Week 5 — Lecture Outline · Probability Foundations

Course: Introduction to Statistics (MATH 11) · Silver Oak University (fictional sample) · Prof. Rivera
Objectives covered: Objective 4 — Apply basic probability rules, including conditional probability.
SLOs touched: A (reason quantitatively from data) · B (communicate results to a non-technical audience)
Meeting pattern: 2 sessions × 75 min = 150 min. Segment minutes below total ~150; scale to your own pattern.

---

Week at a Glance

The week's big question	"When the outcome isn't settled yet, how do we put an honest number on what's likely — and how does new information change that number?"
By the end of the week, students can…	(1) write a sample space, identify an event, and compute a basic probability as favorable ÷ total; (2) apply the complement rule (P(not A) = 1 − P(A)) and the addition rule — knowing when to subtract the overlap and when two events are mutually exclusive; (3) apply the multiplication rule and decide whether two events are independent; (4) compute a conditional probability P(A | B) from a table or a story, and explain in plain words what "given" changes.
Key vocabulary	random experiment, outcome, sample space (S), event, probability, equally likely outcomes, complement (Aᶜ / not A), addition rule, mutually exclusive (disjoint), multiplication rule, independent / dependent events, conditional probability P(A|B), "given," joint probability, two-way (contingency) table, marginal vs. joint probability, base rate
Materials	slides (Deck 5), the week's readings + video links, a spreadsheet (Google Sheets or Excel), one approved chatbot (Gemini / Claude / ChatGPT) for the AI-critique moment and the tutorial
Timing note	8 segments, ~150 min total. Session 1 = Segments 1–4 (~75). Session 2 = Segments 5–8 (~75).

---

Segment 1 — Hook & the Promise (8 min) · Session 1 opens

Hook. Put one sentence on the board: "You are 'due' for a win." Let it sit. Then: "A roulette wheel just landed on black eight times in a row. A gambler bets big on red — 'it has to come now.' How much should red's chance have changed?" Wait for guesses, then drop the answer: not at all. The wheel has no memory; red is still 18/38, exactly what it was on spin one.
- "That feeling — that the world owes you a balance — has a name, the gambler's fallacy, and it has emptied real bank accounts. This week we replace the feeling of likely with a number for likely, and we learn the rules that number obeys."
- Second jab: "Your weather app says 30% chance of rain. What's the chance it doesn't rain? If '70%' jumped to mind instantly — congratulations, you already used the first rule we'll name today."

The promise (write it on the board): "By the end of this week you can take any uncertain situation — a card draw, a campus survey, a medical test result — and compute an honest probability, combine probabilities with the right rule, and update a probability the moment new information arrives."

Why it matters line (memory hook): "Probability isn't about predicting the future. It's about being honest, in a single number, about what you don't yet know."

---

Segment 2 — Sample Spaces, Events & Basic Probability (22 min)

Plain language first. A random experiment is any action whose outcome isn't settled in advance — flip a coin, roll a die, draw a card, pick a random student. The sample space, written S, is the complete list of every possible outcome. An event is any part of that list we care about — a subset of S.
- When every outcome is equally likely, probability is just bookkeeping:

P(event) = (number of outcomes in the event) ÷ (number of outcomes in S) — favorable over total.
- Two iron rules every probability obeys: each probability sits between 0 and 1 (0 = impossible, 1 = certain), and the probabilities of all outcomes in S add up to exactly 1.

Memory hook (put it on a slide):

List the whole space first. Then count what you want over what's possible. "Favorable over total."

One fully worked example (do every step out loud).

Experiment: flip a fair coin twice.
- Sample space S = { HH, HT, TH, TT } — list it before computing anything. That's 4 equally likely outcomes.
- Event A = "exactly one head." Which outcomes qualify? HT and TH. That's 2 outcomes.
- P(A) = 2 ÷ 4 = 0.5 = 50%.
- Event B = "at least one head." Outcomes: HH, HT, TH — 3 of them. P(B) = 3 ÷ 4 = 0.75 = 75%.
- Notation preview (notation comes after the idea): we write the sample space S, an event with a capital letter like A, and its probability P(A).

Land the key idea: the hardest part of a basic probability problem is writing the sample space correctly — once the list is right, the answer is division. Slow down on the list; the arithmetic is the easy part.

