Probability and Probability Distributions

Introduction

Human resource managers routinely make decisions under uncertainty. An organisation may not know whether an employee will accept a promotion, whether absenteeism will exceed a threshold, how many applicants will pass a selection test, or whether training will improve performance. Probability provides a formal language for describing uncertainty, while probability distributions provide models for analysing uncertain outcomes.

In business mathematics and statistics, probability is not a prediction of certainty. It is a numerical measure of the likelihood of an event, based on equally likely outcomes, observed frequencies, or informed judgement. Probability distributions extend this idea by assigning probabilities to the possible values of a random variable. They support workforce planning, risk assessment, recruitment analytics, quality control, compensation analysis and evidence-based decision-making.

Learning Objectives

By the end of this chapter, students should be able to:

• explain experiments, sample spaces, events and probability;
• apply the addition, multiplication and conditional-probability rules;
• distinguish independent and mutually exclusive events;
• use Bayes’ theorem in classification and diagnostic problems;
• distinguish discrete from continuous random variables;
• calculate and interpret expected value, variance and standard deviation;
• apply the binomial, Poisson and normal distributions;
• evaluate assumptions and limitations of probability models in HR decisions.

1. Foundations of Probability

1.1 Experiments, outcomes and sample spaces

A random experiment is a process whose outcome cannot be known with certainty in advance, although its possible outcomes can be identified. Examples include selecting an applicant at random, recording the number of absences in a month, or observing whether an employee leaves within one year.

An outcome is one possible result. The sample space, denoted by S, is the set of all possible outcomes. An event is a subset of the sample space. For example, if an employee’s performance rating is classified as {low, satisfactory, high}, then the event “rating is high” contains the single outcome high, while “rating is at least satisfactory” contains satisfactory and high.

If outcomes are equally likely, the classical probability of an event A is:

Here, n(A) is the number of outcomes in event A and n(S) is the number of outcomes in the sample space. This formula applies only when the outcomes are equally likely and the sample space is correctly specified.

Probability satisfies three fundamental axioms:

1. 0 ≤ P(A) ≤ 1;
2. P(S) = 1;
3. If A and B cannot occur together, then P(A ∪ B) = P(A) + P(B).

The complement of A, written Aᶜ, is the event that A does not occur:

1.2 Addition rules

The union A ∪ B means that A or B or both occur. In general:

The intersection A ∩ B means that both A and B occur. The subtraction term prevents the overlap from being counted twice.

Events are mutually exclusive if they cannot occur simultaneously, so P(A ∩ B) = 0. For mutually exclusive events:

Mutual exclusivity is different from independence. For example, an employee cannot simultaneously be classified as both “new hire” and “not a new hire”; these events are mutually exclusive. Two events are independent when the occurrence of one does not change the probability of the other.

1.3 Conditional probability and multiplication rules

Conditional probability measures the likelihood of A given that B has occurred:

Rearranging gives the general multiplication rule:

If A and B are independent, then P(A | B) = P(A) and:

For more than two events, the multiplication rule is applied sequentially. Independence must be justified rather than assumed merely because events concern different people or departments.

Worked Example 1: Recruitment screening

A company reports that 60% of applicants meet a minimum qualification, 40% have prior industry experience, and 25% have both. Let Q denote qualification and E denote experience.

The probability that an applicant has either qualification or experience is:

P(Q ∪ E) = 0.60 + 0.40 − 0.25 = 0.75

Thus, 75% meet at least one of the two conditions. The probability that an applicant has experience given qualification is:

P(E | Q) = 0.25 ÷ 0.60 ≈ 0.4167

Approximately 41.67% of qualified applicants have industry experience.

2. Counting Methods and Probability Trees

Counting methods are useful when outcomes are numerous. The factorial of a positive integer n is:

The number of arrangements of n distinct objects is n!. The number of ways of selecting r objects from n without regard to order is:

This combination formula applies when selection order does not matter.

A probability tree represents sequential events. Branch probabilities along a path are multiplied; probabilities of separate paths leading to the same result are added. Trees are especially useful for recruitment stages, promotion decisions and employee-retention sequences.

3. Bayes’ Theorem

Bayes’ theorem updates a probability when new evidence is observed:

For mutually exclusive and exhaustive events A₁, A₂, …, Aₖ:

P(Aᵢ) is the prior probability, P(B | Aᵢ) is the likelihood of evidence B under condition Aᵢ, and P(Aᵢ | B) is the posterior probability.

