← Back to the Introduction to Psychology outline The Course Maker
← Introduction to Psychology outline
Week 2 · Lecture outline

Week 2 — Lecture Outline · Research Methods & Ethics

Introduction to Psychology · PSYC 1 Fall 2026 · Prof. Bennett Fictional sample

Course: Introduction to Psychology (PSYC 1) · Silver Oak University (fictional sample) · Prof. Bennett
Objectives covered: Objective 2 — Evaluate psychological research methods and ethics, distinguishing correlation from causation and identifying sources of bias.
SLOs touched: A (apply concepts to real-world behavior) · B (reason scientifically about claims regarding mind and behavior)
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 "How do psychologists actually know what they claim — and why does a link between two things almost never prove that one causes the other?"
By the end of the week, students can… (1) walk the scientific method from theory → testable hypothesis → operational definition → data → replication; (2) tell apart the three research designs — descriptive, correlational, experimental — and say what each can and can't claim; (3) explain why correlation ≠ causation (the third-variable problem and directionality) and why only an experiment with random assignment supports a cause; (4) distinguish random sampling (generalizability) from random assignment (causation); (5) name the core research ethics — informed consent, right to withdraw, protection from harm, justified deception + debriefing, confidentiality, and IRB review.
Key vocabulary scientific method, theory, hypothesis, operational definition, empiricism, replication; descriptive research (case study, naturalistic observation, survey), correlational research, correlation coefficient (−1 to +1), positive/negative correlation, third-variable problem, directionality problem; experiment, independent variable (IV), dependent variable (DV), experimental & control groups, random assignment, confounding variable, placebo effect, blinding; population, sample, representative sample, random sampling, sampling bias; reliability, validity; informed consent, right to withdraw, protection from harm, deception, debriefing, confidentiality, IRB
Materials slides (Deck 2), the week's readings + video link, 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 headline on a slide and ask the room to react fast: "Studies show: students who use a study app earn higher grades." Hands up — does the app work? Most rooms lean "yes." Then the twist: "What if the students who choose the app are just the more motivated ones to begin with? Then the app didn't raise the grades — motivation did, and the app just came along for the ride."

Follow with a second one that's obviously silly: "As ice-cream sales go up, so do drowning deaths." Nobody thinks ice cream drowns people — so what's the real driver? (Hot summer weather: more swimming and more ice cream.) "Same logic, less obvious — that's the trap."

The promise (write it on the board): "By Friday you'll be able to look at any 'studies show' claim and say exactly what kind of study it was, what it can honestly conclude, and whether anyone is allowed to say one thing caused another."

Why it matters line (memory hook): "Last week we said we trust evidence over our gut. This week is the fine print — because not all evidence is created equal."

Segment 2 — The Scientific Method, in Plain Steps (18 min)

Plain language first. Science isn't a place or a lab coat — it's a loop for turning a hunch into something the world can talk back to.

The loop (put it on a slide, one line each):
- Theory — a well-supported explanation that organizes many findings (last week's word). It's the big idea.
- Hypothesis — one specific, testable prediction the theory makes, stated so it could turn out false. "If sleep strengthens memory, then students who sleep 8 hours after studying will recall more words tomorrow than students kept awake."
- Operational definition — exactly how you'll measure each fuzzy word, in numbers anyone could copy. "Memory" becomes "number of words recalled from a 20-item list after 24 hours"; "stress" becomes "score on a 1–7 self-report scale" or "resting heart rate in beats per minute."
- Collect data — run the study and record what actually happens.
- Replicate — do it again, and let other researchers do it again. A result you can't reproduce isn't yet knowledge.

Memory hook (put it on a slide):

"Empiricism in five moves: theory → hypothesis → operational definition → data → replicate. Intuition starts the question; evidence settles it."

Why operational definitions matter (say it out loud): "If I claim 'music boosts focus,' you can't agree or disagree until I tell you what 'focus' means in numbers. Pin the words down, or we're just trading opinions — exactly the thing we left behind last week."

Segment 3 — Three Research Designs, and What Each Can(’t) Claim (25 min)

Plain language first. There are three big families of study. They differ in how much control the researcher has — and that controls what you're allowed to say at the end.