---

Segment 3 — Two Rules: Complement & Addition (25 min)

Plain language first — the complement (the lazy genius move).
- The complement of event A — written Aᶜ or "not A" — is everything in S that is NOT A. Since A and "not A" together make up the whole sample space (which sums to 1):

P(not A) = 1 − P(A).
- Why it matters: sometimes the thing you want is a nightmare to count directly, but its opposite is easy. "At least one" problems are the classic case — count the none and subtract.

One worked example (complement):

The weather app says P(rain tomorrow) = 0.30. Then P(no rain) = 1 − 0.30 = 0.70. Done — no listing required.
Harder one where it earns its keep: roll a die twice; P(at least one 6)? Counting "at least one" directly is fiddly. Flip it: P(no 6 either roll) = (5/6) × (5/6) = 25/36, so P(at least one 6) = 1 − 25/36 = 11/36 ≈ 0.31. "At least one" → 1 minus "none."

Plain language — the addition rule (the "OR" rule).
- For P(A or B) — the chance that A happens, or B happens, or both — you add, but you must not double-count the overlap:

P(A or B) = P(A) + P(B) − P(A and B).
- If A and B can't happen at the same time, they're mutually exclusive (disjoint) — the overlap is 0 — and the rule collapses to the easy version:
P(A or B) = P(A) + P(B) (only when mutually exclusive).

One worked example (addition, both versions) — draw one card from a 52-card deck:

• P(King or Queen)? A King and a Queen can't be the same card → mutually exclusive. P = 4/52 + 4/52 = 8/52 = 2/13 ≈ 0.154.
• P(King or Heart)? A card can be both (the King of Hearts) → NOT mutually exclusive. Subtract the overlap: P = 4/52 + 13/52 − 1/52 = 16/52 = 4/13 ≈ 0.308. Forget to subtract and you'd wrongly get 17/52 — you'd have counted the King of Hearts twice.

Memory hook: "OR means add — but subtract the overlap, unless the two events can't coexist."

---

Segment 4 — Misconceptions + Quick Interaction (20 min) · Session 1 closes (~75)

Name the misconceptions out loud, then cure each:

• ❌ "After a streak, the other outcome is 'due.'" (The gambler's fallacy.)
  ✅ Cure: independent trials have no memory. A fair coin that landed heads five times is still 50/50 on the sixth — the coin can't remember and doesn't owe you. Past independent results never change the next probability.
• ❌ "For 'A or B,' just add the two probabilities."
  ✅ Cure: that's only true when A and B are mutually exclusive. If they can overlap, adding double-counts the overlap — King-or-Heart, not 17/52 but 16/52. Ask first: can both happen at once?
• ❌ "Mutually exclusive and independent mean the same thing."
  ✅ Cure: opposite ideas. Mutually exclusive = they can't both happen (one card can't be both a King and a Queen). Independent = one happening doesn't change the odds of the other (two separate coin flips). In fact, two events that can occur are usually one or the other, not both.
• ❌ "A probability can be 1.4, or 140%."
  ✅ Cure: every probability lives between 0 and 1 (0% to 100%). A number outside that range is a signal you mis-added — most often by forgetting to subtract an overlap.

Interaction — Think-Pair-Share (rapid-fire classification, ~10 min):
Put 6 pairs of events on a slide; for each, students decide solo (30 sec) whether the two events are mutually exclusive (can't both happen) or not, compare with a neighbor (1 min), then the class votes by thumbs. Suggested items:
drawing one card: "red card" vs "King" · one die roll: "even" vs "5" · today's weather: "rains" vs "sunny all day" · one student: "is a senior" vs "owns a car" · one coin flip: "heads" vs "tails" · one card: "a Heart" vs "a face card."
(Answers: not exclusive — King of Hearts/Diamonds exist · exclusive — 5 is odd · exclusive · not exclusive · exclusive · not exclusive — Jack/Queen/King of Hearts.)
Debrief the two that split the room: "red vs King" (not exclusive — two red Kings) and "senior vs owns a car" (not exclusive — plenty of car-owning seniors).

---

Segment 5 — Multiplication & Independence (25 min) · Session 2 opens

Hook back in: "Last session we combined chances with OR — adding. Today we combine them with AND — multiplying — and meet the word that decides whether multiplying is even legal: independent."