Worked Example 2: Interpreting a screening result

Suppose 10% of applicants would perform successfully after appointment. A screening tool identifies 80% of successful applicants, but it also gives a positive result to 20% of unsuccessful applicants. Let S mean success and + mean a positive screening result.

The probability of a positive result is:

P(+) = P(+|S)P(S) + P(+|Sᶜ)P(Sᶜ)

P(+) = (0.80)(0.10) + (0.20)(0.90) = 0.26

Therefore:

P(S|+) = (0.80)(0.10) ÷ 0.26 ≈ 0.3077

Only about 30.77% of applicants with a positive result are expected to succeed. The result illustrates why base rates matter and why a screening tool should not be treated as infallible.

4. Random Variables and Probability Distributions

A random variable assigns a numerical value to each outcome of a random experiment. A discrete random variable has countable values, such as the number of absences or hires. A continuous random variable can take any value in an interval, such as time to fill a vacancy or training expenditure per employee.

A probability mass function (PMF) gives probabilities for a discrete random variable X:

A valid PMF must satisfy p(x) ≥ 0 for every x and ∑ p(x) = 1.

A cumulative distribution function (CDF) is:

For a continuous variable, probabilities are represented by a probability density function f(x). The density itself is not a probability at a point; rather:

A continuous variable has P(X = x) = 0 for any single exact value. Probabilities concern intervals or ranges.

Expected value and variance

The expected value is the probability-weighted average:

For a continuous variable, the corresponding expression is E(X) = ∫ x f(x) dx. Expected value need not be an outcome that can actually occur.

Variance measures spread around the mean:

An equivalent computational form is:

The standard deviation is:

Variance is expressed in squared units, whereas standard deviation has the same units as X.

Worked Example 3: Number of monthly absences

Suppose X, the number of unscheduled absences for a randomly selected employee in a month, has the following distribution:

The expected number is:

E(X) = (0)(0.40)+(1)(0.30)+(2)(0.20)+(3)(0.10)=1.00

Also, E(X²) = 0²(0.40)+1²(0.30)+2²(0.20)+3²(0.10)=2.00. Hence:

Var(X)=2.00−1.00²=1.00

and σ=√1=1 absence. The mean is one absence and the standard deviation is one absence.

5. Important Discrete Distributions

5.1 Bernoulli distribution

A Bernoulli variable records one trial with two outcomes, conventionally success (1) and failure (0). If the probability of success is p, then:

Its mean is p and its variance is p(1 − p). Examples include whether a candidate accepts an offer or whether an employee completes training.

5.2 Binomial distribution

The binomial distribution counts successes in n Bernoulli trials. Its assumptions are:

• a fixed number n of trials;
• exactly two outcomes per trial;
• constant success probability p;
• independent trials.

The probability of exactly x successes is:

Its mean and variance are E(X)=np and Var(X)=np(1−p).

Worked Example 4: Training completion

If 70% of employees complete an online course and 8 employees are selected independently, the probability exactly 6 complete it is:

P(X=6)=C(8,6)(0.70)⁶(0.30)²

=28(0.117649)(0.09)≈0.2965

The probability is approximately 29.65%. The independence assumption may be questionable if employees influence one another or share the same manager.

5.3 Poisson distribution

The Poisson distribution models the number of events in a fixed interval when events occur independently at a constant average rate λ. Its probability function is:

Both the mean and variance equal λ. It may model calls to an HR help desk per hour or workplace incidents per month, provided the constant-rate and independence assumptions are reasonable.

Worked Example 5: HR service requests

If an HR service centre receives an average of 3 requests per hour, the probability of exactly 2 requests in an hour is:

P(X=2)=e⁻³3²÷2!≈0.2240

Thus, the probability is approximately 22.40%.

6. The Normal Distribution

The normal distribution is continuous, bell-shaped and symmetric about its mean μ. It is determined by μ and standard deviation σ, with σ > 0. Its density is:

The standard normal distribution has mean 0 and standard deviation 1. A value x is standardised using:

Then probability calculations can use the standard normal table or statistical software. The empirical rule states that approximately 68%, 95% and 99.7% of observations lie within 1, 2 and 3 standard deviations of the mean, respectively, when the normal model is appropriate.

Worked Example 6: Time to fill vacancies

Suppose time to fill a vacancy is approximately normal with mean 30 days and standard deviation 6 days. The probability that a vacancy is filled within 36 days is:

z=(36−30)÷6=1

For the standard normal distribution, P(Z≤1)≈0.8413. Therefore, approximately 84.13% of vacancies are expected to be filled within 36 days. This conclusion depends on the normal approximation and on the stability of the recruitment process.

