STAT110
Chapter 17: Thinking About Chance
In everyday life, chance is often conveyed through probabilities, for instance, "there is a 10% chance of rain tomorrow."
The National Weather Service indicates that one has a 1 in 15,300 chance of being struck by lightning over a lifetime.
17.1 The Idea of Probability
Coin Tossing in Soccer:
Each team tosses a coin before the game to decide which goal to attack (FIFA rules).
The probability of getting "heads" from a fair coin is 0.5, indicating that if tossed multiple times, outcomes will trend towards 0.5.
The probability isn't merely based on there being two sides but on underlying data patterns (Buf- fon, Pearson, Kerrich).
Page 1
Long-run proportion of heads in repeated trials will tend to 0.5.
Page 2: Randomness and Probability
Coin tossing outcomes are binary: heads or tails.
Experiment outcomes:
After 5 tosses: 3 heads; proportion = 3/5 = 0.6.
After 100 tosses: 49 heads; proportion = 49/100 = 0.49.
After 10,000 tosses: heads proportion = 0.4994, very close to probability 0.5.
Proportion is highly variable initially but stabilizes around 0.5 with more tosses.
This stabilization is described by the Law of Large Numbers.
Page 3: Law of Large Numbers
States that repeated probability experiments will have outcomes that get closer to their expected probability.
Random phenomena have unpredictable individual outcomes, but exhibit predictable long-term distributions.
Probability value ranges from 0 to 1, where 0 means never occurs and 1 means always occurs.
17.2.1 Random Phenomenon
Examples include:
Coin flipping (2 outcomes: H, T).
Rolling a die (6 outcomes: 1, 2, 3, 4, 5, 6).
Predicting stock market indices.
Page 4: More Ideas on Probability
Outcomes with probability 0 never occur; probability 1 outcomes always occur.
Probability serves as language for long-term random behavior.
17.2.3 Probability Myths
Myth #1: Short-run Regularity
Misconception that randomness must show regular patterns in short runs.
Example: Tossing a coin does not guarantee heads in any defined pattern.
Myth #2: Surprising Coincidence
Unlikely coincidences may occur frequently in large sample spaces.
Myth #3: Law of Averages
Misbelief that previous results must influence future outcomes (gambling example).
Page 5: Probability and Risk
Risks associated with various probabilities, such as:
Asbestos exposure - low risk for teachers, significant compared to car accidents.
Statistically, individuals face higher probability of death from heart disease vs. terrorism.
Psychological factors influence perception of risk based on control over events.
Pages 6-16: Probability Models
Basic Concepts
Probability models consist of a sample space and assigned probabilities.
Probability rules dictate:
Probabilities must total 1.
The probability of not occurring equals 1 minus the probability of occurring.
Disjoint Events
Disjoint/non-overlapping events: Their probabilities collectively summed (e.g., never married vs. married).
Intersection of Events
Events can occur simultaneously or not based on specific conditions.
Independent Events
Occurrence of one does not affect the other.
Complement of an Event
The complement includes all outcomes leaving out the specific event.
Chapter 18: Use of Probability Models
Understanding probability models requires consideration of random variables and events.
Involves estimating probabilities through sampling distributions and interpreting results effectively.
Page 17: Simulation
Simulation models chance behaviors to estimate outcomes for complex probabilities.
Important for practical investigations where theory fails or is complicated to compute.
Tree Diagrams
Useful for visually representing probability events. Each path/branch signifies a possible outcome, aiding in calculation.
Page 22-39: Expected Values and Statistical Significance Tests
Expected values reflect long-run averages of outcomes, calculated based on probabilities assigned to events.
Hypothesis testing provides a framework for inferring population parameters based on sample data.
Significant tests assess probability against null hypothesis, determining whether observed outcomes deviate enough to reject it.
Practical Application of Statistical Inference
Evaluation consistency in data interpretation through statistical significance, identifying Type I and Type II errors.
Understand the balance between statistical and practical significance, ensuring that results are both meaningful and reliable.