Plain language first — the multiplication rule (the "AND" rule).
- P(A and B) is the chance that both happen. When A and B are independent — one happening doesn't change the other's odds — you simply multiply:

P(A and B) = P(A) × P(B) (only when independent).
- Independent = knowing one outcome tells you nothing about the other. Separate coin flips, separate dice, a card drawn with replacement. Dependent = one outcome shifts the other's odds (drawing a second card without putting the first back changes what's left).

One worked example (independence):

Roll a fair die twice. P(two 6s)? The rolls are independent — the first die can't influence the second. P = (1/6) × (1/6) = 1/36 ≈ 0.028.
Coin-and-spinner: P(heads) = 1/2, P(spinner lands red) = 1/3, and they're independent. P(heads and red) = (1/2) × (1/3) = 1/6 ≈ 0.167.

The independence test (this is the bridge to next segment): two events A and B are independent exactly when knowing B doesn't change A's probability — i.e., P(A | B) = P(A). If the "given" version differs from the plain version, they're dependent. (We compute P(A | B) next; hold that thought.)

Memory hook: "AND means multiply — but only when the events are independent. If one nudges the other, you can't just multiply."

---

Segment 6 — Conditional Probability: When New Information Arrives (22 min)

Plain language first. Conditional probability, written P(A | B) and read "the probability of A given B," answers: now that I know B happened, what's the chance of A? The word "given" is the whole game — it shrinks the world down to just the cases where B is true, and asks how often A happens inside that smaller world.

P(A | B) = P(A and B) ÷ P(B) — of the times B happened, how often did A also happen?

The two-way table is conditional probability's home (worked example — do every step).

A campus survey of 200 students records where they live and whether they own a car:

	Owns a car	No car	Total
Lives on campus	30	70	100
Lives off campus	80	20	100
Total	110	90	200

• Marginal (plain) probability — use the totals. P(owns a car) = 110 ÷ 200 = 0.55.
• Conditional — the "given" shrinks the world to one row. P(owns a car | lives off campus) = look only at the off-campus row (100 students): 80 of them own a car → 80 ÷ 100 = 0.80.
• Flip the condition and the answer changes: P(lives off campus | owns a car) = look only at the car column (110 owners): 80 live off campus → 80 ÷ 110 ≈ 0.727. P(A|B) and P(B|A) are different questions — don't swap them.

Land the key idea: "given" means cover up the rest of the table and work inside one row (or one column). The denominator stops being the grand total (200) and becomes the subgroup total (100, or 110).

Misconception + cure:
- ❌ "P(A | B) is the same as P(B | A)." (The "confusion of the inverse.")
✅ Cure: they answer different questions and usually differ. P(car | off-campus) = 0.80 but P(off-campus | car) ≈ 0.727. The thing after the bar is the world you're standing in — switch it and you've changed the question.

---

Segment 7 — Independence, Revisited + the Base-Rate Trap (18 min)

Tie the room together — the independence check, now that we can compute conditionals.

Back to the survey. Is owning a car independent of living off campus? Compare the plain probability to the conditional:
- P(owns a car) = 0.55 (from the totals).
- P(owns a car | off campus) = 0.80 (from the row).
- They're not equal → the events are dependent. Living off campus raises the chance of owning a car, which makes real-world sense. Independence is a number check, not a vibe: does the condition move the probability?

The base-rate trap (the example that sticks — a real misuse of conditional probability):

A screening test for a rare disease is "99% accurate." You test positive. Panic? Let's count honestly with natural frequencies instead of percentages. Imagine 1,000 people, where the disease is rare — 1% have it, so 10 are truly sick and 990 are healthy.
- Of the 10 sick, the test catches essentially all → about 10 test positive (true positives).
- Of the 990 healthy, even a good test fires falsely ~5% of the time → about 50 test positive (false positives).
- So roughly 60 people test positive, but only 10 are actually sick.
- P(disease | positive test) = 10 ÷ 60 ≈ 0.17 — about 17%, not 99%. The test is accurate; the positive is still probably a false alarm, because the base rate (how rare the disease is) dominates.
Lesson: P(positive | sick) is NOT P(sick | positive). Ignoring the base rate is one of the most expensive conditional-probability mistakes in medicine, law, and the news.

Callback: this is the inverse-confusion misconception from Segment 6, now with real stakes. Point back to it explicitly: "after-the-bar matters — and the rarer the thing, the more it matters."

