₂ * .5² * (1-.5)^3-2 = 3 * 0.25 * (.5)¹ = 0.375
• P(X=3) = ₃C₃ * .5³ * (1-.5)^3-3 = 1 * 0.125 * (.5)⁰ = 0.125

The Poisson Distribution PMF

In contrast to the Binomial distribution, the Poisson Distribution models the number of events occurring within a fixed interval of time or space, provided these events occur independently and at a constant average rate. This distribution is highly valuable for modeling rare events or counts of incidents like customer arrivals, phone calls received by a call center, or traffic accidents. If a random variable (X) follows a Poisson distribution, the probability that (X) equals (k) occurrences is determined by the following PMF formula:

P(X=k) = λ^k * e^{– λ} / k!

Understanding the parameters in the Poisson PMF is essential for applying it correctly:

• λ: Represents the mean number of successes (or occurrences) that happen during the specified interval. Lambda ((lambda)) is often referred to as the rate parameter.
• k: Represents the specific number of successes (occurrences) we are interested in.
• e: Represents Euler’s constant, an irrational number equal to approximately 2.71828, which is fundamental to exponential growth and decay models.

Consider a hypothetical scenario where a particular hospital experiences an average of (lambda = 2) births per hour. We can utilize the Poisson PMF formula to calculate the probability of observing 0, 1, 2, or 3 births (or any integer count) in a given hour:

• P(X=0) = 2⁰ * e^{– 2} / 0! = 0.1353 (The probability of zero births)
• P(X=1) = 2¹ * e^{– 2} / 1! = 0.2707 (The probability of exactly one birth)
• P(X=2) = 2² * e^{– 2} / 2! = 0.2707 (The probability of exactly two births)
• P(X=3) = 2³ * e^{– 2} / 3! = 0.1805 (The probability of exactly three births)

Visualizing a PMF: Clarity through Bar Charts

To convey the distribution of probabilities effectively, probability mass functions are most often visualized using bar charts or histograms. Since a PMF deals exclusively with discrete values, the visualization clearly shows distinct bars for each outcome, where the height of the bar corresponds directly to the probability of that outcome occurring. This visual approach is far more intuitive for understanding the shape and skewness of the distribution than simply viewing a list of numerical probabilities.

For example, the following bar chart illustrates the probabilities associated with the number of births per hour, calculated using the Poisson distribution example detailed in the preceding section. This graphical representation immediately highlights that 1 and 2 births per hour are the most likely outcomes, while 0, 3, and 4 births become progressively less likely:

It is important to note that for distributions like the Poisson, the number of potential occurrences could theoretically extend to infinity. However, as the number of occurrences increases significantly past the mean ((lambda)), the corresponding probabilities rapidly decrease, often becoming negligibly small. This is why, in practical visualizations, probabilities for outcomes beyond a certain threshold (e.g., beyond 10 in this case) may become so low that they are practically invisible or truncated on a standard bar chart.

Summary of the Defining Properties of a PMF

A function must satisfy three fundamental criteria to be rigorously defined as a probability mass function (P(x)) for a discrete random variable (X). These properties ensure that the function adheres to the universal laws of probability:

1. Non-negativity within the Support: All probabilities for values within the defined support must be positive or zero. Specifically, (P(x) geq 0) for all possible values (x). For the dice example, the probability that the die lands on a number between 1 and 6 is a positive fraction, while the probability of all outcomes outside the support is exactly zero.
2. Boundedness between Zero and One: Every individual outcome must have a probability that lies between 0 and 1, inclusive. Formally, (0 leq P(x) leq 1) for all (x). In the case of the standard die roll, the probability for each outcome is 1/6, which is approximately 0.1666666, satisfying this condition.
3. The Law of Total Probability: The sum of all probabilities for every value in the sample space must be equal to 1. This is mathematically expressed as (sum_{x} P(x) = 1). Continuing the dice example, the sum of probabilities is calculated as 1/6 + 1/6 + 1/6 + 1/6 + 1/6 + 1/6 = 1, ensuring the PMF is complete and valid.