The central limit theorem

The central limit theorem states that, under broad conditions, the distribution of the sample mean becomes approximately normal as sample size increases, even when individual observations are not normally distributed. If a population has mean μ and standard deviation σ, the sampling distribution of the mean has:

where n is the sample size. The observations should generally be independent, and a finite variance is required. For sampling without replacement, independence is often approximated when the sample is a small fraction of the population.

7. Applications in Human Resource Management

Probability distributions support:

• workforce planning: estimating staffing requirements and vacancy volumes;
• recruitment: assessing applicant success, offer acceptance and time to hire;
• attendance management: modelling absence counts and identifying unusual patterns;
• training: estimating completion rates and learning outcomes;
• retention: estimating turnover risk while recognising uncertainty;
• compensation: analysing variable-pay outcomes and payroll risk;
• employee relations: planning HR service capacity and case volumes.

A model should assist judgement, not replace it. Managers must distinguish a probability for a group from a deterministic label for an individual.

8. Illustrative Case: Workforce Planning for a Contact Centre

Background

A fictional customer-service centre has 40 agents. Historical records suggest that each agent has a 0.15 probability of being absent on a particular day. For a first approximation, absences are treated as independent and the probability is assumed constant.

Problem

Management wants to estimate the probability that more than 8 agents are absent and decide whether reserve staffing is justified.

Analysis

Let X be the number absent. Under the stated assumptions, X~Binomial(n=40,p=0.15). The expected number absent is:

E(X)=np=40(0.15)=6

The standard deviation is:

√[np(1−p)] = √[40(0.15)(0.85)] ≈ 2.258

Management may calculate P(X>8) by summing binomial probabilities for x = 9 through 40, using software or a reliable statistical table. A normal approximation may be considered because np = 6 and n(1 − p) = 34; however, the success expectation is relatively small, so the exact binomial calculation is preferable. If a normal approximation is used, a continuity correction is required: P(X>8)≈P(Y>8.5), where Y is normal with mean 6 and standard deviation approximately 2.258.

Approach and decision

The centre should compare the probability with the operational cost of reserve staff, overtime and service delays. It should also examine whether absences cluster by shift, season or supervisor. If absences are correlated, the binomial model understates risk.

Learning points

The case demonstrates that a probability model requires explicit assumptions, that expected value is not a worst-case forecast, and that operational decisions should combine statistical evidence with cost, fairness and contextual information.

9. Limitations, Ethics and Governance

Probability models can be misleading when data are biased, definitions change, observations are dependent or the process is non-stationary. Selection data may reflect unequal access to opportunities rather than genuine differences in potential. A model trained on historical promotion decisions can reproduce discriminatory patterns.

HR practitioners should:

• document variables, data sources and assumptions;
• test calibration and error rates across relevant groups;
• avoid using sensitive characteristics as unjustified predictors;
• protect confidentiality and limit access to employee data;
• provide human review and an avenue for explanation or appeal;
• monitor model performance after implementation;
• distinguish association from causation;
• comply with applicable employment, privacy and equality law.

AI systems frequently output probabilities, such as predicted turnover risk. A probability is not a fact about an individual. Transparency, validation, auditability and proportional use are essential. Automated recommendations should not become a hidden basis for adverse employment decisions.

Best-Practice Checklist

1. Define the random experiment and unit of analysis.
2. Identify the population, sample space and event precisely.
3. Select a distribution whose assumptions fit the process.
4. Check independence, constant rates and data quality.
5. Report assumptions alongside numerical results.
6. Communicate uncertainty rather than presenting false precision.
7. Validate results against new data and operational experience.
8. Review ethical, legal and fairness implications before action.

Chapter Summary

Probability measures uncertainty through axioms, complements, unions, intersections and conditional probabilities. Addition rules apply to alternative events, while multiplication rules apply to joint events. Independence is a substantive condition, not a default assumption. Bayes’ theorem updates prior beliefs using evidence.

Random variables may be discrete or continuous. Their distributions describe possible values and associated probabilities. Expected value measures the probability-weighted centre; variance and standard deviation measure dispersion. The Bernoulli, binomial, Poisson and normal distributions provide useful models for HR events, but each requires specific assumptions. The central limit theorem explains why sample means often become approximately normal.

In practice, probability models inform staffing, recruitment, attendance, training and service planning. Their responsible use requires careful modelling, validation, awareness of dependence and bias, and ethical governance of employee data and automated decisions.