① Descriptive research — what's happening? Goal: capture behavior as it is.
- Case study — an in-depth look at one person or small group (e.g., a single individual with rare brain damage). Rich and detailed — but you can't generalize from one case, and it can't establish cause.
- Naturalistic observation — watch behavior in its natural setting without interfering (kids on a playground, drivers at a stop sign). Great for real-world behavior; watch for the observer effect (people act differently when watched).
- Survey — ask many people questions. Fast and wide-reaching — but vulnerable to question wording ("Should the city finally fix the roads?" loads the answer) and sampling bias (who actually answered?).
- Bottom line: descriptive research describes. It can't explain why.

② Correlational research — are two things related? Goal: measure whether two variables move together.
- The correlation coefficient runs from −1 to +1. The sign is the direction; the absolute value is the strength.
- Positive (toward +1): they rise and fall together (hours studied ↑, grades ↑).
- Negative (toward −1): one goes up as the other goes down (hours of partying ↑, grades ↓).
- Near 0: no linear relationship.
- −0.85 is STRONGER than +0.30 — strength ignores the sign. (Flag this; it's a classic trap and a quiz item.)
- Bottom line: correlation finds a link, never a cause — for two reasons we'll hit in Segment 4.

③ Experimental research — does X cause Y? Goal: actually test cause and effect.
- Independent variable (IV) — the thing the researcher manipulates (the suspected cause).
- Dependent variable (DV) — the thing the researcher measures (the outcome).
- Experimental group gets the treatment; control group doesn't (or gets a placebo).
- Random assignment — every participant is sorted into a group by chance (coin flip / random number), so the groups start out roughly equal on everything else.
- Bottom line: the experiment is the only design that supports a cause-and-effect claim — because random assignment evens out the other explanations.

Memory hook:

"Describe = watch and report · Correlate = measure a link · Experiment = manipulate and compare. Only the last one earns the word cause."

Segment 4 — Correlation ≠ Causation + Quick Interaction (22 min) · Session 1 closes (~75)

Name the headline misconception out loud, then cure it:

  • ❌ "A strong correlation proves causation."
    ✅ Cure: even a perfect correlation can fail to be causal, for two reasons:
  • The third-variable problem — some unmeasured factor drives both. Ice-cream sales correlate with drowning deaths; the third variable is hot summer weather (more swimming and more ice cream). The two outcomes never touched each other.
  • The directionality problem — even if A and B are causally linked, a correlation can't tell you which way the arrow points. Stress correlates with poor sleep — but does stress wreck sleep, or does bad sleep cause stress? A correlation is silent on direction.
  • "A correlation is a clue, not a verdict. To claim cause, you need an experiment with random assignment."

The signature worked example (do it on the board — students will see it again in the tutorial):

Claim: "A study app raises grades."
- Correlational version: survey students; app users average higher grades than non-users. Tempting! But it's a link. The likely confound: more motivated students self-select into using the app. Motivation could be raising both app use and grades. We cannot say the app caused anything.
- Experimental version: take 200 students and randomly assign them — half use the app for a month, half don't — then compare final grades. Because assignment was random, the two groups started out similar on motivation and everything else, so a grade difference can be pinned on the app. Now — and only now — "the app causes higher grades" is a fair claim.
Land it: "Same question, two designs, two very different rights to the word 'cause.'"

Interaction — Think-Pair-Share (rapid-fire, ~8 min):
Put four "studies show" snippets on a slide; for each, students decide design (descriptive / correlational / experimental) and what it can conclude, solo (30 sec), compare with a neighbor (1 min), then vote. Suggested items: "People who eat breakfast have higher GPAs." · "Researchers randomly gave half a class a standing desk and measured focus." · "A psychologist spent a month logging how often shoppers say thank-you." · "Cities with more bookstores report more coffee shops."
(Answers, roughly: correlational/link only · experimental/can test cause · naturalistic observation/descriptive · correlational/third variable = city size.)

Segment 5 — Sampling vs. Assignment: The Two Randoms (22 min) · Session 2 opens

Hook back in: "Last session: only an experiment with random assignment earns the word cause. Today: the single most-confused pair of words in this whole unit — random sampling and random assignment. They sound alike. They do completely different jobs."

