Comprehensive Guide to Probability: Empirical vs. Theoretical, Addition and Multiplication Rules, and Conditional Probability

Foundations of Probability: Empirical vs. Theoretical

• Pareidolia is a psychological phenomenon driven by the brain's pattern-recognition system, causing individuals to perceive familiar patterns—specifically faces—in inanimate, non-face objects such as onions, grilled cheese sandwiches, or bell peppers.

• Human cognition recognizes patterns across diverse domains, particularly within sequences of events, forming the perceptual backdrop for the study of probability.

• Probability has distinct statistical definitions that differ from everyday colloquial usage, primarily divided into two classifications:

  • Empirical Probability:

  • Defined as probability calculated from actual, observed data (such as calculating the proportion of female children born within individual families).

  • Inherently contains uncertainty because it is derived from samples that represent only a fraction of all possible data.

  • Acts analogously to sample statistics, providing an estimate or glimpse of the true theoretical probability without necessarily matching it exactly due to sampling error and random variability.

  • Example (Slot Machine): Playing a slot machine repeatedly allows an observer to estimate the probability of winning the jackpot by dividing the total number of wins by the total number of attempts. For instance, hitting the jackpot $6$ times in $100$ plays produces an empirical probability of $\frac{6}{100}$ or $6\%$. While this empirical result does not rule out true underlying values such as $5\%$ or $10\%$, it provides high confidence that the probability is near $6\%$ rather than an extreme value such as $99\%$.

  • Theoretical Probability:

  • Defined as the underlying, ideal truth residing within the universe that cannot be observed directly.

  • Functions analogously to true population parameters (such as the population mean or standard deviation), which are inferred through sample data.

The Addition Rule and Mutually Exclusive Events

• Probability Notation:

  • The probability of an event occurs is written as $P(\text{Event})$.

  • The probability of Event A or Event B occurring is denoted as $P(\text{A or B})$.

  • The probability of both Event A and Event B occurring together is denoted as $P(\text{A and B})$.

• Mutually Exclusive Events:

  • Two events are mutually exclusive if they cannot occur at the same time. For mutually exclusive events, $P(\text{A and B}) = 0$.

  • Example (Bag of Skittles):

  • A standard bag contains $5$ distinct colors, each comprising an approximately equal proportion of $20\%$ ($0.2$).

  • Because a single randomly drawn Skittle cannot be two colors simultaneously, color selection is mutually exclusive: $P(\text{Red and Purple}) = 0$.

• Simplified Addition Rule (for Mutually Exclusive Events):

  • Formula: $P(\text{A or B}) = P(\text{A}) + P(\text{B})$

  • Applying this to Skittles: $P(\text{Red or Purple}) = P(\text{Red}) + P(\text{Purple}) = 0.2 + 0.2 = 0.40$ ($40\%$).

  • If an individual likes all $5$ colors equally, the probability of drawing an acceptable Skittle is the sum of all individual color probabilities: $0.2 + 0.2 + 0.2 + 0.2 + 0.2 = 1.00$ ($100\%$).

• Non-Mutually Exclusive Events:

  • Events are non-mutually exclusive when both outcomes can happen simultaneously during the same trial, meaning $P(\text{A and B}) \neq 0$.

• Example (Coin Flip and Die Roll):

  • Tossing a coin and rolling a standard six-sided die yields $12$ total possible combined outcomes.

  • Flipping a tails and rolling a $6$ can occur on the same turn, which means $P(\text{tails or } 6) \neq 0$.

  • Across the sample space, there are $6$ outcomes containing tails and $2$ outcomes containing a $6$.

  • Simply summing these outcomes yields $6 + 2 = 8$, but an examination of the sample space reveals only $7$ unique outcomes that feature either a tails, a $6$, or both.

  • The single overlapping outcome (tails and a $6$) is double-counted if the sets are summed independently.

• Full Addition Rule (General Addition Rule):

  • Formula: $P(\text{A or B}) = P(\text{A}) + P(\text{B}) - P(\text{A and B})$

  • Die and Coin calculation: $P(\text{tails or } 6) = P(\text{tails}) + P(6) - P(\text{tails and } 6)$.

  • Visualized with a Venn diagram, one circle represents the instances of flipping tails and the second represents rolling a $6$. The overlapping intersection represents outcomes containing both, which must be subtracted once to eliminate double counting.

The Multiplication Rule and Independent Events

• The Multiplication Rule is utilized to calculate the probability of two or more events occurring simultaneously.

• Independent Events:

  • Two events are independent if the occurrence of one event does not alter the probability of the other event occurring.

  • Formally, if events are independent, knowing that Event 2 occurred gives no new information about Event 1.

• Hypothetical Dining Scenario (Cole Sprouse and IHOP):

  • Assume actor Cole Sprouse visits a local IHOP for dinner on any given night with a probability of $20\%$ ($0.2$).

  • The restaurant runs a promotion where nights are randomly selected as "Free Ice Cream Night" with a probability of $10\%$ ($0.1$).