Key Terms

Probability: Numerical measure of the likelihood of an event.

Random experiment: Process with an uncertain outcome.

Outcome: A possible result of an experiment.

Sample space: Set of all possible outcomes.

Event: Subset of a sample space.

Complement: Event that an event does not occur.

Union: Event that at least one of two events occurs.

Intersection: Event that two events occur together.

Mutually exclusive events: Events that cannot occur simultaneously.

Independent events: Events for which occurrence of one does not alter the probability of the other.

Conditional probability: Probability of an event given another event.

Bayes’ theorem: Rule for updating probability using new evidence.

Random variable: Numerical assignment to outcomes.

Probability mass function: Function giving probabilities for a discrete variable.

Probability density function: Function describing probabilities over intervals for a continuous variable.

Expected value: Probability-weighted mean of a random variable.

Variance: Mean squared deviation from the expected value.

Binomial distribution: Distribution of successes in a fixed number of independent trials.

Poisson distribution: Distribution of event counts in a fixed interval at an average rate.

Normal distribution: Continuous symmetric bell-shaped probability distribution.

Review Questions

Short Answer Questions

1. What is the difference between a sample space and an event?
2. State the complement rule and explain its meaning.
3. Distinguish mutually exclusive events from independent events.
4. What does the expected value of a random variable represent?
5. Identify two assumptions of the binomial distribution.

Long Answer Questions

1. Explain the addition and multiplication rules of probability, including their special cases.
2. Describe Bayes’ theorem and discuss its relevance to HR screening decisions.
3. Compare discrete and continuous random variables, including their probability functions.
4. Explain the assumptions and applications of the binomial, Poisson and normal distributions.
5. Discuss the ethical and governance risks of using probability-based AI systems in HR.

Analytical / Application Questions

1. A training programme has an 80% completion rate. For 10 independent employees, calculate the probability that exactly 8 complete the programme and interpret the result.
2. A help desk receives an average of 4 HR queries per hour. Use a Poisson model to calculate the probability of exactly 3 queries in one hour.
3. Recruitment time is normally distributed with mean 24 days and standard deviation 5 days. Calculate the z-score for 34 days and explain what the result indicates.

Multiple Choice Questions with Answers

Question 1

If P(A) = 0.35, what is P(Aᶜ)?

A. 0.35

B. 0.65

C. 1.35

D. 0.05

Correct Answer: B. 0.65

Question 2

Which condition defines mutually exclusive events?

A. They always have equal probabilities

B. Their probabilities must both be zero

C. They cannot occur simultaneously

D. Occurrence of one causes the other

Correct Answer: C. They cannot occur simultaneously

Question 3

If P(A)=0.5 and P(B)=0.4 and the events are independent, what is P(A ∩ B)?

A. 0.10

B. 0.20

C. 0.45

D. 0.90

Correct Answer: B. 0.20

Question 4

In Bayes’ theorem, the posterior probability is the probability of:

A. The evidence before observing data

B. The condition after observing evidence

C. Two unrelated events together

D. The complement of the sample space

Correct Answer: B. The condition after observing evidence

Question 5

A binomial distribution requires:

A. A variable with infinitely many categories

B. Dependent trials with changing probabilities

C. A fixed number of independent trials with two outcomes

D. A continuous measurement only

Correct Answer: C. A fixed number of independent trials with two outcomes

Question 6

If X has a Poisson distribution with λ = 5, what are its mean and variance?

A. Mean 5 and variance 25

B. Mean 25 and variance 5

C. Mean 5 and variance 5

D. Mean 0 and variance 5

Correct Answer: C. Mean 5 and variance 5

Question 7

A normal variable has mean 50 and standard deviation 10. What is the z-score for x = 70?

A. −2

B. −0.5

C. 2

D. 7

Correct Answer: C. 2

Question 8

Which statement about a continuous random variable is correct?

A. P(X=x) is always 1

B. P(X=x) is normally 0 for an exact point

C. Its probabilities must sum over countable points

D. It can take only integer values

Correct Answer: B. P(X=x) is normally 0 for an exact point

Question 9

If a random variable has E(X)=4 and E(X²)=25, what is Var(X)?

A. 9

B. 16

C. 21

D. 29

Correct Answer: A. 9

Question 10

The standard deviation of the number of successes in a binomial distribution is:

A. np

B. np(1−p)

C. √[np(1−p)]

D. n+p

Correct Answer: C. √[np(1−p)]