Plain language first.
- Population = the whole group you care about (e.g., all PSYC 1 students in the country). Sample = the smaller group you actually study.
- Random sampling — who gets into the study. Drawing your sample so that everyone in the population has an equal chance of being picked. Its job: a representative sample → results you can generalize to the population. Skip it and you get sampling bias (e.g., only surveying students leaving the library says little about students who never go).
- Random assignment — who gets the treatment. Once people are in the study, sorting them into experimental vs. control by chance. Its job: balance out confounds so a difference can be caused by the IV.

Memory hook (put it on a slide):

"Sampling = who's studied (generalize) · Assignment = who's treated (cause). One gets you out to the world; the other gets you in to causation."

Worked mini-example (do it out loud):

A team wants to know if a 10-minute meditation raises focus.
- Random sampling would mean every PSYC 1 student in the country had an equal shot at being one of the participants — so the finding generalizes nationally. (In practice most studies use a convenience sample — whoever's enrolled — which limits how far the result travels.)
- Random assignment means whoever is in the study gets flipped into "meditate" or "don't" by chance — so the focus difference can be pinned on the meditation.
"You can have one without the other. A campus study can randomly assign (clean cause) yet not randomly sample (limited generalizability)."

Reliability vs. validity (brief — ~4 min):
- Reliability = consistency. A measure gives the same result on repeat (a bathroom scale that reads the same weight twice in a row).
- Validity = accuracy. A measure captures what it claims to (a scale that reads your true weight, not 10 lbs heavy).
- "A broken scale can be perfectly reliable — consistently wrong by 10 pounds — and completely invalid. Consistency isn't accuracy."

Segment 6 — Research Ethics (the fully worked walk-through) (20 min)

Set it up: "We could learn a lot by lying to people or stressing them out. We don't — because participants are humans first and data second. Here's the floor every psychology study has to clear."

The core protections (one line each; put them on a slide):
- Informed consent — before agreeing, participants are told what the study involves, its risks, and that taking part is voluntary.
- Right to withdraw — anyone can stop at any time, for any reason, with no penalty (and may pull their data).
- Protection from harm — risk is minimized and kept no greater than everyday life; the researcher steps in if distress appears.
- Deception — only when justified + debriefing. Some questions collapse if people know the true aim (they'd change their behavior). Deception is allowed only as a last resort, must be cleared by review, and requires a debriefing afterward — telling participants the real purpose, why the deception was necessary, and undoing any distress.
- Confidentiality — data is kept private and reported so individuals can't be identified.
- IRB review — an Institutional Review Board (an independent ethics committee) must approve a study before it runs, weighing benefits against risks.
- Ethical treatment of animals — when animals are studied, their care and use are regulated to minimize suffering and justify the research.

One fully worked example (walk it through):

A researcher wants to study obedience — will people follow an authority's instructions even when uneasy? If participants knew the real aim, they'd act "appropriately" and the study would measure nothing. So the design uses deception (a cover story). To run it ethically: get IRB approval first; obtain informed consent to "a study on learning"; make clear participants may withdraw at any time; protect them from harm and stop if distress runs high; and debrief fully afterward — reveal the true purpose, explain why deception was needed, and check that everyone leaves okay. Confidentiality protects their identities in any report.
Land it: "Ethics isn't a hoop — it's the difference between studying people and using them."

Misconception + cure:
- ❌ "Psychology studies don't really need ethics oversight — they're just surveys and games."
✅ Cure: every study with human (or animal) participants goes through IRB review, informed consent, and debriefing when deception is used. The protections exist because past studies caused real harm.

Segment 7 — Why the Design Decides the Claim (18 min)

Plain language first. Pull the week together into one habit: read the design first, then judge the claim. The design is what gives a claim its ceiling.

The claim ceiling (put it on a slide):
- Descriptive → "Here's what we observed." (No cause, limited generalizing.)
- Correlational → "These two things are linked." (No cause — third variable & direction unknown.)
- Experimental → "X caused a change in Y" — if it had random assignment and a control group.

  • Operational definitions, again: when you read a headline, silently ask "how did they measure that?" "Happiness," "aggression," and "intelligence" mean nothing until someone says how they were counted.
  • Watch the verbs. "Linked to," "associated with," "predicts" = correlational language (a link). "Causes," "boosts," "leads to" = causal language that only an experiment earns. Headlines love to quietly upgrade a link into a cause.