---

Segment 8 — Technology Workflow + AI-Critique, Callback & Hand-off (10 min) · Session 2 closes (~75)

Technology workflow — simulate probability in a spreadsheet (exact steps):
1. We claimed P(at least one 6 in two rolls) ≈ 0.31. Let's check it by experiment. In A2 type =RANDBETWEEN(1,6) (roll one die) and in B2 =RANDBETWEEN(1,6) (roll the second).
2. In C2 type =IF(OR(A2=6,B2=6),1,0) — a 1 if at least one die shows a 6, else 0.
3. Select A2:C2 and fill down to row 1001 (1,000 simulated trials).
4. In E1 type =AVERAGE(C2:C1001) — the fraction of trials with at least one 6.
- Google Sheets and Excel are identical here. Re-calculate (Ctrl/Cmd-R or edit any cell) to reshuffle. You'll land near 0.31 — not exactly, because it's a simulation. The math gives the truth; the simulation circles it.

AI-critique moment (students verify, not consume):

Paste this to an approved chatbot: "In a 52-card deck, what is the probability of drawing a King OR a Heart?" and "A test is 99% accurate for a disease that 1% of people have. I tested positive — what's the chance I actually have it?"
Then check its work. Chatbots often forget to subtract the overlap on King-or-Heart (giving 17/52 instead of 16/52), and many fall for the base-rate trap, answering "99%" instead of ~17%. Your job all semester: the tool drafts, you judge. This is exactly how the weekly Lecture Tutorial works — you'll catch the model, not trust it.

Callback + tease:
- Callback: "Week 4 we asked whether two variables move together in a two-way table; this week that same table gave us conditional probability. The table never changed — our question got sharper."
- Tease next week: "Now that we can put a number on a single uncertain outcome, Week 6 turns the outcomes themselves into numbers — random variables, expected value, and what you'd win on average if you played a game a thousand times."

Hand-off (the week's graded work):
- Lecture Tutorial 5 (AI tutor, share-link submission) — sample spaces, complement & addition, multiplication & independence, conditional probability.
- Quiz 5, Discussion 5 ("Due for a win?" — reason about a probability misconception with your chatbot), and Assignment 5 (four coached problems) — all due end of week.

---

Instructor FAQ — Common Stumbles

Student says / does	Quick cure
"It landed on red five times, so black is due."	Independent trials have no memory. Each spin is the same probability as the first; the gambler's fallacy is believing otherwise. Past results don't pay future debts.
Adds for "A or B" and forgets the overlap.	Ask: can both happen at once? If yes (King or Heart), subtract P(A and B). If no (King or Queen), plain addition is fine. A probability over 1 is the telltale you double-counted.
Confuses mutually exclusive with independent.	Mutually exclusive = they can't both occur (one card, King vs Queen). Independent = one doesn't change the other's odds (two separate flips). Different questions entirely.
Multiplies P(A)×P(B) when events are dependent.	Plain multiplication needs independence. Cards drawn without replacement are dependent — the second draw's odds shifted. When in doubt, use P(A and B) = P(B) × P(A|B).
Swaps P(A|B) and P(B|A).	The thing after the bar is the world you're inside. P(car | off-campus)=0.80 ≠ P(off-campus | car)≈0.727. Switching the condition changes the question.
On a two-way table, divides by the grand total for a conditional.	"Given" shrinks the world to one row or column. The denominator becomes that subgroup total, not 200. Cover up the rest of the table.
Hears "99% accurate test" and assumes a positive means 99% sick.	That's P(positive | sick), not P(sick | positive). With a rare disease the base rate dominates — natural frequencies show ~17%. Always ask how common the thing is to begin with.
Writes a probability like 1.3 or −0.2.	Impossible. Every probability is between 0 and 1. Re-check your arithmetic — you likely forgot to subtract an overlap or mis-listed the sample space.

---

Scope flag

This outline stays within Objective 4 (basic probability rules + conditional probability). The base-rate / screening-test example and the gambler's-fallacy roulette framing are added context — not strictly required by the objective — kept because they make the conditional-probability and independence ideas stick and because they show probability used in the wild. Trim them for a leaner 60-minute version. We deliberately keep no combinatorics-heavy counting (no permutations/combinations), per the course's depth decisions — every sample space here is small enough to list.