  • Cole Sprouse has no prior knowledge of the promotion, meaning his attendance is independent of Free Ice Cream Night.

• Calculating Joint Probability of Independent Events:

  • To find the probability that Cole Sprouse is present and it is Free Ice Cream Night, multiply their individual probabilities: $P(\text{Cole and Free Ice Cream}) = P(\text{Cole}) \times P(\text{Free Ice Cream}) = 0.2 \times 0.1 = 0.02$ ($2\%$).

  • Conceptual basis: Cole visits on $20\%$ of all nights. Within that subset of Cole nights, only $10\%$ feature free ice cream ($10\%$ of $20\%$), yielding $2\%$.

• Calculating "Either/Or" for Independent Events:

  • Using the full addition rule, the probability that a patron either sees Cole Sprouse or receives free ice cream (or both) is: $P(\text{Cole or Free Ice Cream}) = P(\text{Cole}) + P(\text{Free Ice Cream}) - P(\text{Cole and Free Ice Cream}) = 0.2 + 0.1 - 0.02 = 0.28$ ($28\%$).

Conditional Probability and Medical Screening Applications

• Conditional Probability Definition and Notation:

  • Denoted as $P(\text{Event 1} \mid \text{Event 2})$, read as "the probability of Event 1 given that Event 2 has occurred."

  • If two events are independent, conditional probability collapses to marginal probability: $P(\text{Cole} \mid \text{Ice Cream Night}) = P(\text{Cole})$.

  • Formula for calculating conditional probability: $P(\text{Event 2} \mid \text{Event 1}) = \frac{P(\text{Event 1 and Event 2})}{P(\text{Event 1})}$

• Non-Reciprocal Nature of Conditional Probability:

  • Conditional probabilities are not symmetrical: $P(\text{A} \mid \text{B}) \neq P(\text{B} \mid \text{A})$.

• Medical Screening Context (Cervical Cancer Screening):

  • Historic medical standards previously recommended that all adult women receive screening for cervical cancer once per year.

  • Diagnostic classifications:

  • False Positive: The screening indicates an abnormality when no cancer is present.

  • False Negative: The screening returns an all-clear or normal result when cancer is actually present.

  • Statistical baseline parameters in the United States:

  • Baseline disease rate (prevalence): approximately $0.0081\%$ (around $8$ in every $100{,}000$ women).

  • False positive rate (among patients without cancer): $3\%$.

  • False negative rate (among patients with cancer): $46\%$.

  • Diagnostic Breakdown per $100{,}000$ Women:

  • Women with cancer ($8$ total): Given a $46\%$ false negative rate, the true positive rate is $54\%$ ($100\% - 46\%$). Therefore, approximately $4$ women ($8 \times 0.54 = 4.32$) receive a positive test.

  • Women without cancer ($99{,}992$ total): Given a $3\%$ false positive rate, approximately $3{,}000$ women ($99{,}992 \times 0.03 \approx 3{,}000$) receive a positive test.

  • Total positive tests in the population: approximately $3{,}004$.

  • Actual cancer cases among positive tests: approximately $4$.

  • Posterior Conditional Probability:

  • The probability of actually having cancer given a positive test result is: $P(\text{Cancer} \mid \text{Positive Test}) = \frac{4}{3{,}004} \approx 0.1\%$.

  • In contrast, the reciprocal probability—the probability of testing positive given that an individual has cancer—is $P(\text{Positive Test} \mid \text{Cancer}) \approx 50\%$ (specifically around $54\%$).

  • Healthcare Policy Implications:

  • Because a positive test carries only a $0.1\%$ probability of true abnormality while triggering expensive, invasive follow-up procedures and psychological distress, medical organizations updated guidelines to recommend less frequent routine screenings.

Practical Decision-Making and Real-World Probability

• Limitations of Real-World Quantification:

  • Many everyday situations cannot be cleanly assigned exact numerical probabilities due to human unpredictability (e.g., predicting whether Cole Sprouse visits IHOP, whether a teacher will call in sick, or whether one will catch all red lights during a morning commute).

• Probability as a Heuristic for Setting Expectations:

  • Weekend Planning: Establishing contingency plans avoids poor outcomes (such as unintentionally staying home to watch the film Sandy Wexler).

  • Movie Ticket Purchases: When attempting to watch Black Panther in a theater, requiring center-row seating drastically lowers the probability of success compared to being willing to accept any available seat; alternative backup options include streaming films like Get Out.

  • Higher Education Applications: Applying to multiple colleges leverages the addition rule of probability to raise the overall probability of receiving at least one admission offer.

  • Creative Expectations: Probability contextualizes long odds, explaining why authors should not anticipate that their very first short story will simultaneously receive an 'A' grade and be published in The New Yorker.

  • Epidemiological Risk: Conditional probability models disease transmission risk in personal interactions, such as assessing the conditional probability of contracting mononucleosis given that a significant other has mononucleosis.