Worked mini-example:

Headline: "Coffee drinkers score higher on memory tests." Design? Almost certainly correlational (nobody randomly assigned people to be lifelong coffee drinkers). So the honest read is a link, with obvious third-variable candidates — maybe coffee drinkers sleep less but study more, or are younger, or wealthier. To claim coffee improves memory you'd need to randomly assign people to coffee vs. no-coffee and measure memory. "The headline used a causal verb the study never earned."

Memory hook: "Design first, claim second. Match the verb to the method."

Segment 8 — Technology Workflow + AI-Critique, Callback & Hand-off (12 min) · Session 2 closes (~75)

Technology workflow — the "interpret-the-study" habit, on demand:
1. Find any "studies show" claim (a headline, an ad, a friend's post).
2. Ask three questions in order: What design was it? (descriptive / correlational / experimental) → What can it honestly conclude? (describe / link / cause) → What's a plausible third variable if it's correlational?
3. Check the verb against the method. If the design is correlational but the verb is "causes," the claim overreached.

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

Paste a correlational finding to an approved chatbot and ask it to interpret — for example: "A study found that teenagers who use social media more report feeling more lonely. What does this prove?"
Then check its work against today's lesson. If the model says social media causes loneliness, it overreached — the correct read is "a link; loneliness could drive more social-media use (directionality), or a third variable like poor sleep could drive both. You'd need a randomized experiment to claim cause." Your job all semester: the tool drafts, you judge. Catching an AI sliding from correlation to causation is exactly the skill this week builds.

Callback + tease:
- Callback: "Last week we said psychology is a science because it trusts evidence over intuition. This week is the fine print on the word evidence — the design decides what a study is allowed to say."
- Tease next week: "We've been talking about behavior from the outside. Next week we go inside the skull — neurons, neurotransmitters, and the brain structures that make all of this behavior possible."

Hand-off (the week's graded work):
- Lecture Tutorial 2 (AI tutor, share-link submission) — the scientific method, the three designs, correlation vs. causation, the two randoms, and research ethics.
- Quiz 2 (end of week) and Discussion 2 ("Correlation or Cause?" — interrogate a real "studies show" claim).
- Assignment 2 — identify IV/DV/design, critique a flawed study, classify designs, and explain a misleading headline to a friend.

Instructor FAQ — Common Stumbles

Student says / does Quick cure
"A strong correlation proves causation." Two reasons it doesn't: a third variable may drive both, and a correlation can't tell you the direction. Only a randomized experiment supports cause.
Treats random sampling and random assignment as the same thing. Sampling = who gets into the study (→ generalizing). Assignment = who gets the treatment (→ cause). Different jobs entirely.
Confuses IV and DV. The IV is what the researcher changes (the suspected cause); the DV is what they measure (the outcome). "I manipulate the IV; I depend on the DV to tell me what happened."
Thinks a vivid case study proves a general rule. A case study is rich but n = 1 — it generates hypotheses, it doesn't establish them.
Reads −0.85 as weaker than +0.30. Strength = absolute value. The sign is only direction. −0.85 is the stronger relationship.
"Surveys are objective — people just answer." Question wording and sampling bias shape survey results. Who answered and how you asked can flip the finding.
"Deception in research is just lying / never allowed." Deception is permitted only as a last resort, with IRB approval and a required debriefing afterward.
Says an experiment with no control group "proves" the treatment works. Without a control group (and random assignment), you can't rule out that people would have improved anyway. Compare against a baseline.
"Psychology research doesn't need ethics oversight." Every study clears IRB review, informed consent, and debriefing/withdrawal protections — precisely because past studies harmed people.

Scope flag

This outline stays within Objective 2 (research methods and ethics; correlation vs. causation; sources of bias). It deliberately keeps statistics to interpretation — what a correlation or a "significant" result means — and leaves the computation (calculating r, running significance tests) to the Introduction to Statistics course. The brain mechanisms behind behavior are Week 3; classic studies referenced here (e.g., obedience research) are named only to illustrate ethics and design, not taught in depth. The instructor and institution remain fictional; the research concepts and any real psychologists named are factual.

~ Prof. Bennett's edition · Fall 2026 · built with thecoursemaker.com